Abstracts of Oak Ridge Posters - Clinical Chemistry

Abstracts of Oak Ridge Posters Quantification of HIV-1 RNA Using a Homogeneous Single-Tube Assay, Rajesh Patel,1* Rinie van Beuningen,2 Dianne van Strijp,2 and Nurith Kurn1 (1 Dade Behring Inc., San Jose, CA 95135; 2 Organon Teknika, NL 5281 RM Boxtel, The Netherlands; * address correspondence to this author at: Advanced Diagnostics Group, Dade Behring Inc., San Jose, CA 95135; fax 408-239-2707, e-mail Rajesh_ [email protected]) Quantification of HIV RNA has become an essential tool in the management of HIV infection and monitoring of gene expression. Although direct quantification of nucleic acids has been demonstrated, the principal method of quantifying nucleic acids still involves nucleic acid amplification and quantification of the specific amplification products. Most commonly, quantification of nucleic acids involves either real-time monitoring of the amplification products [with fluorescently labeled probes (TaqMan; Molecular Beacons)] or postamplification quantification [e.g., electrochemiluminescence (Nuclisens™)] (1–3 ). Real-time monitoring increases the dynamic range of quantification, but it also leads to more complex instrumentation, and the absence of an internal control in the reaction may lead to erroneous quantification because of sample interference. Quantitative nucleic acid sequencebased amplification (NASBA™) is carried out using internal controls added to the sample, but the presence of three internal controls and a postamplification electrochemiluminescence detection scheme makes this method susceptible to contamination of samples with previously amplified products. We have combined highly sensitive NASBA with a highly sensitive chemiluminescent detection technology, Luminescent Oxygen Channeling Immunoassay (LOCI™), for homogeneous quantification of nucleic acids. NASBA is an isothermal exponential nucleic acid amplification system, based on an activity of avian myeloblastosis virus reverse transcriptase, RNase H, and T7 RNA polymerase, which predominantly produces a singlestranded RNA product. The specificity of HIV-1 amplification and detection is ensured by the use of specific primers and is further enhanced by the use of specific probe sequences for detection. The product concentration can be as high as 8.0 ␮mol/L after a typical NASBA reaction. Because the amplification product is singlestranded, there is no need to denature at the end of amplification for the binding of specific probes for either detection or quantification. Addition of an internal control (Qa), which bears sequence resemblance and is coamplified with the same primers as the wild type (WT), allows amplification of WT and Qa with equal efficiency. The ratio of the two amplicons at the end of amplification reflects the ratio of the two targets, WT and Qa, present at the beginning of amplification. The two amplicons competitively bind to a common probe added to the amplification reaction. The distribution of the common probe between the WT and Qa amplicons is determined by LOCI. The ratio of the WT- and Qa-specific LOCI signals obtained from any NASBA/LOCI reaction reflects the

ratio of the two amplicons present in the reaction. The WT analyte present in the sample can be easily determined from the ratio of the WT and Qa amplification products and the amount of Qa target added to the reaction. LOCI is a highly sensitive homogeneous chemiluminescent assay involving a pair of ⬃200-nm latex bead particles: one dyed with an olefin (chemiluminescent particle) and the other dyed with a phthalocyanine photosensitizer dye (4 ). When excited by 680 nm light, the photosensitizer particle emits singlet oxygen (1O2) with a half-life of ⬃4 ␮s. Formation of analyte-specific bead pairs between the sensitizer and chemiluminescent particles allows an efficient transfer of 1O2 to the chemiluminescent particles, where it reacts with the olefin and generates a chemiluminescent signal with a half-life in the range of 0.5–30 s. In the absence of an analyte, the 1O2 decays in the medium without generating any chemiluminescent signal. In the present assay for nucleic acid detection and quantification, the LOCI particles are coated with specific oligonucleotide sequences that are capable of binding to their complementary sequences present on the probes. Quantitative NASBA/LOCI is based on homogeneous and simultaneous detection of two analytes in a single tube. The two signals are obtained by using a pair of chemiluminescent particles, which were basically prepared as described by Ulman et al. (4 ) except that the first chemiluminescent particle was dyed with thioxene and 9,10-diphenyl anthracene (DPA) and the second particle was dyed with 2H-1,4-oxazine-3,4-dihydro-4,5-diphenyl6-dimethylaminophenyl and Rubrene (N-phe). The DPA particle emits at ⬃400 nm with a half-life of ⬃1.0 s, and the N-phe particle emits ⬃560 –590 nm with a half-life of ⬃30 s. The DPA particle binds to the WT-specific probe for the WT-specific HIV amplicon generated, whereas the N-phe particle binds to the Qa-specific probe (Fig. 1A). In addition to the specific probes, both amplicons compete for binding to a common probe added to the reaction, which in turn binds to a sensitizer bead. The trimolecular complex of the amplicon strand and the two probes hybridizes to oligonucleotides on the corresponding sensitizer and chemiluminescent latex particles. The ratio of the WT to Qa amplicons bound to the sensitizer particles is reflected by the ratio of the two LOCI signals generated in the assay. Detection of the two distinct chemiluminescent signals generated by the amplicon-bound bead pairs was carried out in a LOCI instrument (built in-house) that irradiates the sample tube with a 680 nm laser and collects the signal in two phases. The DPA emission is collected using a 395– 415 nm bandpass filter. The N-phe signal is collected after the entire DPA signal has decayed, using a 550 – 650 nm bandpass filter. The DPA signal is corrected for the contributing N-phe signal by use of the known decay time constant of N-phe. Initially, the HIV WT and Qa amplicons generated using the NASBA technology were quantified using the LOCI probes and beads. The LOCI signal was linear over 3 logs of amplicon concentrations. With LOCI, ⬃3 ⫻ 106 amplicons could be detected in a 50-␮L reaction with a

Clinical Chemistry 46, No. 9, 2000

1487

1488

Abstracts of Oak Ridge Posters

signal-to-background ratio of 1.7. The effect of LOCI reagents on NASBA was also assessed, and no inhibition of amplification efficiency was observed. LOCI reagents, consisting of the three LOCI oligonucleotide-coated particles and the three probes, were added to the NASBA Amplification kit (Organon Teknika), which included the HIV-specific primers. HIV target and Qa internal controls were added at appropriate concentrations, and amplification was carried out for 90 min as described in the package insert. The reaction tubes were then introduced into the LOCI reader, and the two chemiluminescent signals were collected as described. The WT signal was first obtained by irradiating the reaction tubes with a 680 nm laser for 1.0 s and then collecting the chemiluminescent signal for 0.5 s. This was repeated for a total of three times, after which the DPA signal was allowed to decay for 30 s before the Qa signal from the slow-decaying N-phe particle was collected for 10 s. The following algorithm was used to quantify the WT target present in the sample: Log HIV quantified (QT) ⫽ v{log [Qa/(WT ⫹ Qa)]} ⫹ w ⫹ log[WT/(WT ⫹ Qa)], where WT and Qa are

the corrected chemiluminescent LOCI signals. Ideally, the value of v is ⫺1, and w is the log of Qa added to the reaction. However, v and w are derived from a fit-to-astandard curve for a specific combination of probes, LOCI reagents, and assay conditions. Purified WT target and Qa were used to optimize the concentration of LOCI particles and the relevant probes for efficient quantification of the HIV target sequence. This assay was used to quantify HIV-1 viral RNA (1 ⫻ 102 to 1 ⫻ 106 copies/mL) from plasma samples (Fig. 1B). HIV-1 plasma calibrators (Advanced Bioscience Laboratories) were used to prepare nucleic acid extracts based on Nuclisens extraction (Organon Teknika). The calibrator used is an HIV-1 producing cell line, which has been quantified with several technologies. The concentration of the HIV-1 calibrator used is equivalent to VQC (Red Cross Blood Banks, The Netherlands) and WHO standards. The NASBA/LOCI detection limit of 320 copies per Nuclisens extraction converts into an amplification analytical sensitivity of 6 HIV-1 copies, based on the use of only 4% of the extract volume in the amplification and the typical nucleic acid recovery of 50% (data not shown). Quantification was carried out using the NASBA/ electrochemiluminescence (ECL) or the NASBA/LOCI method. The log of the average intra-experiment imprecision for NASBA/LOCI was 0.16 (n ⫽ 6 duplicates with input above 1 ⫻ 103 viral copies/mL). The accuracy of the assay, based on the use of known quantities of HIV-1 VQC calibrators, was 0.09 log (n ⫽ 6 duplicates) for NASBA/LOCI and 0.25 log for NASBA/ECL (n ⫽ 7 single measurements). The correlation between both quantification methods was 0.988. The NASBA/LOCI assay procedure described is a simple method of nucleic acid quantification with a dynamic range of three orders of magnitude and a lower limit of detection of 1 ⫻103 HIV copies/mL. The system provides for containment and is easily amenable to automation. The single end-point readout allows simple instrumentation. In addition, LOCI provides an ultrasensitive homogeneous method for direct detection of nucleic acids with a limit of detection of 2 ⫻ 102 to 4 ⫻ 102 molecules.

References

Fig. 1. Quantitative NASBA/LOCI. (A), in NASBA, DPA (DP) particles bind to the WT-specific probe (W) for the WT-specific HIV amplicon generated (top), whereas N-phe (n-phe) particles bind to the Qa-specific probe (bottom). Amplicons also bind to a common probe, which then binds to the sensitizer (Sen). Em, emission. (B), quantification of HIV-1 viral RNA from plasma samples. 䡺, NASBA/ECL quantification; u, NASBA/LOCI quantification. Bars, SD.

1. Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res 1996;4:357– 62. 2. Lewin SR, Vesanen M, Kostrikis L, Hurley A, Duran M, Zhang L, et al. Use of real-time PCR and molecular beacons to detect virus replication in human immunodeficiency virus type 1-infected individuals on prolonged effective antiretroviral therapy. J Virol 1999;73:6099 –103. 3. van Gemen B, van Beuningen R, Nabbe A, Jurrians S, Lens P, Kievits T. A one-tube quantitative HIV-1 RNA NASBA nucleic acid amplification assay using electrochemiluminescent (ECL) labeled probes. J Virol Methods 1994; 49:157– 68. 4. Ulman EF, Kirakossian H, Switchenko AC, Ishkanian J, Ericson M, Wartchow CA, et al. Luminescent oxygen channeling assay (LOCI™): sensitive, broadly applicable homogenous immunoassay method. Clin Chem 1996;42:1518 – 26.


Reflectometry Interference Spectroscopy in Detection of Hepatitis B Surface Antigen, Fang Yu, Danfeng Yao, and Weiping Qian* (National Laboratory of Molecular and Biomolecular Electronics, Southeast University, Nanjing 210096, Peoples Republic of China; * author for correspondence: fax 86-25-7712719, e-mail [email protected]) Optical biosensors are of increasing interest for real-time detection of analytes without the use of additional labeled reagents. The sensing system is based on different interfacial physical properties that change upon the binding reaction. Combining high sensitivity and selectivity, labelfree biosensors have the potential to provide low-cost detection and measurement technology for accurate and highly specific quantification of low-concentration analytes. Reflectometry interference spectroscopy (RIfS) is such a typical optical system that has been systematically developed and investigated by Brecht et al. (1 ). Based on the white-light interference effects occurring at thin transparent films, RIfS has shown high precision and high stability in monitoring the change of optical thickness attributable to interfacial molecular interaction and has been shown to be a useful approach for the detection of herbicides in river water and hydrocarbons in air (1, 2 ). At present, we are trying to find possible applications of RIfS in clinical diagnosis. ELISAs, the widely used label immunoassays in clinical laboratories, have inherent drawbacks, such as numerous pipetting and incubation steps. Additionally, the technique does not always show dose dependence in an extensive concentration range of analytes, exemplified by the so called high-dose hook effect (3 ). One of important factors that induces the high-dose hook effect is the introduction of labeled antibody, so it is expected that improved performance can be gained with RIfS technology, which is a label-free measurement. We describe here a strategy based on RIfS and spincoating, which can in principle eliminate several steps of the traditional ELISA. The molecule-substrate and intermolecule interactions can be recorded in real time with relative high sensitivity and reproducibility. The optical principle of RIfS measurement has been described in detail by Brecht et al. (4 ). The change of optical thickness caused by interfacial binding can be probed with a spectrometer, using white-light interference. In our apparatus, a CCD spectrometer (PC-1000; Ocean Optics Inc.) is used instead of a diode array spectrometer to decrease size. The optics use Y-shaped reflective optical fibers to illuminate and collect the reflected light of the transducer chip on a windowless Teflon cell (flow cell) with volume of ⬃30 ␮L. Self-developed software is used for the realtime acquisition of full spectra and online analysis. Solution sample handling was carried out using a peristaltic pump with a minimal flow rate of 100 ␮L/min. A schematic of the transducer is shown in Fig. 1A. A 10-nm Ta2O5 film was deposited on glass substrate to enhance the interference contrast (5 ). After the above treatment, polystyrene was spin-coated on the chip at 2000 rpm and dried in 60 °C for 1 h. A transparent

1489

polystyrene layer with thickness of ⬃650 nm was then generated and served as both the interference film and the hydrophobic surface for immobilization of antibodies. In our experiments, yeast hepatitis B surface antigen (HBsAg) and its monoclonal antibodies (anti-HBs) were chosen as a model system. The adsorption of antibody occurred in self-designed Teflon cells mounted on the polystyrene-coated chips. Anti-HBs was dissolved in 50 mmol/L carbonate buffer, pH 9.6, at a concentration of 5 mg/L, and 200 ␮L of this solution was pipetted into each cell. After overnight incubation at 4 °C, the chips were subsequently blocked for 1 h at 37 °C with carbonate buffer supplemented with 10 g/L bovine serum albumin (BSA) and then washed five times with carbonate buffer before use. With a “blank” chip, the coating and blocking process can be monitored by RIfS in real time at room temperature (20 ⫾ 2 °C). When the anti-HBs molecules were physically adsorbed onto a polystyrene film, an increase of ⬃3.4 nm in total optical thickness was detected. The following blocking process with BSA increased thickness only ⬃0.3 nm because of the small molecular weight of BSA and the dense distribution of antibody molecules previously adsorbed on the polystyrene surface. Immunoreaction was conducted at room temperature. The chips were first rinsed with 50 mmol/L phosphatebuffered saline (pH 7.4) until a stable prerun baseline was recorded. The standard deviation of the baseline fluctuation was ⬃17.8 pm. A series of antigen solutions (0.1, 1, 10,

Fig. 1. Schematic of the RIfS transducer (A) and correlation between association rate constant (Ks) and HBsAg concentrations (B). Bars, SD.

1490


100, 250, 1000, and 5000 ␮g/L in phosphate-buffered saline with 500 mg/L BSA added) were then injected at 400 ␮L/min, and each binding curve was recorded. The curves were analyzed with the “CurveExpert” system (Ver. 1.34), using a pseudo-first-order exponential association model in an effort to determine each association rate (Ks) and hence the calibration curve. As negative control, BSA solution at a concentration of 1000 mg/L was injected and had only a negligible effect, indicating the specificity of the RIfS detection of immunoreaction. The RIfS signal was totally obscured by noise when the concentration of HBsAg sample was 1 ␮g/L. The calibration curves extended into the ␮g/L range (Fig. 1B). The interaction between HBsAg and immobilized Anti-HBs in BSA buffer provided a good signal at 10 ␮g/L and a linear window range of 20 –5000 ␮g/L (Fig. 1B). By contrast, conventional ELISA has a lower linear range, usually 1–100 ␮g/L for HBsAg quantitative analysis. However, the present RIfS is by no means optimized. It is expected that substantial improvement can be obtained by the use of a compensator to carefully control the system noise. Application of RIfS in the clinical detection of HBsAg is possible with the advantages of direct detection of analytes, shorter measurement time (a few minutes), no interference from the mobile phase, a favorable linear range, and reusability of the transducer.

This work was supported by the Natural Science Foundation of China. References 1. Brecht A, Piehler J, Lang G, Gauglitz G. A direct optical immunosensor for atrazine detection. Anal Chim Acta 1995;311:289 –99. 2. Kraus G, Gauglitz G. A reflectometric sensor for ammonia and hydrocarbons. Fresenius Z Anal Chem 1993;346:572– 6. 3. Fernando SA, Wilson GS. Multiple epitope interaction in the two-step sandwich immunoassay. J Immunol Methods 1992;151:67– 86. 4. Brecht A, Gauglitz G, Polster J. Interferometric immunoassay in a FIA system: a sensitive and rapid approach in label-free immunosensing. Biosens Bioelectron 1993;8:387–92. 5. Brecht A, Gauglitz G. Optimized layer systems for immunosensors based on the RIFS transducer. Fresenius J Anal Chem 1994;349:360 – 6.

Purification of Prostate-specific Antigen from Serum by Indirect Immunosorption and Elution with a Hapten, Wolfgang Hoesel,1* Jochen Peter,2† Helmut Lenz,1 and Carlo Unverzagt2‡ (1 Roche Diagnostics GmbH, Nonnenwaldstrasse 2, 82372 Penzberg, Germany; 2 Technical University Munich, 85748 Garching, Germany; * author for correspondence: fax 49-8856-603341, e-mail wolfgang.hoesel@ roche.com; † present address: National Institute of Environmental Health Sciences (NIH/NIEHS), Bldg. 101, Room F011, Research Triangle Park, NC 27709; ‡ present address: Lehrstuhl fu¨r Bioorganische Chemie, Universita¨t Bayreuth, Gebaüde NW 1, 95440 Bayreuth, Germany) Preparative purification of macromolecules, e.g., proteins, has reached a high standard because of efficient separa-

tion materials that are available for this task. However, this is limited to samples containing substantial amounts of the desired protein. When proteins that are present only in minute amounts (e.g., in the ␮g/L range) must be isolated from a limited supply of complex biological material (e.g., human serum), a one-step method is recommended to obtain the protein in sufficient yield. When suitable antibodies are available, immunopurification often is the method of choice (1 ), but nonspecific binding of proteins to affinity matrices is a substantial problem because elution with acid or chaotropic agents very often washes off the impurities as well as the analyte (2 ). Thus, there is a need for methods that allow the specific elution of the desired proteins without eluting the impurities. Specific elution by competing with a low-molecular weight analog is appealing, but suitable analogs may not be available. We wanted to isolate prostate-specific antigen (PSA) from serum and analyze its structure by mass spectrometry. The concentration of PSA in serum from patients with prostate cancer (PCa) ranges from ⬃3 to ⬎3000 ␮g/L, ⬃105–107 times less than that of other serum proteins, e.g., albumin. We have developed a general indirect immunosorption method that follows to a certain extent an immunoassay principle developed by Hashida et al. (3 ) and Ishikawa et al. (4 ) and makes use of a digoxigenylated anti-analyte antibody. Initially, the PSA from the biological sample is bound to magnetic beads by an array of antibodies. The key step is the competitive release of the PSA-antibody pair by digoxigenin-lysine under neutral conditions. These conditions leave the impurities almost entirely bound to the matrix and yield the antibody-PSA complex in a pure form (5 ). When we began isolating PSA from serum, we first tried a standard immunosorption method using magnetic beads combined with acidic elution. To this end, streptavidin-coated magnetic beads were loaded with a biotinylated anti-PSA monoclonal antibody (Mab) and incubated with serum from a PCa patient. After magnetic collection of the beads and several washing steps, the bound protein was released by a small amount of a mixture of either formic acid-water-acetonitrile (1:3:2, by volume) or 1 mol/L propionic acid, and the eluted material was analyzed by nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and matrix-assisted laser desorption/ionization time-offlight mass spectrometry (MALDI-TOF MS). It was found that mainly nonspecifically bound serum proteins were liberated from the beads (see lanes A1–A4 in Fig. 1). The high degree of impurities did not allow the unambiguous detection of free PSA and the PSA/␣1-antichymotrypsin (ACT) complex. This applied especially to the elution with formic acid-water-acetonitrile, a solvent used for preparing the matrix solution for MALDI-TOF MS and which would therefore be very suitable for that type of analysis. But when the elution mixture with this solvent was applied to the MALDI-TOF MS, almost all of the proteins detected in the mass spectrum were impurities. Free PSA was barely detectable, and the PSA/ACT complex could


not be seen at all. Because the direct immunosorption method using acidic elution was not suitable for the analysis of proteins either by SDS-PAGE or directly by MALDI-TOF MS, we reasoned that we should use a specific elution step at neutral pH to avoid coelution of nonspecifically bound proteins. Again using streptavidin-coated magnetic beads, we constructed a complex consisting of biotinylated antidigoxigenin IgG/digoxigenylated anti-PSA IgG/PSA on the beads. Specific elution of the complex consisting of digoxigenylated anti-PSA IgG/PSA was accomplished using a solution of a digoxigenin-lysine conjugate (2.5 g/L) at pH 7.3. The eluted material was analyzed in comparison with that eluted by acid (see lanes B1 and B2 in Fig. 1). Below a molecular mass of 100 kDa, only the target analytes PSA and PSA/ACT were present in the indirect immunosorption eluates, except for a protein band at 65 kDa, which very likely represents albumin. Evidently, even these very mild conditions elute small amounts of this abundant serum protein from the beads. The protein bands above 100 kDa consisted mostly of the digoxigenylated anti-PSA antibody and did not interfere with separation of the analytes because of their high molecular masses. Analysis of free PSA and PSA/ACT complex in the eluates by MALDI-TOF MS revealed that except for serum albumin, only the target analytes could be detected in the spectra (free PSA, 28.3 kDa; PSA/ACT, 83 kDa). In the first experiments, an analyte recovery of ⬃20%

1491

was obtained. It became obvious that the affinity of the anti-digoxigenin antibody used was a key factor because most of the losses occurred when the beads were washed. By switching to a different anti-digoxigenin antibody with higher affinity and lower Koff rate, the overall yield could be increased to ⬃40%. A second set of antibodies was also evaluated, which consisted of biotinylated anti-ruthenium-bispyridyl Mab and anti-PSA Mab labeled with ruthenium-bispyridyl complex. With this system, a recovery of ⬃60% could be obtained when the free rutheniumbispyridyl complex was used for elution. The improved yield in comparison to the digoxigenin/anti-digoxigenin system was very likely attributable to the better binding characteristics of the anti-rubispyridyl Mab and a higher concentration of the more water-soluble hapten in the competitive elution step. The comparatively hydrophobic digoxigenin-lysine molecule could also be interfering with the crystallization step in the MALDI matrix preparation, so it should probably be removed (e.g., by dialysis) before the crystallization. This is probably not necessary when the rubispyridyl complex is used. After removal of the free PSA from a PCa serum, the PSA/ACT complex was cleaved by treatment with ethanolamine under alkaline conditions. The released PSA was then isolated by immunosorption as described above, and the intact PSA as well as the peptides obtained by an endoproteinase Lys C digest were analyzed by MALDITOF MS and compared with PSA from seminal fluid. The intact PSA as well as the peptides obtained after digestion (⬃80% coverage of the sequence) did not reveal any structural difference between the PSA released from the PSA/ACT complex of human serum and PSA from seminal fluid. These data were published recently in this Journal (6 ). The indirect immunosorption method described above allows the isolation of small amounts of analytes from complex biological material (e.g., serum) in a form that is almost free of impurities. The method is generally suitable if a ligand of high affinity for an analyte is available (e.g., antibody, receptor, lectin, or aptamer) that can be haptenylated. It is especially suitable for the isolation of proteins for analysis by SDS-PAGE and mass spectrometry (e.g., MALDI-TOF MS of the proteins or their peptides).

References

Fig. 1. SDS-PAGE of direct (Lanes A1–A4) and indirect (Lanes B1 and B2) immunosorption eluates from PCa and control sera. Lanes A1 and A2, elution with formic acid-water-acetonitrile (1:3:2, by volume); lanes A3 and A4, elution with 1 mol/L propionic acid; lanes B1 and B2, elution with digoxigenin-lysine. Lanes A1, A3, and B1, serum blank (no PSA present); lanes A2, A4, and B2, PCa serum (containing free PSA and PSA/ACT).

1. Jack GW. Immunoaffinity chromatography. In: Kenney A, Fowell S, eds. Methods in molecular biology 11: practical protein chromatography. Totowa, NJ: Humana Press, 1992:125–71. 2. Hermanson GT, Mallia AK, Smith PK. Immobilized affinity ligand techniques. San Diego: Academic Press, 1992;307–11. 3. Hashida S, Tanaka K, Kohno T, Ishikawa E. Novel and ultrasensitive sandwich enzyme immunoassay (sandwich transfer enzyme immunoassay) for antigens. Anal Lett 1988;21:1141–54. 4. Ishikawa E, Kohno T. Development and applications of sensitive enzyme immunoassay for antibodies: a review. J Clin Lab Anal 1989;3:252– 65. 5. Peter J, Unverzagt C, Hoesel W. Purification of prostate-specific antigen from human serum by indirect immunosorption and elution with a hapten. Anal Biochem 1999;273:98 –104. 6. Peter J, Unverzagt C, Hoesel W. Analysis of free prostate-specific antigen (PSA) after chemical release from the complex with ␣1-antichymotrypsin (PSA-ACT). Clin Chem 2000;46:474 – 82.

1492


Application of Bispecific F(abⴕ)2␮ Fragments Prepared from IgMs against Carcinoembryonic Antigen and Alkaline Phosphatase, Koichi Morimoto1* and Kuniyo Inouye2 (1 Department of Biotechnological Science, Kinki University, 930 Nishi-mitsuya, Uchita, Naga-gun, Wakayama 649-6493, Japan; 2 Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto 6068502, Japan; * author for correspondence: fax 81-736-774754, e-mail [email protected]) Bispecific monoclonal antibodies (BsMAbs) have two different binding regions that simultaneously recognize two different antigens (1, 2 ). BsMAbs have been studied for use in immunodiagnosis (3, 4 ) and immunotherapy (5, 6 ). Most BsMAbs are derived from the IgG class, whereas most hybridomas raised against carbohydrate antigens, including the tumor-associated antigens, are IgM MAb producers. IgM antibody-enzyme conjugates are of limited use in enzyme immunoassays because of steric effects, reduction of their activities, and high nonspecific binding. For these reasons, BsMAbs prepared from IgM MAbs have been especially used for cancer immunodiagnosis. Their production by hybrid-hybridoma fusion is one method (7 ), but it is not an effective one. F(ab⬘)2␮ fragments have been successfully prepared from their IgMs, and their performance has been evaluated (8 –11 ). In this report, homogeneous bispecific F(ab⬘)2␮ fragments (BsF␮ fragments) were efficiently prepared from IgMs. Mouse IgM monoclonal antibodies 9CA10 and APM05 recognize carcinoembryonic antigen (CEA) and calf intestine alkaline phosphatase (AP), respectively. Their F(ab⬘)2␮ fragments were prepared by pepsin digestion and purified by hydrophobic interaction HPLC (8, 10, 11 ). The purities of the fragments from the IgMs were ⬎96% by size exclusion HPLC using a TSKgel G3000SWXL column (7.8 mm i.d ⫻ 30 cm; Tosoh), and the yields were 54% (APM05) and 53% (9CA10), respectively. The F(ab⬘)2␮ of 9CA10 was reduced to Fab⬘␮-SH with 50 mmol/L 2-mercaptoethylamine (Sigma) in phosphate-buffered saline (PBS) containing 10 mmol/L EDTA at 37 °C for 1 h, and was separated from 2-mercaptoethylamine by size exclusion HPLC on a TSKgel G3000SWXL column. Immediately after HPLC, the Fab⬘␮-SH was converted to a thionitrobenzonic acid (TNB) derivative with dithionitrobenzene, and Fab⬘␮-TNB was applied to a TSKgel G3000SWXL column to remove any excess TNB and dithionitrobenzene. Fab⬘␮-TNB (9CA10) was reacted with Fab⬘-SH (APM05) to form SOS bonds between Fab⬘␮ (9CA10) and Fab⬘␮ (APM05) at a molar ratio of 2.5:1, followed by incubation at 4 °C for 16 h under N2 gas. Ammonium sulfate (70% saturation) was added to the reaction mixture, and the mixture was centrifuged at 10 000g for 20 min. The precipitate was dissolved with 50 mmol/L phosphate buffer (pH 7.4) containing 2.0 mol/L ammonium sulfate. The solution was applied to a TSKgel Ether-5PW column (7.5 mm i.d. ⫻ 7.5 cm; Tosoh) equilibrated with the same buffer and eluted with a linear gradient of ammonium sulfate from 2.0 to 0 mol/L at a flow rate of 0.5 mL/min. Three peaks were eluted at 32

min (peak 1), 40 min (peak 2), and 46 min (peak 3). The proteins of peaks 1 and 3 were Fab⬘␮-TNB and Fab⬘␮-SH, respectively. The protein eluted in peak 2 showed immunoreactivities against both antigens (data not shown) and was identified as BsF␮. The elution time of peak 2 suggests that the hydrophobicity of the BsF␮ fragment is the average of those of the parental F(ab⬘)2␮ fragments. The BsF␮ fragments were purified to homogeneity in a onestep procedure using hydrophobic interaction HPLC (Fig. 1A), and the final yields of the BsF␮ fragments were 35– 40% of the theoretical values. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed in a 6% slab gel under nonreducing conditions (12 ), and the proteins were stained with Coomassie Brilliant Blue R-250. The BsF␮ fragment and its parental F(ab⬘)2␮ fragments showed a similar migration pattern (144 –146 kDa) during SDS-PAGE, suggesting that the BsF␮ is composed of four chains [two pairs of a truncated heavy chain and an intact light chain derived from the parental F(ab⬘)2␮].

Fig. 1. Purification of BsF␮ by hydrophobic interaction HPLC with TSKgel Ether-5PW (A), and the immunoreactivity of BsF␮ against CEA (B). (A), elution was monitored by measuring absorbance at 280 nm. The BsF␮ peak is indicated by an arrow. (B), 100 ␮L of anti-CEA IgM covalently labeled with AP (䡺), anti-CEA F(ab⬘)2␮ fragment covalently labeled with AP (E), or BsF␮ bound to AP (‚) was added to the well, followed by the addition of 100 ␮L of CEA solution at the concentration indicated on the x axis. p-Nitrophenylphosphate was added, and the absorbance at 405 nm was measured.


The dissociation constants (KD) of the BsF␮ and its parental F(ab⬘)2␮ fragments with their specific antigens were measured by surface plasmon resonance using a BIAcore biosensor (Pharmacia Biosensor). The BsF␮ and its parental F(ab⬘)2␮ fragments were injected separately and immobilized on the chip coated with goat anti-mouse (IgG ⫹ IgM) antibodies (Jackson ImmunoResearch Laboratories). When CEA (50 mg/L; Scripps Laboratories) was injected on the chips coated with BsF␮ or its parental F(ab⬘)2␮ fragments, the intensities increased 29 and 30 relative response units, respectively. The intensities of BsF␮ (50 mg/L) and its parental F(ab⬘)2␮ fragments (50 mg/L) against AP increased 23 and 24 relative response units, respectively. For CEA, the KDs were 249 nmol/L (BsF␮) and 212 nmol/L [F(ab⬘)2␮ (9CA10)], respectively. The KDs against AP were 13 nmol/L (BsF␮) and 10 nmol/L [F(ab⬘)2␮ (APM05)], respectively. The immunoreactivity of BsF␮ bound to AP was evaluated by sandwich enzyme immunoassay and compared with that of the F(ab⬘)2␮ covalently conjugated to AP. A 96-well microtiter plate (Nunc-intermed; MaxiSorp) was coated with goat anti-CEA polyclonal antibody (prepared in our laboratory) and blocked by incubation with 2 g/L bovine serum albumin in PBS (pH 7.4). After washing, the plate was incubated at 25 °C for 60 min with 100 ␮L of CEA at various concentrations (0, 25, 50, 100, 200, 400 ␮g/L) and 100 ␮L of the BsF␮ bound to AP (absorbance at 280 nm in PBS ⫽ 0.004), IgM covalently labeled with AP (absorbance at 280 nm in PBS ⫽ 0.002), or F(ab⬘)2␮ covalently labeled with AP (A280 nm in PBS ⫽ 0.004). After another washing, 100 ␮L of the substrate solution (2 g/L p-nitrophenylphosphate in 50 mmol/L carbonate buffer, pH 9.5, containing 10 mmol/L MgCl2) was added to the plate for 20 min at 25 °C, and the reaction was terminated by adding 100 ␮L of 0.5 mol/L NaOH. The AP activity was measured by the absorbance at 405 nm (Fig. 1B). The detection limits of the CEA assay (defined as 2 SD above the zero concentration calibrator) were 0.90 ␮g/L for BsF␮ and 3.4 ␮g/L for F(ab⬘)2␮, respectively. The immunoreactivity of BsF␮ against CEA was almost the same as that of the F(ab⬘)2␮ covalently labeled with AP in the range 0.90 – 400 ␮g/L. The nonspecific binding of BsF␮ was 27% (0.90 ⫻ 100/3.4 ⫽ 27) lower than that of F(ab⬘)2␮. Because the covalently labeled conjugate that was prepared from F(ab⬘)2␮ and AP is a complex polymer molecule and has multiple antigen binding sites in one conjugate molecule, the conjugate cannot bind to the antigen at a molar ratio of 1:1. The BsF␮ bound to the AP conjugate can bind more efficiently to the antigen at a 1:1 molar ratio. To detect lower concentrations of the antigen, BsF␮ may be more useful than the covalently labeled conjugate. The BsF␮ bound to the AP conjugate was stable for at least 3 months at 4 °C. In conclusion, we showed that homogeneous BsF␮ fragments prepared from IgMs recognized both antigens without any loss in immunoreactivity. The BsF␮ bound to AP allowed sensitive and reliable measurement of the antigen and may be generally applicable for immunodiagnosis.

1493

References 1. Milstein C, Cuello AC. Hybrid hybridomas and their use in immunohistochemistry. Nature 1983;305:537– 40. 2. Brennan M, Davison PF, Paulus H. Preparation of bispecific antibodies by chemical recombination of monoclonal immunoglobulin G1 fragments. Science 1985;229:81–3. 3. Suresh MR, Cuello AC, Milstein C. Advantages of bispecific hybridomas in one-step immunocytochemistry and immunoassays. Proc Natl Acad Sci U S A 1986;83:7989 –93. 4. Morimoto K, Inouye K. A sensitive enzyme immunoassay of human thyroidstimulating hormone (TSH) using bispecific F(ab⬘)2 fragments recognizing polymerized alkaline phosphatase and TSH. J Immunol Methods 1997;205: 81–90. 5. Posey JA, Raspet R, Verma U, Deo YM, Keller T, Marshall JL, et al. A pilot trial of GM-CSF and MDX-H210 in patients with erbB-2-positive advanced malignancies. J Immunother 1999;22:371–9. 6. Koelemij R, Kuppen PJ, van de Velde CJ, Fleuren GJ, Hagenaars M, Eggermont AM. Bispecific antibodies in cancer therapy, from the laboratory to the clinic. J Immunother 1999;22:514 –24. 7. Takahashi M, Fuller SA. Production of murine hybrid-hybridomas secreting bispecific monoclonal antibodies for use in urease based immunoassays. Clin Chem 1988;34:1693– 6. 8. Morimoto K, Inouye K. Method for the preparation of bispecific F(ab⬘)2␮ fragments from mouse monoclonal antibodies of the immunoglobulin M class and characterization of the fragments. J Immunol Methods 1999;224: 43–50. 9. Morimoto K, Inouye K. Flow cytometric analysis of sialyl Lewis A antigen on human cancer cells by using F(ab⬘)2␮ fragments prepared from a mouse IgM monoclonal antibody. Cytotechnology 1997;24:219 –26. 10. Inouye K, Morimoto K. Single-step purification of F(ab⬘)2␮ fragments of mouse monoclonal antibodies (immunoglobulins M) by hydrophobic interaction high-performance liquid chromatography using TSKgel Ether-5PW. J Biochem Biophys Methods 1993;26:27–39. 11. Inouye K, Morimoto K. Preparation of F(ab⬘)2␮ fragments from rat IgM monoclonal antibodies and their application to the enzyme immunoassay of mouse interleukin-6. J Immunol Methods 1994;171:239 – 44. 12. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680 –5.

Toward Reagent-free Clinical Analysis: Quantitation of Urine Urea, Creatinine, and Total Protein from the Mid-Infrared Spectra of Dried Urine Films, R. Anthony Shaw,1* Sarah Low-Ying,1 Michael Leroux,2 and Henry H. Mantsch1 (1 National Research Council of Canada, Institute for Biodiagnostics, 435 Ellice Ave., Winnipeg, Manitoba, R3B 1Y6 Canada; 2 Department of Clinical Chemistry, Health Sciences Centre, 820 Sherbrook St., Winnipeg, Manitoba, R2H 2A6 Canada; * author for correspondence: fax 204-984-5472, e-mail [email protected]) Infrared (IR) spectroscopy offers an approach to clinical analysis that is conceptually very appealing. Whereas countless assays rely on the use of chemical agents to “recognize” the analyte of interest and to react with the analyte to produce specific color changes, IR-based analysis is founded on the rich IR absorption patterns that characterize the analytes themselves. These absorption patterns provide the basis to distinguish among the constituents and to separately quantify them. The most obvious distinguishing feature is that no reagents are required. In addition, IR-based analytical methods require very small sample volumes (typically microliters), show good precision over the entire physiological range, and are well suited for automation. Several previous studies have illustrated potential roles for IR spectroscopy in the clinical laboratory. For example,

1494


six serum analytes have been shown to be suitable for IR-based analysis, namely albumin, total protein, glucose, triglycerides, urea, and cholesterol (1– 6 ). Studies of amniotic fluid have yielded IR models to quantify the lecithin/sphingomyelin ratio and the surfactant/albumin ratio, establishing IR spectroscopy as an attractive option for the assessment of fetal lung maturity (7, 8 ). There are several approaches to IR-based analysis, with the first choice being whether to use the near-IR (750 –2500 nm) or mid-IR (2.5–100 ␮m) spectral range. Near-IR spectroscopy has gained notoriety within the clinical chemistry community through the many efforts to develop a noninvasive blood glucose monitor based on this technology [see e.g., Refs.(9, 10 )], and in that vein it has been shown that glucose concentrations can be recovered from the near-IR spectrum of native serum (3 ).

Fig. 1. Scatterplots comparing IR-predicted urea (top), creatinine (bottom), and protein (middle) concentrations with the concentrations provided by accepted clinical analytical methods. Regression lines (y ⫽ ax ⫹ b, where y is the IR-based analysis, and x is the reference analysis): for creatinine, a ⫽ 0.99, b ⫽ 0.03 mmol/L, r ⫽ 0.98; for urea, a ⫽ 0.99, b ⫽ 7.7 mmol/L, r ⫽ 0.98; for total protein, a ⫽ 0.97, b ⫽ 0.08 g/L, r ⫽ 0.94.

The main reason for the focus on near-IR spectroscopy is that tissue is quite transparent to near-IR light, hence the attraction for in vivo work. However, this is obviously not a factor for in vitro analysis. The mid-IR spectrum offers some potential advantages. Near-IR spectroscopy typically requires a sample volume of at least 0.1– 0.2 mL, whereas a mid-IR assay can be carried out with ⱕ10 ␮L. Although water contributes enormous absorption bands in the mid-IR, these can be eliminated by simply drying the sample to a film and using the spectrum of the dry film as the basis for analysis (6 – 8 ). This film may then be archived for subsequent reanalysis. The present study was conducted to evaluate the sensitivity and accuracy of mid-IR spectroscopy in the determination of urine urea, creatinine, and total protein. The IR-based quantification methods were calibrated by comparison with the results provided by standard clinical chemistry assays. To that end, urea [enzymatic (urease) conductivity], creatinine (Jaffe´ rate), and total protein (benzethonium chloride reaction) concentrations were determined for 200 urine samples. Urea concentrations were 40 – 440 mmol/L, creatinine concentrations were 1.5–18 mmol/L, and total protein was 0.02–20 g/L. Samples were prepared for IR spectroscopy by first adding 0.1 mL of aqueous (4 g/L) potassium thiocyanate solution to 0.5 mL of the urine sample. Duplicate films were prepared by drying 12 ␮L of this mixture onto IR-transparent BaF2 substrates, and mid-IR absorption spectra were acquired at ambient temperature for the dry films (Bio-Rad FTS40A Fourier transform IR spectrometer operating at 4 cm⫺1 resolution, with 512 scans averaged for both the sample and background spectra). An isolated thiocyanate absorption at 2060 cm⫺1 then provided the basis to normalize all spectra to a common effective optical pathlength. Quantification methods were derived by using partial least-squares regression (PLS) to establish relationships between the IR spectra and the reference analyses. A training set of 133 specimens (266 spectra) was used to calibrate quantification methods for each of the three analytes. The test set, comprising the remaining 67 specimens (134 spectra), served to test the validity of the IR-based assays. The accuracy of the PLS quantification models was improved by using spectral subregions rather than the entire 800-5000 cm⫺1 range that was available. The appropriate spectral regions for PLS were determined by first carrying out a series of exploratory trials using limited spectral ranges and fine-tuning those ranges based on the standard errors in the training and test sets. The number of PLS factors in the final model was set at the point where (a) the addition of more factors produced either no improvement or a deterioration in the concentrations predicted for the test set, and (b) the predicted concentrations were equally accurate for the training and test sets. The final quantification models were based on the spectral region 900-1500 cm⫺1 for protein (16 PLS factors), 1400 –1800 cm⫺1 for creatinine (11 factors), and 3100 –3550 cm⫺1 for urea (7 factors). Scatterplots comparing the IR-predicted protein, creatinine, and urea concen-


trations to the reference analyses for this set of test specimens are shown in Fig. 1. The IR-based analytical methods yielded creatinine concentrations with a Sy兩x [the root mean square difference between IR-predicted and reference analyte concentrations for the test set only] of 0.58 mmol/L (r ⫽ 0.98) for creatinine, 14.1 mmol/L (r ⫽ 0.98) for urea, and 0.48 g/L (r ⫽ 0.94) for protein. The distribution of protein concentrations is skewed heavily, with the majority of specimens showing concentrations well below 1 g/L (Fig. 1, middle panel). As a result, the best approach to IR-based protein quantification is to use two models rather than one. A second PLS quantification model was optimized for those samples with concentrations ⬍1 g/L, yielding Sy兩x ⫽ 0.13 g/L. although this still falls short of the performance required for accurate quantification at typical low protein concentrations, the method is sufficiently accurate to serve as a coarse screening test. The ultimate accuracy of the IR-based methods is influenced in part by the accuracy of the reference methods used to calibrate them. This is not a factor for the protein analysis, where the reference method is clearly more accurate than the IR-based method, but it may play a role for both urea and creatinine. This possibility is suggested by the precision of the IR-based assays: SDdup ⫽ 0.18 mmol/L for creatinine, 6.8 mmol/L for urea, 0.14 g/L for protein (including all samples), and 0.05 g/L for protein concentrations ⬍1 g/L.3 At least part of the gap between the precision and accuracy of the urea (Sy兩x ⫽ 14.5; SDdup ⫽ 6.8 mmol/L) and creatinine (Sy兩x ⫽ 0.54; SDdup ⫽ 0.18 mmol/L) assays may be attributable to scatter in the reference methods themselves. The mid-IR quantification methods presented here match or exceed the performance of the near-IR methods presented previously (12 ). Both approaches yield analyses that are accurate enough to serve as a routine method for urine urea and creatinine analyses. Although protein concentrations are too low for accurate quantification using IR spectroscopy, the method may serve as a screen to detect concentrations above ⬃0.5 g/L and to quantify at those concentrations. The practical implementation of this and other clinical IR-based assays requires two key developments. One of these is the discovery of an inexpensive substrate to substitute for the costly BaF2 windows that were used as part of this work. Although these windows can be cleaned and used repeatedly, this is probably impractical in highvolume laboratories. A surprising alternative has emerged recently, as we have shown recently that many analyses can be carried out using ordinary glass as the substrate, despite its limited transparency in the mid-IR region (13, 14 ). The stumbling block that remains in place is a practical one, that being automation of the method. The practical benefits of IR-based methods are being realized in an extraordinary range of analytical applications (15 ), and it would seem to be

3 SDdup ⫽ (⌺d2/2n)1/2, where d is the difference between concentrations determined for duplicate aliquots and n is the number of samples (11 ).

1495

only a matter of time before these methods find their way into the clinical realm. References 1. Hall JW, Pollard A. Near-infrared spectrophotometry: a new dimension in clinical chemistry. Clin Chem 1992;38:1623–31. 2. Hall JW, Pollard A. Near-infrared spectroscopic determination of serum total proteins, albumin, globulins, and urea. Clin Biochem 1993;26:483–90. 3. Hazen KH, Arnold MA, Small GW. Measurement of glucose and other analytes in undiluted human serum with near-infrared transmission spectroscopy. Anal Chim Acta 1998;371:255– 67. 4. Heise HM, Marbach R. Koschinsky T, Gries FA. Multicomponent assay for blood substrates in human plasma by mid-infrared spectroscopy and its evaluation for clinical analysis. Appl Spectrosc 1994;48:85–95. 5. Janatsch G, Kruse-Jarres JD, Marbach R, Heise HM. Multivariate calibration for assays in clinical chemistry using attenuated total reflection infrared spectra of human blood plasma. Anal Chem 1989;61:2016 –23. 6. Shaw RA, Kotowich S, Leroux M, Mantsch HH. Multianalyte serum analysis using mid-infrared spectroscopy. Ann Clin Biochem 1998;35:624 –32. 7. Liu KZ, Dembinski TC, Mantsch HH. Rapid determination of fetal lung maturity from infrared spectra of amniotic fluid. Am J Obstet Gynecol 1998;178:234 – 41. 8. Liu KZ, Shaw RA, Dembinski TC, Reid GJ. Low Ying S, Mantsch HH. A comparison of the accuracy of the infrared spectroscopy and TDx-FLM assays in the estimation of fetal lung maturity. Am J Obstet Gynecol 2000;183:181–7. 9. Heise HM. Non-invasive monitoring of metabolites using near infrared spectroscopy: state of the art. Horm Metab Res 1996;28:527–34. 10. Khalil OS. Spectroscopic and clinical aspects of noninvasive glucose measurements. Clin Chem 1999;45:165–77. 11. Koch DD, Peters T. Selection and evaluation of methods: with an introduction to statistical techniques. In: Burtis CA, Ashwood ER, eds. Tietz fundamentals of clinical chemistry. Philadelphia: WB Saunders, 1996:170. 12. Shaw RA, Kotowich S, Mantsch HH, Leroux M. Quantitation of protein, creatinine, and urea in urine by near-infrared spectroscopy. Clin Biochem 1996;29:11–9. 13. Shaw RA, Mantsch HH. Multianalyte serum assays from mid-IR spectra of dry films on glass slides. Appl Spectrosc 2000;54:885–9. 14. Shaw RA, Eysel HH, Liu KZ, Mantsch HH. Infrared spectroscopic analysis of biomedical specimens using glass substrates. Anal Biochem 1998;259: 181– 6. 15. Davies AMC, Williams P, eds. Near infrared spectroscopy: the future waves. Chichester, UK: NIR Publications, 1996:742pp.

Miniaturization of the Luminescent Oxygen Channeling Immunoassay (LOCITM) for Use in Multiplex Array Formats and Other Biochips, Alan Dafforn,* Hrair Kirakossian, and Kaiqin Lao† (Advanced Diagnostics Group, Dade Behring Inc., PO Box 49013, San Jose, CA 95161-9013; * author for correspondence: fax 408-239-2707, e-mail [email protected]; † present address: PE Biosystems, 850 Lincoln Centre Dr., Foster City, CA 94404) Many of the emerging technologies in clinical chemistry and research require the ability to perform hundreds or thousands of measurements on a single sample such as amplified DNA, typically by contacting the sample with an array of different probes or other reagents. If these array approaches are to be practical, the underlying technology must be simple, robust, inexpensive, and amenable to automation. The Luminescent Oxygen Channeling Immunoassay (LOCITM) is a recently developed homogeneous assay method that should be suitable for arrays because of its simplicity. However, to perform large numbers of measurements on reasonable sample sizes (e.g., 500 different measurements on aliquots of a

1496


50-␮L volume), it must be possible to detect LOCI signals from very small volumes. Miniaturization has also become a central theme in other areas of clinical chemistry (1 ). Accordingly, we have constructed a LOCI microscope

Fig. 1. Calibration curves for three assays determined in 10 nL (⽧) or a large volume (f; 80 ␮L in A and B, 20 ␮L in C). The ratio of specific signal (signal ⫺ background) to background in the absence of analyte [(S ⫺ B)/B] is plotted against the final concentration of analyte in the reaction (or dilution for Chlamydia amplicon). In each case, the dashed line represents the signal at 3 SD above background. (A), assay in which two beads are cross-linked by an oligonucleotide; (B), immunoassay for TSH; (C), assay for amplified Chlamydia DNA.

and used it to demonstrate sensitive detection of analytes in small volumes for three types of assays of potential interest in arrays: detection of a single-stranded DNA fragment, detection of a double-stranded DNA amplicon, and an immunoassay for a protein. LOCI is a sensitive (femtomolar) detection method that uses chemiluminescence to quantify latex agglutination (2 ). This technique utilizes one latex particle dyed with a photosensitizer and a second dyed with a chemiluminescent dye, both having binding ligands on their surfaces. Particle suspensions are mixed with the sample, and cross-linking by any analyte present leads to formation of bead pairs or higher aggregates. When the suspension is then illuminated, singlet oxygen is generated by the sensitizer particle, migrates to the chemiluminescent particle, and generates light. Nonspecific signals are low because singlet oxygen decays before it can reach unpaired particles. A small-volume LOCI reader was constructed by modifying a fluorescence microscope to allow sample illumination with a 678 nm laser and monitoring of chemiluminescence with a photomultiplier tube. Illumination was accomplished either through the objective, using a beam splitter, or from below the stage, using a shutter to protect the photomultiplier; similar results were obtained in either mode. Samples were read by adding a 10-␮L aliquot to a hemocytometer cell and imaging a single field. The volume imaged was defined by the depth of the cell (100 ␮m) and the diameter of the field (365 ␮m) as ⬃10 nL. Three different assays were compared by performing incubations as described previously (2 ), and then removing aliquots of the final mixture and reading either using the microscope or using our regular LOCI readers. These prototype readers illuminate and read 20- or 80-␮L aliquots through standard optics. Reagents and procedures for two of the analytes, an oligonucleotide linker and thyroid-stimulating hormone (TSH), were essentially as described previously (2 ). Briefly, in one assay an oligonucleotide linker included the sequences 5⬘-(TACT)5 and T20 separated by a short spacer. This linker forms bead pairs between sensitizers with conjugated A24 and chemiluminescent particles with conjugated 5⬘-(AGTA)6. In the other assay, TSH binds first to a chemiluminescent bead conjugated with one monoclonal anti-TSH antibody and to a second biotinylated monoclonal anti-TSH antibody. Streptavidin beads are then added to form bead pairs if TSH is present. Assays were read by three cycles (five for TSH) of illuminating for 1 s at 678 nm followed by 1 s of light collection. Detection of double-stranded DNA was demonstrated using an amplicon from Chlamydia trachomatis. A 518-bp sequence from cryptic plasmid pLGV440 was amplified by conventional PCR using the primers 5⬘-GGA CAA ATC GTA TCT CGG GTT ATT-3⬘ and 5⬘-GGA AAC CAA CTC TAC GCT GTT-3⬘. The final concentration of the amplified DNA was estimated as ⬃40 nmol/L by comparison to an earlier LOCI calibration curve. Various dilutions of the amplicon were mixed with 2 ␮g of each


LOCI bead (same as for single-stranded DNA detection above). Solutions also contained each of the following probes at a concentration of 50 nmol/L in a total volume of 40 ␮L of PCR buffer (70 mmol/L KCl, 10 mmol/L Tris-HCl, 4 mmol/L MgCl2, 0.2 g/L acetylated bovine serum albumin, pH 8.2): 5⬘-CTC ACA GTC AGA AAT TGG AGT ACT TAC TTA CTT ACT TAC T-X 5⬘-TTT TTT TTT TTT TTT TTT TTA GAC TTT TTC TAT TCG CAG CGC-X Target-specific binding regions are underlined. X represents OCH2CH((CH2)4NH2)CH2OH, a group introduced to block the 3⬘ end against possible enzymatic extension. Solutions were covered with 20 ␮L of mineral oil in 200-␮L MicroAmpTM tubes (PE BioSystems), and then heated to 95 °C for 2 min to denature the amplicon, cooled to 50 °C for 15 min to allow probes to bind, then cooled to 37 °C for 70 min to allow formation of the complex with beads. Aliquots were then read as above. The results for all three assays are summarized in Fig. 1. The ratio of specific signal (signal ⫺ background) to background is presented as a function of analyte concentration in each case to allow comparisons over many orders of magnitude. In general, the assays lost only 14- to 64-fold in signal/background over a 2000- to 8000-fold decrease in volume. As expected, ⬃1000-fold decreases in signal were partially compensated by decreased background. (Raw background counts for large volume and 10 nL are as follows: DNA Linker, 6506 and 60; Chlamydia DNA, 4777 and 68; and TSH, 14 934 and 183.) As they should be, the curve shapes in Fig. 1 are in general parallel and log-linear except when very little signal is present. The decreasing signal seen at high concentrations of Chlamydia amplicon is frequently observed when LOCI is used to detect DNA amplification in a closed tube. It is believed to result from saturation of bead surfaces, but it can be avoided by changes in protocol (3 ). Detection limits for each assay were defined as the concentrations at which observed signals exceeded background by 3 SD and were estimated from linear plots of low concentration points. The change in the lower limit of detection as a function of volume was larger than that in signal/background because the statistical CV of the background increased at small volumes. The limits for large volumes and for 10 nL were as follows: Chlamydia amplicon: 1:1 000 000 and 1:12 000 dilutions (⬃3 pmol/L) DNA Linker: 6 fmol/L and 0.8 pmol/L TSH: 7 fmol/L and 2 pmol/L Because most array applications are likely to involve detection and quantification of undiluted amplicons or of expressed proteins, LOCI has ample limits of detection for most applications even in very small volumes. The imprecision of the specific signal (signal ⫺ background) was determined in the TSH assay as 82 ⫾ 10

1497

counts (n ⫽ 5; CV ⫽ 12%) at 3.57 pmol/L TSH and 731 ⫾ 68 counts (n ⫽ 5; CV ⫽ 9%) at 35.7 pmol/L. The imprecision was also estimated using a single bead dyed with both sensitizer and chemiluminescent dye according to the basic procedures described previously (2 ). Ten replicate aliquots of a bead suspension gave 2450 ⫾ 328 counts (CV ⫽ 13%). The most likely source of the observed variability is manual positioning of each well of the slide under the microscope objective in this prototype instrument. Realization of practical LOCI arrays will also require a sample cassette. This could be as simple as a matrix of closed wells, each containing LOCI reagents for a specific measurement and interconnected by channels to distribute the sample. Fortunately, the dimensions required are easily obtainable by inexpensive molding techniques (4 ). For example, a well with a 50-nL volume could be 125 ␮m deep and 714 ␮m in diameter. If 0.5 cm of channel on the average was required to connect each well and the channel had a cross-sectional area of 104 ␮m2, then another 50 nL per well would be required to fill the channel. Thus, a 50-␮L sample volume would be sufficient to fill 500 wells and the necessary connecting channels. In summary, LOCI offers several advantages for signal detection from arrays and other miniaturized devices: The assay retains ample sensitivity for analytes of likely interest in such devices. An oligonucleotide could be detected at ⬃1 pmol/L (6000 molecules), the protein TSH could be detected at 2 pmol/L, and a DNA amplicon could be detected even at a 1:10 000 dilution. In addition, arrays large enough for clinical diagnostic purposes should be feasible (500 or more measurements/sample). Homogeneous assay arrays should also be much simpler to manufacture than many types of arrays because no surface chemistry must be performed on a chip. The absence of surface chemistry or absorption should also give greater reproducibility compared with spotting technologies and simplify quality control. The use of generic reagents also simplifies preparation of large arrays. Finally, homogeneous assays offer relatively fast kinetics and simplicity of protocol. We thank Neal DeChene, John Pease, Sharat Singh, and Raj Singh for supplying many of the reagents used in this work and Sam Rose for many helpful insights.

References 1. Kricka LJ. Miniaturization of analytical systems. Clin Chem 1998;44:2008 – 14. 2. Ullman EF, Kirakossian H, Switchenko AC, Ishkanian J, Ericson M, Wartchow C, et al. Luminescent oxygen channeling immunoassay (LOCITM): sensitive, broadly applicable homogeneous immunoassay method. Clin Chem 1996; 42:1518 –26. 3. Patel R, Pollner R, de Keczer S, Pease J, Pirio M, DeChene N, et al. Quantification of DNA by use of the luminescent oxygen channeling assay. Clin Chem 2000;46:1471–7. 4. McCormick RM, Nelson RJ, Alonso-Amigo MG, Benvegnu DJ, Hooper HH. Microchannel electrophoretic separations of DNA in injection-molded plastic substrates. Anal Chem 1997;69:2626 –30.

1498


Application of the Luminex LabMAP in Rapid Screening for Mutations in the Cystic Fibrosis Transmembrane Conductance Regulator Gene: A Pilot Study, Sherry A. Dunbar* and James W. Jacobson (Luminex Corporation, 12212 Technology Blvd., Austin, TX 78727; * author for correspondence: fax 512-258-4173, e-mail sdunbar@ luminexcorp.com) Cystic fibrosis (CF) is an autosomal recessive disorder that occurs in ⬃1 in 2500 Caucasians (1–3 ). CF can result from combinations of ⬎750 known mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, and in certain populations, as many as 1 person in 25 is a carrier (1–3 ). To date, screening methods have either been narrow, requiring multiple methodologies to be brought to bear for comprehensive coverage, or are cost prohibitive (1, 4 ). This study describes the Luminex LabMAPTM system and its potential for simultaneous, rapid, sensitive, and specific screening for mutations in the CFTR gene. The LabMAP system incorporates polystyrene microspheres that are internally dyed with two spectrally distinct fluorochromes. Using precise ratios of these fluorochromes, an array is created consisting of 100 different microsphere sets with specific spectral addresses. Each microsphere set can possess a different reactant on its surface. Because microsphere sets can be distinguished by their spectral addresses, they can be combined, allowing up to 100 different analytes to be measured simultaneously in a single reaction vessel. A third fluorochrome coupled to a reporter molecule quantifies the biomolecular interaction that has occurred at the microsphere surface. Microspheres are interrogated individually in a rapidly flowing fluid stream as they pass by two separate lasers in the Luminex100 analyzer. High-speed digital signal processing classifies the microsphere based on its spectral address and quantifies the reaction on the surface in a few seconds per sample. To evaluate the utility of this platform for CF screening, we constructed a multiplexed assay to detect wild-type and mutant DNA sequences in the CFTR gene. The selected mutations consisted of both single-base changes and small deletions, and are the five most common CF mutations found in North America: ⌬F508 (66.1%), W1282X (2.35%), G542X (2.24%), 621⫹1G3 T (1.48%), and N1303K (1.25%) (1 ). A panel of patient DNA samples characterized for mutations in the CFTR gene was obtained from Coriell Cell Repositories for use in this study. Presumed wildtype genomic DNA samples were obtained from Sigma Chemical Company. Regions of exon 10 (upstream primer 5⬘ to 3⬘, TCTGTTCTCAGTTTTCCTGG; downstream primer 5⬘ to 3⬘, TTGGCATGCTTTGATGACGC), exon 11 (TAGGACATCTCCAAGTTTGC and CAATAATTAGTTATTCACCTTGC), exon 20 (GAGACTACTGAACACTGAAG and TTCTGGCTAAGTCCTTTTGC), exon 21 (TGCTATAGAAAGTATTTATTTTTTCTGG and AGCCTTACCTCATCTGCAAC), and intron 4 (CTTCATCACATTGGAATGCAG and ACTTGTACCAGCTCACTACC) containing the mutation sites were amplified by

multiplexed PCR to yield amplicons 106 –147 bp in length. Each of the upstream primers was labeled at the 5⬘ terminus with biotin. Targets were amplified in 50-␮L reactions containing 1⫻ PCR buffer (Qiagen), 200 ␮mol/L each dNTP, 0.2 ␮mol/L each primer (Sigma Genosys), 1.5 mmol/L MgCl2, 2.5 U of HotStarTaq DNA polymerase (Qiagen), and 100 ng of template DNA. PCR reactions were incubated at 95 °C for 15 min to activate the enzyme and then cycled for 40 cycles with denaturation at 94 °C for 30 s, annealing at 50 °C for 1 min, and extension at 72 °C for 1 min. Final extension was done at 72 °C for 7 min, and reactions were soaked at 4 °C. Capture oligonucleotide probes specific for each wildtype and mutant sequence [⌬F508 (wild-type 5⬘ to 3⬘, AACACCAAAGATGATATTTT; mutant 5⬘ to 3⬘, GAAACACCAATGATATTTTC), W1282X (GGCTTTCCTCCACTGTTGCA and GGCTTTCCTTCACTGTTGCA), G542X (CACCTTCTCCAAGAACTATA and CACCTTCTCAAAGAACTATA), 621⫹1G3 T (GAAGTATTACCTTCTTATAA and GAAGTATTAACTTCTTATAA), and N1303K (GGGATCCAAGTTTTTTCTAA and GGGATCCAACTTTTTTCTAA)] were synthesized with a 5⬘ amine Uni-Link modification (Oligos Etc). Probes were 20 nucleotides in length, complementary in sequence to the biotinylated strand of the amplicon, and designed with the specific wild-type or mutant nucleotide(s) centered within the probe sequence. Probes were coupled to carboxylated microspheres using a previously described carbodiimide coupling method (5 ). Briefly, for each probe and microsphere set combination, 5 ⫻ 106 carboxylated microspheres were suspended in 50 ␮L of 0.1 mol/L MES, pH 4.5. One nmol of amine-substituted probe oligonucleotide was added, followed by the addition of 25 ␮g of N-(3dimethylaminopropyl)-N⬘-ethylcarbodiimide (EDC) and incubation in the dark for 30 min. The addition of EDC and incubation were repeated, and the microspheres were washed once with 0.2 mL/L Tween 20 and once with 1 g/L sodium dodecyl sulfate. Coupled microspheres were stored in MES containing 1 mmol/L EDTA, pH 8.0, at 2– 8 °C in the dark. After PCR amplification, 5 ␮L of each reaction was denatured at 95 °C for 10 min and added to hybridization solution [3 mol/L tetramethylammonium chloride, 50 mmol/L Tris-HCl (pH 8.0), 4 mmol/L EDTA (pH 8.0), 1 g/L Sarkosyl] containing a mixture of 5000 of each probe-coupled microsphere set in a 50-␮L total reaction volume. Reactions were hybridized at 55 °C for 10 min, pelleted by microcentrifugation, and resuspended in 50 ␮L of hybridization solution. Hybridized amplicons were labeled with 120 ng of streptavidin-R-phycoerythrin (Molecular Probes) at 55 °C for 5 min (62 ␮L total reaction volume). Reactions were then analyzed on the Luminex100. Assay results were obtained ⬃30 min after amplification. DNA samples were amplified twice, and two aliquots of each reaction were analyzed in four independent experiments. Typical results are presented in Fig. 1. The bar graph (Fig. 1, top) indicates the net median fluorescence for each probe-coupled microsphere set for a wild-


1499

Fig. 1. Identification of CFTR mutations by the Luminex LabMAP system. The bar graph (top) demonstrates the net median fluorescence for wild-type (normal) and mutant probes for each of the five targets in a normal DNA sample. Net median fluorescence values obtained for all 15 patient DNA samples are listed at the bottom. Shaded boxes indicate positive results.

1500


type DNA sample. The net median fluorescence values obtained for all 15 patient DNA samples are also shown (Fig. 1, bottom). The Luminex LabMAP simultaneously and correctly identified the genotype of the five selected CFTR mutations in all patient DNA samples tested. Net fluorescence ranged from 121 to 737 above background for all positive alleles. The signal-to-noise ratio for each wild-type and corresponding mutant probe pair was ⱖ7.2 in all cases. This study demonstrates the potential of the Luminex LabMAP system for simultaneous, rapid, and specific molecular genetic analyses of multiple nucleic acid targets. Because the LabMAP system is capable of screening a single sample for 100 different analytes simultaneously, probes for up to 50 CFTR mutations (and corresponding wild-type sequences) could be combined in a single reaction well. Screening for additional mutations could be performed in multiple wells, and additional probe-coupled microsphere sets could be included as internal controls for sample integrity and yield. In summary, this pilot study demonstrates the potential of the Luminex LabMAP system as an analysis platform suitable for routine multiplexed screening for diverse CF mutations. Analysis with the system is rapid, requires relatively few sample manipulations, and produces adequate resolution to reliably genotype five of the most commonly occurring CF mutations.

Using AmplifluorTM technology, we have developed a rapid and extremely sensitive assay for cytokine mRNA quantification. The high signal-to-noise ratio of the assay that is achieved by the unique structure of Amplifluor primers allows for closed-tube fluorescence detection and quantification during PCR amplification (real time) or at endpoint. Elimination of laborious post-PCR sample processing enables high-throughput analysis and greatly reduces the risk of carryover contamination. The Amplifluor primer system is a molecular switch for detecting DNA amplification by energy transfer between fluorophore and quencher (2, 3 ). The OFF-to-ON transition occurs when the conformation of the Amplifluor primer changes from “closed” intramolecular stem-loop structure to an “open” extended structure. This structural change is achieved when the Amplifluor primers are incorporated into a double-stranded DNA molecule by primer-medi-

References 1. Schwarz M, Malone G. Methods for screening in cystic fibrosis. In: Elles R, ed. Methods in molecular medicine: molecular diagnosis of genetic diseases. Totowa, NJ: Humana Press, 1996:99 –119. 2. Centers for Disease Control and Prevention. Newborn screening for cystic fibrosis: a paradigm for public health genetics policy development. Morbid Mortal Wkly Rep 1997;46:24pp. 3. Grody WW. Cystic fibrosis: molecular diagnosis, population screening, and public policy. Arch Pathol Lab Med 1999;123:1041– 6. 4. Rowley PT, Loader S, Kaplan RM. Prenatal screening for cystic fibrosis carriers: an economic evaluation. Am J Hum Genet 1998;63:1160 –74. 5. Fulton R, McDade R, Smith P, Kienker L, Kettman J. Advanced multiplexed analysis with the FlowMetrix system. Clin Chem 1997;43:1749 –56.

Rapid and Sensitive Closed-Tube Quantification of Human Interferon-␥ mRNA by Reverse Transcription-PCR Utilizing Energy-Transfer Labeled Primers, Hiroshi Uehara,* Glenn Nardone, Sandra Randall, Yuri Khripin, and Dan St. Louis (Intergen Discovery Products, 202 Perry Pkwy., Gaithersburg, MD 20877; * author for correspondence: fax 301-963-5017, e-mail [email protected]) Quantification of cytokine gene expression is essential for understanding the biological roles of cytokines in immune responses. Quantitative reverse transcription-PCR has become popular for the sensitive detection and quantification of mRNA (1 ). However, the necessity of postPCR sample processing with the potential for carryover contamination of the amplicon makes it a difficult methodology for routine clinical testing.

Fig. 1. Calibration curves for human IFN-␥ control target obtained by endpoint analysis (A) and the Prism 7700 Sequence Detection System (B). (A), each curve corresponds to the calibration curve obtained with different PCR cycles (31, 33, 35, and 37, respectively). The calibration curves show the typical amplification dynamics of the Amplifluor system with the number of PCR cycles performed. (B), thermocycling conditions identical to those described for endpoint analysis were used, and the fluorescence was monitored at the annealing step.


ated DNA amplification such as PCR. The unique features of the Amplifluor system, direct incorporation into amplification products and low fluorescence background observed with unincorporated primers, allow closed-tube quantification by endpoint or real-time analysis. Here we present results obtained for human interferon ␥ (IFN-␥) by endpoint analysis (Wallac plate reader) and by real-time analysis (ABI Prism 7700). For endpoint analysis, the 42-nt-long human IFN-␥specific Amplifluor primer consisted of a 20-nt-long 3⬘ end sequence (5⬘-TTGCTTTGCGTTGGACATTC-3⬘) complementary to the cDNA sequence in the fourth exon of IFN-␥ and a 5⬘ end 22-nt-long stem-loop structure. The 22nt-long 5⬘ end sequence of the Amplifluor primer, which has no homology to the IFN-␥ sequence, was designed to form a stable intramolecular hairpin structure. The 5⬘ end of the Amplifluor primer is labeled with fluorescein (donor) in the last step of chemical synthesis. The DABSYL [4-(4⬘-dimethylamino-phenylazo)benzene sulfonic acid; quencher], a nonfluorescent chromophore, was linked to a T residue containing a C5-C6 amino group that is complementary to the 5⬘ end A linked to the fluorescein. The sequence of the 20-nt-long reverse primer (nonlabeled oligomer; 5⬘-GACCAGAGCATCCAAAAGAG-3⬘) was identical to the cDNA sequence in the third exon of IFN-␥. Amplification of the IFN-␥ cDNA with these primers generated a 174-bp cDNA fragment that spans across the 2424-nt-long third intron. We used serially diluted linearized IFN-␥ control template to generate a calibration curve by endpoint analysis (Fig. 1A). PCR amplification was performed in a reaction mixture (final volume, 25 ␮L) containing 400 nmol/L primers, 200 ␮mol/L dNTP mixture, 1 U of rTaq polymerase (Intergen Company), 10 mmol/L Tris-HCl, pH 8.5, 1.8 mmol/L MgCl2, and 50 mmol/L KCl. The amplification was performed using a thermocycler (PerkinElmer GeneAmp PCR System 9700) with the following cycling condition: denaturation step at 94 °C for 5 min, followed by 31–37 cycles of 94 °C for 15 s, 55 °C for 40 s, and 72 °C for 1 min. After the PCR was completed, the relative fluorescence in reaction vessels was measured using a Wallac-Perkin-Elmer Victor 1420 fluorescence plate reader (excitation/emission filters 485 ⫾ 7.5 nm /535 ⫾ 15 nm). The calibration curve showed a detection limit of 100 copies of target IFN-␥ with a log-linear dynamic range that spanned three orders of magnitude (102–105) after 35 PCR cycles. It also indicated that the dynamic range and sensitivity are dependent on the number of amplification cycles. A linear dynamic range between 103 and 106 copies was observed after 31 PCR cycles, whereas increased sensitivity but a narrower dynamic range (102–104) were obtained by increasing the amplification cycles to 37. The direct correlation between the relative fluorescence in reaction vessels and the amount of amplicon was confirmed by an agarose gel electrophoresis of the amplification products (data not shown). To demonstrate the versatility of the Amplifluor system in closed-tube detection of PCR amplicons, we subjected

1501

serially diluted IFN-␥ target to real-time analysis with the ABI Prism 7700 Sequence Detection System. As shown in Fig. 1B, a linear calibration curve with a wide dynamic range (101–107) was obtained. Similar results were also obtained using LightCycler (Roche Molecular Biochemicals) under slightly different conditions (data not shown). In summary, the methodology described here uses Amplifluor technology to provide versatile and sensitive detection of cytokine mRNAs. Direct incorporation into the amplicon and the high signal-to-noise ratio of the Amplifluor primers enable quantification of the amplicon by multiple instrument platforms in both endpoint and real-time assays, a feature difficult to achieve with genespecific hybridization probes. The Amplifluor methodology is simple and inexpensive and is well suited for high-throughput PCR assays.

We thank Ray A. Martin for expert technical assistance. References 1. Siebert P, ed. The PCR technique: RT-PCR. Biotechniques update series. Natick, MA: Eaton Publishing, 1998:300pp. 2. Nazarenko IA, Bhatnagar SK, Hohman RJ. A closed tube format for amplification and detection of DNA based on energy transfer. Nucleic Acids Res 1997;25:2516 –21. 3. Uehara H, Nardone G, Nazarenko IA, Hohman RJ. Detection of telomerase activity utilizing energy transfer primers: comparison with gel- and ELISAbased detection. Biotechniques 1999;26:552– 8.

Thin Film Biosensor for Rapid Detection of mecA from Methicillin-resistant Staphylococcus aureus, Robert Jenison, Ayla Haeberli, Shao Yang, Barry Polisky, and Rachel Ostroff * (BioStar, Inc., 6655 Lookout Rd., Boulder, CO 80301; * author for correspondence: fax 303-581-6405, email [email protected]) The reported incidence of methicillin-resistant Staphylococcus aureus (MRSA) isolates in hospitals has increased from 2% in 1974 to 50% in 1997 (1 ). In addition, MRSA is emerging as a community-acquired pathogen. Resistance is most often mediated by the mecA gene, which encodes an altered penicillin-binding protein (PBP-2A) with low affinity for ␤-lactam antibiotics. Rapid identification of the mecA gene is important for implementation of appropriate antibiotic therapy. MRSA is identified by either culture or molecular methods. Routine culture methods require two sequential steps, one to isolate S. aureus and the second to determine antibiotic susceptibility (2 ). Recently, molecular methods, including PCR, branched DNA, and cycling probe assays, have been described that identify mecA sequences from individual S. aureus colonies (3–7 ). These methods are sensitive and specific, but they all require instrumentation to interpret the results. In this report, we describe a thin film biosensor for qualitative visual detection of mecA either directly from a single S. aureus colony or from a PCR amplification

1502


reaction. The technology allows direct visual detection of the interaction of target DNA sequences with complementary oligonucleotide probes immobilized to an optically coated silicon chip (8 ). The key assay steps for thin film formation are: (a) simultaneous hybridization of the target sequence to covalently attached surface capture probe and solution-phase biotinylated detector probes; (b) binding of anti-biotin antibody enzyme conjugate to the annealed detector probes; and (c) mass deposition through enzyme-catalyzed precipitation (Fig. 1A). The thin film alters the interference pattern of light on the surface, producing a perceived color change. Composition of the optical layers is designed such that small increases in thickness produce a color change from gold to purple, producing contrast in the range where the human eye is

most sensitive. Thickness changes as small as 10Å can be detected visually. The thin film biosensor may be interpreted visually for a qualitative result. Alternatively, quantification may be obtained by either of two instrumented methods, chargecoupled device (CCD) imaging or ellipsometry (9, 10 ). CCD imaging software measures color intensity, and ellipsometry measures thin film thickness. Each method provides a measurement that is proportional to the amount of target bound to the biosensor surface. The nucleic acid detection surface described in these studies was a multilayer optical surface composed of a silicon wafer layered with 515Å of silicon nitride applied by plasma vapor deposition and 130Å of a hydrophobic polymer (polydimethylsiloxane; United Chemical Tech-

Fig. 1. Schematic of thin film biosensor surface and nucleic acid hybridization assay (A), assay for mecA from a single MRSA or MSSA colony lysate (B), and color difference values and images of thin film biosensor results for dilutions of the mecA PCR reaction (C). (B), graphical data are color difference values calculated by CCD imaging and comparison of color intensity within the reaction zone and the surrounding unreacted area. Below the graph are CCD images of representative biosensor chips. Data are the mean ⫾ 1 SD for 10 determinations. (C), below the biosensor chips is an ethidium bromide-stained 2% agarose gel with the equivalent amount of DNA loaded per lane as was tested with the biosensor. ppt, precipitate.


nologies) applied by spin coating. A final layer of polyphenylalanine-polylysine (Sigma) was applied by passive adsorption to provide functional amines. Covalent attachment of the mecA capture probe (5⬘-GTCATTTCTACTTCACCATTACCAAC-3⬘) was accomplished by reacting the homobifunctional cross-linker disuccidimidyl suberate (Pierce) with the 3⬘ amine of the capture probe and amines on the biosensor surface, producing a stable amide linkage. Hybridization of the target sequence to the biosensor surface and to biotinylated detector probes is followed by reaction with an anti-biotin antibody conjugated to horseradish peroxidase (HRP) and precipitating substrate. This precipitation event deposits additional mass onto the surface, triggering a color change from gold to purple. This thin film deposition effectively transduces the hybridization reaction into a visible result. The mecA biosensor determined the correct genotype of methicillin-sensitive S. aureus (MSSA) and MRSA strains from a single colony in a simple 3-h procedure (Fig. 1B). MRSA (ATCC strain 33592) and MSSA (ATCC strain 11632) were incubated at 37 °C for 24 h on a trypticase soy-blood agar plate (Remel). A single colony was suspended in lysis buffer (10 mg/L lysostaphin (Sigma), 75 mmol/L NaCl, 25 mmol/L EDTA, 20 mmol/L Tris, pH 7.5) and incubated at room temperature for 20 min. Samples were then denatured at 95 °C for 10 min in the presence of three biotinylated detector probes (1.33 ␮mol/L each): 422d: 2aminoACACCATTTTATATTGAGCATCTACT; 581d: 2aminoACACCATTTTACCACGTTCTGATTTT; 603d: 2aminoACACCATTCCACATTGTTTCGGTCTAA. The sample was diluted with an equal volume of hybridization buffer [10⫻ standard saline citrate (SSC), 2 g/L sodium dodecyl sulfate, 10 g/L BlockAidTM (BioStar reagent)], incubated for 10 min at 95 °C, and then immediately added to the biosensor surface. Simultaneous hybridization of the S. aureus genomic DNA to the biosensor surface and the biotinylated detector probes was carried out for 2 h at 53 °C, followed by 10 min at 23 °C. The surfaces were washed with 0.1⫻ SSC containing 1 g/L sodium dodecyl sulfate, followed by 0.1⫻ SSC. Anti-biotin antibody conjugated to HRP (1 mg/L) was incubated on the surface for 10 min at room temperature; the surface was washed, and a precipitating 3,3⬘,5,5⬘-tetramethylbenzidine (TMB) substrate (BioFX) was added for 15 min. The mecA gene was specifically detected from the MRSA strain, illustrating the utility of the mecA biosensor for direct detection of mecA sequences from a single S. aureus colony without purification or amplification of the genomic DNA (Fig. 1B). A second rapid biosensor format was developed to analyze the products of mecA PCR amplification. In this 10-min assay, the mecA hybridization target was a 617-bp PCR amplicon rather than total genomic DNA. Because the PCR amplicon target was less complex and more abundant than chromosomal DNA, hybridization times were shortened from 2 h to 3 min. In addition, the

1503

increased copy number of the specific targeted sequences allowed for reduction in timing of all assay steps. PCR amplification was performed with purified MRSA genomic DNA, (forward primer, 5⬘-TAATAGTTGTAGTTGTCGGGTTTG-3⬘; reverse primer, 5⬘-GGTTTTAAAGTGGAACGAAGGTAT-3⬘). PCR conditions were as follows: 95 °C for 4 min, followed by 25 cycles of 95 °C for 45 s, 53 °C for 45 s, 72 °C for 60 s, which amplified a 617-bp fragment from the mecA gene. The PCR reaction mixture was diluted in hybridization buffer, mixed with 1 ␮mol/L each of the three biotinylated detector probes, and denatured at 95 °C for 3 min. The sample was incubated on the biosensor surface at 50 °C for 3 min, washed, and reacted with the anti-biotin antibody HRP conjugate for 2 min at room temperature. The precipitating TMB substrate was then added for 2 min at room temperature. The assay was complete in 10 min. We compared the relative sensitivity of the rapid mecA biosensor PCR amplicon detection protocol with traditional agarose gel electrophoresis and ethidium bromide staining (Fig. 1C). The mecA biosensor was 10-fold more sensitive than the electrophoretic method with results available in 10 min. The lower limit of detection for the mecA biosensor corresponded to ⬃30 –50 fmol of amplified target. Because a positive result with the biosensor requires specific nucleic acid hybridization, amplification of the correct sequence was confirmed. The thin film biosensor provides rapid, noninstrumented detection of nucleic acid target sequences. Detection of mecA from a single S. aureus colony eliminates the need for confirmatory culture methods and reduces the time for MRSA determination from 24 – 48 h to ⬍3 h. Sequence-specific detection of mecA PCR amplicons was complete in 10 min with sensitivity exceeding electrophoretic analysis. The thin film biosensor is a rapid, sensitive alternative to traditional methods for PCR amplicon detection such as gel electrophoresis or microwell hybridization assays. Because the biosensor sensitivity provides for detection of relatively few surface capture hybridization reactions, targets as complex as bacterial chromosomal DNA can be analyzed without purification. The assays may be configured for multitarget detection simply by attaching multiple capture probes to the silicon surface in discrete locations. The thin film biosensor can be formatted for single-gene or multigene detection to provide information for multiple antibiotic resistance mutations with one biosensor surface.

We thank Chris High for preparation of the figures. References 1. CDC. Four pediatric deaths from community-acquired methicillin resistant Staphylococcus aureus—Minnesota and North Dakota, 1997–1999. Morbid Mortal Wkly Rep 1999;48:707–10. 2. Murray P, Baron E, Pfaller M, Tenover F, Yolken R, eds. Manual of clinical microbiology, 7th ed. Washington: ASM Press, 1999:1555–92. 3. Murakami K, Minamide W, Wada K, Nakamura E, Teraoka H, Watanabe S. Identification of methicillin-resistant strains of staphylococci by polymerase chain reaction. J Clin Microbiol 1991;29:2240 – 4. 4. Geha D, Uhl J, Gustaferro C, Persing D. Multiplex PCR for identification of

1504

5.

6.

7.

8.

9.

10.


methicillin-resistant staphylococci in the clinical laboratory. J Clin Microbiol 1994;32:1768 –72. Salisbury S, Sabatini L, Spiegel C. Identification of methicillin-resistant staphylococci by multiplex polymerase chain reaction assay. Am J Clin Pathol 1995;107:368 –73. Kolbert C, Arruda J, Varga-Delmore P, Zheng X, Lewis M, Kolberg J, et al. Branched-DNA assay for detection of the mecA gene in oxacillin-resistant and oxacillin-sensitive staphylococci. J Clin Microbiol 1998;36:2640 – 4. Bekkaoui F, McNevin J, Leung C, Peterson G, Patel A, Bhat R, et al. Rapid detection of the mecA gene in methicillin resistant staphylococci using a colorimetric cycling probe technology. Diagn Microbiol Infect Dis 1999;34: 83–90. Ostroff R, Hopkins D, Haeberli A, Baouchi W, Polisky B. Thin film biosensor for rapid visual detection of nucleic acid targets. Clin Chem 1999;45:1659 – 64. Ostroff R, Maul D, Bogart G, Yang S, Christian J, Hopkins D, et al. Fixed polarizer ellipsometry for simple and sensitive detection of thin films generated by specific molecular interactions: applications in immunoassays and DNA sequence detection. Clin Chem 1998;44:2031–5. Trotter B, Moddel G, Ostroff R, Bogart G. Fixed-polarizer ellipsometry: a simple technique to measure the thickness of very thin films. Opt Eng 1999;38:902–7.

Endotoxin Activity in Whole Blood by Neutrophil Chemiluminescence—A Novel Analytical Paradigm, Alexander D. Romaschin* and Paul M. Walker (Department of Lab Medicine and Pathobiology and Department of Surgery, Toronto General Hospital, University Health Network, University of Toronto, Toronto, Ontario M5G 2C4, Canada; * address correspondence to this author at: Dept. of Clinical Biochemistry, ES-3-404, Toronto General Hospital, 200 Elizabeth St., Toronto, Ontario M5G 2C4, Canada) Endotoxin, or lipopolysaccharide (LPS), is the major gram-negative bacterial cell wall toxin that triggers septic shock (1 ). Gram-negative endotoxin is shed from the membrane of rapidly proliferating bacteria, and enhanced release into the blood is associated with antibiotic use (2 ). Endotoxin is also translocated through the gut after periods of hypotension often associated with cardiopulmonary bypass or hypovolemic shock (3 ). LPS has been demonstrated to enter the systemic circulation from the lung in experimental animal studies and human clinical investigations. In the blood, LPS binds to a carrier protein, which acts as a chaperone to carry this molecule to CD14 receptors on immunocompetent cells to trigger pro-inflammatory cytokine synthesis (tumor necrosis factor). To date, the detection of LPS in blood or plasma has

been largely dependent on variations of the limulus amebocyte lysate (LAL) assay, which utilizes the clotting enzyme cascade extracted from the primitive “white cell like” amebocyte cells of horseshoe crabs to detect LPS. The effectiveness of this assay in human blood and blood plasma has been controversial and problematic because of numerous interfering reactions and variations in betweenlot and between-manufacturer reagent performance. We have developed a rapid, homogeneous assay for the detection of endotoxin activity (EA) in whole blood based on in vitro neutrophil activation (4 ). This novel type of assay uses the priming effects of complement opsonized immune complexes on the respiratory burst activity of neutrophils as an analytical platform. Hypochlorous acid generated by the concerted activity of membrane-bound NADPH oxidase and azurophil granule myeloperoxidase of the neutrophil produces luminol chemiluminescence. Although immune complexes formed by the interaction of antibody with antigen do not directly stimulate neutrophil respiratory bursts, they prime neutrophil oxidative machinery to higher activity, which is subsequently elicited by challenging the cells with yeast cell wall zymosan. The EA assay consists of three tubes containing lyophilized reagents that reconstituted with buffer (1 mL) containing luminol and zymosan. Assay reactions are initiated by the addition of 0.04 mL of whole blood to each tube, and light emission is monitored and integrated over 20 min in a temperature controlled (37 °C) photon counting luminometer (Berthold 953). One assay tube (blank) reflects baseline neutrophil activation in the absence of exogenous immune complexes. A second tube (test) contains a specific anti-LPS IgM that stimulates neutrophil activity in proportion to the concentration of LPS in the blood. The third tube (max) contains specific anti-LPS IgM and an excess of LPS so that the chemiluminescence reflects the maximum response of the individual patient sample. The differences in neutrophil activation and cell count between individual samples are normalized by subtracting the light integral of the blank from the test and max tubes and expressing the EA as the ratio of the test (minus blank) to the max (minus blank). EA is a hyperbolic function of the LPS concentration over the range of 0 – 800 ng/L. The sensitivity of the assay, primarily dependent on antibody concentration, was adjusted to yield 50% of maximal signal in the steepest portion of the dose–response curve at borderline pathological endotoxin concentrations (50 ng/L). Detection of

Table 1. Summary of endotoxin values (ng/L) for the LAL and CL assays.a LALc

Infection Gram-negative infection Sepsis a b c

CLc

Present

Absent

P

Present

Absent

P

200 ⫾ 50 180 ⫾ 15 130 ⫾ 37

180 ⫾ 15 232 ⫾ 70 190 ⫾ 18

0.62 0.25 0.89

370 ⫾ 72 455 ⫾ 80 470 ⫾ 57

265 ⫾ 45 265 ⫾ 40 157 ⫾ 140

0.22 ⬍0.05 ⬍0.001

Results are expressed as mean ⫾ SE. LAL assay using whole blood (5 ). CL, chemiluminescent EA assay (4 ).


20 –30 ng/L of Escherichia coli 055:B5 LPS was routinely achieved in donor blood with added LPS, with signals ⬎3 SD above the mean of parallel samples without added LPS. The assay demonstrated broad cross-reactivity with the LPS from pathogenic strains of E. coli, Pseudomonas, Salmonella, Klebsiella, and Serratia and was generally reactive with the widely conserved lipid A structure present in the LPS of most gram-negative organisms. The assay was not reactive with the cell wall products of grampositive bacteria, including preparations of pathogenic strains of heat-killed Staphylococci and Streptococci, lipoteichoic acids, or fungal cell wall extracts. Minimal blood constituents of the assay include blood plasma with an active classical and alternative pathway complement cascade and neutrophils or equivalent cultured phagocytic cells (retinoic acid differentiated HL-60 cells). Heat treatment of plasma to 60 °C for 10 min (which destroys complement activity) completely abolished antibody-dependent LPS signal detection. Previous studies (4 ) have demonstrated that the assay is insensitive to variations in neutrophil count (0.5–20 ⫻ 109 cells/L) and erythrocyte concentration (0 –140 g/L hemoglobin). Recovery studies using whole blood, from healthy volunteers, supplemented with increasing concentrations of LPS have demonstrated reproducible recovery using the EA assay but inconsistent recovery by acid extraction LAL methodology (4 ). Assay specificity was confirmed in studies on patient blood samples obtained from the Medical-Surgical Intensive Care Unit at the Toronto General Hospital after informed consent and approval by the Hospital Ethics Committee. Nineteen blood samples that were endotoxemic by the chemiluminescent EA assay were reassayed after treatment with polymyxin B sulfate, a well-characterized endotoxin-binding agent. Pretreatment of the blood samples with 0.5 mg/L polymyxin B sulfate eliminated the endotoxin-dependent signal in 16 samples and diminished a previously high signal to low endotoxin concentrations in 3 samples. The assay was not affected by lipemia (⬍15 g/L triglycerides), hemolysis (⬍5 g/L hemoglobin), icterus (⬍200 mg/L), or variations in hematocrit from (20 – 60%). Because of the inherent instability of blood samples in the assay beyond 90 min at room temperature, pooled within-run precision was evaluated by simultaneous assay of samples with and without added LPS by three operators on three separate luminometers in replicates of eight. A total of 10 samples (5 without added LPS, 5 with added LPS at 800 ng/L) were evaluated with a pooled CV of 12% achieved in samples without added LPS and 8% in samples with added LPS. Within-run precision was also evaluated in 200 intensive care unit (ICU) patient samples assayed in duplicate at two independent clinical trial sites. A within-run precision profile of CV vs EA indicated that CVs ⱕ10% were achieved across the dynamic range in ⬎75% of samples, with a pooled overall CV of 9.3% at both sites. In an initial clinical trial conducted at three academic centers, a cohort of 169 ICU patients was evaluated. After

1505

approval by the hospital ethics committee and informed consent, patients with suspicion of infection were enrolled into the trial, and EA was assayed daily for a maximum of 7 days by the chemiluminescent assay. The first endotoxin assay result was determined within 8 h of ICU admission and compared with microbiological culture from any body site, including blood harvested within a 12-h window before the endotoxin assay. The following assay performance characteristics were achieved: clinical sensitivity for detection of gram-negative infection in any body site (sputum, blood, abdominal, thoracic or pelvic abscess drainage, urine, bronchoalveolar lavage, tissue biopsy) was 96% (47 of 49), the negative predictive value was 96% (49 of 51), the specificity was 43% (49 of 114), and the positive predictive value with regard to a gram-negative infection was 45% (53 of 118). The seemingly low positive predictive value and specificity was attributable to the frequent presence of endotoxin in the blood in the absence of a culturable gram-negative organism. This phenomenon can occur in clinical situations where prior antibiotic use stuns or kills microorganisms, preventing their growth in culture media, or in situations where translocation of LPS through leaky gut or lung epithelium can occur. Persistent endotoxemia in these situations is likely to have major clinical relevance and is the subject of on-going clinical trials. In a subset of 74 patients from the clinical trial above, LAL assays were also included, using the whole blood method of Tamura et al. (5 ). No correlation between the LAL result and the chemiluminescent endotoxin assay was evident. The chemiluminescent endotoxin assay was significantly higher in patients with gram-negative infection and sepsis syndrome, whereas the LAL assay was not able to discriminate these patient subgroups (Table 1). Failure of the LAL assay to correlate with gram-negative bacteremia in patients with sepsis has recently been documented (6 ). We conclude that this novel method for the analysis of EA in whole blood exemplifies a generic analytical platform that incorporates the specificity of antibodies with the responsiveness of a patient’s immunological effector cells. Within minutes, this EA assay yields key information about the infective status of the patient that may be of prognostic and diagnostic value and may not be otherwise available by microbiological culture for several days. References 1. Suffredini AF, Fromm RE, Parker MM, Brenner M, Kovacs JA, Wesley RA, Parrillo JE. The cardiovascular response of normal humans to the administration of endotoxin. N Engl J Med 1989;321:280 –7. 2. Prins JM, van Agtmael MA, Kuijper EJ, van Deventer SJ, Speelman P. Antibiotic-induced endotoxin release in patients with gram negative urosepsis: a double-blind study comparing imipenem and ceftazidime. J Infect Dis 1995;172:886 –91. 3. Jansen PGM, Te Velthuis H, Oudemans-Van Straaten HM, Bulder ER, van Deventer SJ, Sturk A, et al. Perfusion-related factors of endotoxin release during cardiopulmonary bypass. Eur J Cardiothorac Surg 1994;8:125–9. 4. Romaschin AD, Harris DM, Ribeiro MB, Paice J, Foster DM, Walker PM, Marshall JC. A rapid assay of endotoxin in whole blood using autologous neutrophil dependent chemiluminescence. J Immunol Methods 1998;212: 169 – 85. 5. Tamura H, Tanaka S, Obayashi T, Yoshida M, Kawai T. A new sensitive

1506


method for determining endotoxin in whole blood. Clin Chim Acta 1991;200: 35– 42. 6. Bates DW, Parsonnet J, Ketchum PA, Novitsky TJ, Sands K, Hibberd PL, et al. Limulus amebocyte lysate assay for detection of endotoxin in patients with sepsis syndrome. AMCC Sepsis Project Working Group. Clin Infect Dis 1998;27:582–91.

channeling immunoassay (LOCITM) (5 ) is described. An ELISA for the determination of tHcy based on the modification of sample tHcy by alkylation and detection of the alkylated product was described previously (6 ). The current assay is designed for tHcy determination in serum or plasma at clinically relevant concentrations. Chemical reactions involved in the assay can be divided into three steps:

Homogeneous, Rapid Luminescent Oxygen Channeling Immunoassay (LOCITM) for Homocysteine, Yen Ping Liu,* Steve de Keczer, Svetlana Alexander, Marcel Pirio, Dariush Davalian, Nurith Kurn, and Edwin F. Ullman (Advanced Diagnostics Division, Dade Behring Inc., PO Box 49013, San Jose, CA 95161; * author for correspondence: fax 408-239-2707, e-mail [email protected])

(a) Serum disulfide bonds are reduced to release Hcy as monothiol; (b) A derivative [Hcy-acetylbenzoic acid (Hcy-ABA)] is formed by reacting with the alkylating agent p-(2chloroacetyl)benzoic acid phosphate (CABA-phosphate); and (c) The Hcy-ABA concentration is measured in a competition assay using an anti-Hcy-ABA antibody.

Homocysteine (Hcy) is present in plasma primarily bound as disulfides with itself, Cys, and albumin (⬃70%) (1– 4 ). Total homocysteine (tHcy) in serum or plasma is markedly increased in patients with cobalamin or folate deficiency (3 ), and decreases only when they are treated with the deficient vitamin. tHcy is therefore of clinical relevance, with reference values in fasting subjects of ⬃5–15 ␮mol/L (1 ). In addition, even a moderate increase of Hcy (hyperhomocysteinemia) is a risk factor for premature cardiovascular disease (4 ). These disorders justify introduction of the tHcy assay in the routine clinical chemistry laboratory. The development of a rapid, homogeneous assay for Hcy in serum or plasma using the luminescent oxygen

The first step involves the release of bound Hcy by reduction of serum disulfides with tris-(2-carboxyethyl)phosphine (TCEP); the second step involves the derivatization of Hcy and cysteine with CABA to produce acyclic Hcy-ABA and cyclic Cys-ABA as shown in the alkylation reaction (Fig. 1A); the third step involves the selective binding of Hcy-ABA to anti-Hcy-ABA coated on chemiluminescent latex particles in competition with binding of Hcy-ABA coated on sensitizer latex particles. A two-reagent assay protocol was used in this system: The first reagent contains TCEP, CABA-phosphate (the enol-phosphate derivative of chloroacetylbenzoic acid), and latex particles coated with Hcy-ABA (the product of alkylation of Hcy with CABA). The phosphate protecting

Fig. 1. Alkylation reactions in the LOCI assay for Hcy (A) and correlation for Hcy quantification by HPLC and LOCI (B). (A), the alkylating reagent CABA-phosphate, the enol-phosphate derivative of chloroacetylbenzoic acid, was deprotected with alkaline phosphatase and then combined with sample containing Hcy and Cys in assay buffer. Modified Cys forms a six-member ring, but the modified Hcy does not. Thus, antibodies to the alkylated Hcy (Hcy-ABA) can differentiate Hcy-ABA from the cyclic alkylated cysteine (Cys-ABA). (B), correlation of tHcy results in serum clinical samples assayed by LOCI and HPLC methods. The LOCI assay protocol is described in the text.


group of CABA is required for compatibility with TCEP and greatly increases the stability and solubility of CABA. A second reagent contains alkaline phosphatase (required for the release of CABA) and latex particles coated with an antibody specific to Hcy-ABA. Disulfide bonds in the sample are reduced when the sample is combined with the first reagent. When the second reagent is added, CABA is released and alkylates the sulfhydryl groups of Hcy and Cys. The amino group of the alkylated cysteine but not the alkylated Hcy can react internally with the ketone introduced by CABA to give a cyclic imide (Fig. 1A). The alkylated Hcy derivative (Hcy-ABA) then binds to the antibody that is coated on the chemiluminescent latex particles. The structurally distinct cyclic cysteine derivative (Cys-ABA) is not recognized by the antibody. This overcomes the problem of developing antibodies to Hcy that do not cross-react with cysteine, which usually is present at ⬃25-fold molar excess concentration (7 ). The detection step is superficially similar to latex agglutination but uses a low concentration of particles and a novel photochemically triggered chemiluminescent detection technique (5 ). Two types of latex particles (0.2– 0.3 ␮m in diameter) are used. Both types of particles have a hydrogel coating that protects the particles from nonspecific interactions with matrix components and provides a functionalized surface to which antibodies and analytes can be covalently attached. Binding of the two particles is mediated by Hcy-ABA and is detected by measurement of the chemiluminescence that ensues following brief irradiation of the assay mixture. The chemiluminescence is generated by reaction of an olefinic acceptor in the chemiluminescent particles with singlet oxygen that is generated by a photosensitizer particle in close proximity. The adduct that is formed has a half-life of ⬃0.6 s and decays with emission of light at wavelength ⬎600 nm. Because of the short lifetime of singlet oxygen in water (⬃4 ␮s), it can only diffuse a few microns. A signal can therefore be produced only when the particles are closely associated. The LOCI Hcy assay is performed on an automated instrument (Tecan), which was modified with a pulsed diode laser and a luminometer (5 ). The patient sample is incubated by mixing 5 ␮L of serum or EDTA-treated plasma with 50 ␮L of the first reagent, which contains 2 mmol/L TCEP, 5 mmol/L CABA-phosphate, and 5 ␮g of Hcy-ABA-coated sensitizer particles. After a 7-min incubation at 37 °C and the addition of 50 ␮L of the second reagent, which contains 50 ␮g of alkaline phosphatase and 12.5 ␮g of anti Hcy-ABA monoclonal antibody-coated chemiluminescent particles. After the addition of 145 ␮L of 0.1 mol/L borate buffer, pH 9.2, the mixture is incubated an additional 2.6 min. The chemiluminescent signal is then measured by repetitively irradiating at 680 nm for 1 s and reading at 600 – 620 nm for 1 s. The assay signal is inversely related to the amount of tHcy present in the serum sample. Concentrations are determined using pooled serum calibrators. Calibrators were prepared from a pooled serum sample supplemented with known amounts of Hcy and verified by a HPLC method (3 ). Signal modulation of 65–75% over

1507

the range of 0 – 60 ␮mol/L Hcy was demonstrated. Crossreactivity of l-cysteine and l-methionine, each at 10 mmol/L in assay buffer, was assayed with Hcy LOCI. The observed relative luminescent units corresponded to 0.87 and 0.67 ␮mol/L tHcy, respectively. The intraassay CVs obtained by assaying five replicates on the same carousel were 5.9%, 2.7%, and 3.4% for 10, 30, and 60 ␮mol/L Hcy. This process was repeated twice to determine the interassay CVs, which were 5.3%, 2.7%, and 3.9% for 10, 30, and 60 ␮mol/L Hcy, respectively. The recovery of Hcy from patient samples supplemented with exogenous Hcy (n ⫽ 2 at five concentrations) was 91–106% with a mean of 97.3%. Regression analyses of the results obtained using 97 serum or 50 plasma clinical samples analyzed by a single LOCI measurement (y) and by an established HPLC method (x) gave the following equations: for serum, y ⫽ 1.00x ⫺ 0.50 (r ⫽ 0.98; slope ⫽ 1.00; n ⫽ 97), as shown in Fig. 1B; and for plasma, y ⫽ 0.86x ⫹ 1.24 (r ⫽ 0.96; slope ⫽ 0.90; n ⫽ 50). Total Hcy in the majority of the plasma samples was ⬍20 ␮mol/L, and mostly in the range of 5–15 ␮mol/L. Samples with results ⬎100 ␮mol/L tHcy had to be diluted and reassayed and were excluded from this study. In conclusion, we have demonstrated that LOCI is applicable to an antibody-based assay for tHcy quantification. This unique method is simple, rapid, and highly robust and suitable for routine determinations of serum or plasma tHcy concentrations in clinical laboratories.

We thank Dr. Frederick Van Lente and Ingrid Raulinaitis at the Cleveland Clinical Foundation, Department of Clinical Pathology, Section of Biochemistry (Cleveland, OH) who provided all of the samples and HPLC analyses for this study. We also thank Drs. A. Dafforn and S. Rose for reviewing this manuscript.

References 1. Ueland PM, Refsum H, Stabler S, Malinow MR, Andersson A, Allen RH. Total homocysteine in plasma or serum: methods and clinical applications. Clin Chem 1993;39:1764 –79. 2. Fiskerstrand T, Refsum H, Kvalheim G, Ueland PM. Homocysteine and other thiols in plasma and urine: automated determination and sample stability. Clin Chem 1993;39:263–71. 3. Jacobsen DW, Gatautis VJ, Green R, Robinson K, Savon SR, Secic M, et al. Rapid HPLC determination of total homocysteine and other thiols in serum and plasma: sex differences and correlation with cobalamin and folate concentrations in healthy subjects. Clin Chem 1994;40:873– 81. 4. Ueland PM, Refsum H. Plasma homocysteine, a risk factor for vascular disease: plasma levels in health, disease, and drug therapy. J Lab Clin Med 1989;114:473–501. 5. Ullman EF, Kirakossian H, Switchenko A, Ishkanian J, Ericson M, Wartchow C, et al. Luminescent oxygen channeling assay (LOCITM): sensitive, broadly applicable homogeneous immunoassay method. Clin Chem 1996;42: 1518 –26. 6. Van Atta RB, Goodman TC, Ullman EF. Immunoassay for homocysteine. US Patent No. 5,478,729, December 26, 1995. 7. Guttormsen AB, Mansoor AM, Fiskerstrand T, Ueland PM, Refsum H. Kinetics of plasma homocysteine in healthy subjects after peroral homocysteine loading. Clin Chem 1993;39:1390 –7.

1508


A Distinctive Surface for Phospholipid Assays, Susan K.W. Nanda* and Wendy K. Scholz (Nalge Nunc International, 2000 N. Aurora Rd., Naperville, IL 60563; * author for correspondence: fax 630-416-2519, e-mail snanda@ nalgenunc.com) Solid-phase immunoassay is a commonly used immunological technique in many clinical and research laboratories. It is known that proteins will adsorb to a solid surface such as polystyrene. Many studies have also indicated that surface modifications such as the introduction of different functional groups onto the polystyrene surface will enhance the overall protein-binding properties of the plastic polymer (1 ). Nevertheless, the increased binding also means increased nonspecific binding, producing higher background signals. The currently available polystyrene surfaces work well for most protein antigens. However, these polystyrene surfaces are not adequate for the measurement of all antigenic substances. Certain antigenic lipids are not easily detected and measured using these surfaces (2 ). A commonly used diagnostic test for anti-phospholipid antibodies (aPLs) is an ELISA using cardiolipin, an anionic phospholipid, as the antigen in the solid-phase immunoassay (3 ). aPLs are antibodies associated with the clinical syndrome of thrombosis and recurrent midtrimester fetal death and other, less common manifestations, termed antiphospholipid syndrome (4 ). aPLs are also detected in patients with systemic lupus erythematosus and related autoimmune disorders (5–7 ). There are numerous critical variables for anticardiolipin assays. Critical factors that can influence ELISA results include the amount of phospholipids coated on the surface, the solvent for the phospholipids, choice of blocking agents and serum diluent, incubation temperature, and the type of microwell plates (8 –10 ). To develop a solidphase immunoassay optimized for phospholipids, we examined various modified polystyrene surfaces for the ability to bind protein and phospholipids, including cardiolipin. The following reagents were obtained from Sigma: cardiolipin (diphosphatidylglycerol), sodium salt from beef heart, 4.8 g/L in ethanol; o-phenylenediamine dihydrochloride (OPD), 30-mg tablets; p-nitrophenyl phosphate, disodium; and diethanolamine, ACS reagent grade. Nitrobenzoxadizole (NBD)-labeled 1-palmitoyl-2-[6-(7nitro-2-1,3-benzoxadiazol-4-y)]-sn-glycerol 3-phosphoserine (sodium salt), Mr 796.81, 1 g/L in chloroform; and 1,2-dioleoyl-sn-glycerol-3-[phospho-l-serine] (sodium salt), Mr 810.03, 10 g/L in chloroform were obtained from Avanti Polar Lipids. Acridine orange 10 nonyl bromide (Mr 472.51) was obtained from Molecular Probes. All other laboratory chemicals were analytical grade and were from Sigma and Mallinckrodt. Polyclonal rabbit anti-sheep (18 g/L) and rabbit antihuman ␣1-fetoprotein-horseradish peroxidase (HRP) conjugate (1.3 g/L) were purchased from Dako. Lyophilized positive serum samples for IgG anti-cardiolipin antibod-

ies (aCLs) were obtained from Louisville APL Diagnostics. Alkaline phosphatase-conjugated goat anti-human IgG (␥-chain specific) affinity-isolated antibody and adult bovine serum were obtained from Sigma. For the protein binding assay, the 96-well plates were coated (200 ␮L per well) with an antibody mixture solution containing 3600 ␮g/L rabbit anti-sheep IgG and 33 ␮g/L rabbit anti-human ␣1-fetoprotein-HRP in 0.05 mol/L carbonate buffer, pH 9.6. The plates were sealed with Nunc™ sealing tape and incubated overnight in the dark at room temperature. The plates were washed three times with a washing buffer [0.15 mol/L phosphatebuffered saline (PBS), pH 7.2, containing 0.2 mol/L NaCl and 0.5 mL/L Triton X-100]. The washed plates were incubated with a substrate solution (200 ␮L per well) consisting of 0.6 g/L OPD and 0.5 mL/L H2O2 (300 g/L) in 0.1 mol/L citrate phosphate buffer, pH 5.0. The colorimetric reaction was stopped by the addition of 2 mol/L H2SO4 (150 ␮L per well). Absorbance at 490 nm was determined using a microplate reader. For the cardiolipin binding assay, cardiolipin diluted in ethanol was coated onto the 96-well microplate by evaporation overnight at 4 °C. The plate was washed three times with PBS. One hundred microliters of a fluorescent dye, 10-N-nonyl-acridine orange (15 ␮mol/L), was added to each well. The plate was incubated at room temperature under reduced light. Nonbound dye was removed by washing the plate with PBS. Fluorescence was detected by a fluorescence microplate reader using two filter sets: filter set 1, excitation at 485 nm, emission at 530 nm; and filter set 2, excitation at 485 nm, emission at 645 nm. For the phospholipid binding assay, NBD-labeled phosphatidylserine and nonlabeled phosphatidylserine were mixed at a ratio of 1:20 in chloroform:methanol (1:4, by volume) in a glass tube. Different volumes (15–90 ␮L per well) of the mixed phosphatidylserine solution were dispensed into a 96-well plate. The solutions in the wells were dried under nitrogen to minimize oxidation. The plate was washed three times with PBS, and 100 ␮L of PBS was then added to each well. Fluorescence was detected by a fluorescence plate reader with excitation at 485 nm and emission at 530 nm. For the anticardiolipin ELISA, cardiolipin (50 mg/L in alcohol) was immobilized on a 96-well plate by evaporation overnight at 4 °C. The wells were blocked with bovine serum albumin (BSA) buffer (PBS containing 20 g/L BSA) and incubated for 1 h at 4 °C. The plate was washed three times with PBS. Human sera positive for aCLs were added to the wells (100 ␮L per well) and incubated at 4 °C for 2 h. After incubation, the plate was washed three times with PBS. A secondary antibody (goat anti-human IgG-alkaline phosphatase conjugate diluted 1:1000 in PBS with 100 mL/L adult bovine serum) was added and incubated for 1 h at 4 °C. The plate was washed with PBS, and buffered substrate (1 g/L pnitrophenyl phosphate in diethanolamine buffer) was added. The plate was sealed with tape and incubated for 1 h at room temperature under reduced light. The colorimetric reaction was stopped by the addition of 3 mol/L


1509

Fig. 1. Binding of phosphatidylserine (A) and cardiolipin (B) on modified surfaces. Bars, SD.

NaOH, and the absorbance at 405 nm was measured by a microplate reader. We compared the protein binding characteristics of polystyrene plates with nonmodified surfaces and plates with surfaces modified by two different processes. Nunc plates (96-well) with three different surfaces, nonmodi-

fied, MultiSorpTM, and MaxiSorpTM, were incubated with an antibody mixture solution containing rabbit anti-sheep IgG and rabbit anti-human ␣1-fetoprotein conjugated with HRP as described above. Bound IgG was detected by the colorimetric reaction of OPD and H2O2 with HRP. Absorbance measurements at 490 nm (A490 nm) reflected

1510


attachment of IgG on the surfaces. The mean A490 nm (n ⫽ 96) was 11 077.83 ⫾ 409.51 for plates with the nonmodified surface, 857.04 ⫾ 330.45 for plates with the MultiSorp surface, and 16 630.98 ⫾ 588.46 for plates with the MaxiSorp surface. A comparison of the absorbance readings of the three surfaces clearly indicated that the MultiSorp surface has a low affinity for IgG, i.e., at the given coating concentration and equal coating time as described above, less IgG was bound to the MultiSorp surface compared with the other two surfaces. To explore the binding capability of the modified surfaces for molecules other than IgG, we determined the binding of phosphatidylserine, an anionic phospholipid, on these surfaces. Acyl chain-labeled, fluorescent NBDphosphatidylserine was used to measure phospholipid binding to plates with nonmodified surfaces and plates with MultiSorp and MaxiSorp surfaces. Fluorescent NBDlabeled phosphatidylserine was mixed with nonlabeled phosphatidylserine at a ratio of 1:20, diluted in chloroform:methanol (1:4, by volume) and dried under nitrogen onto plates with different surfaces as described above. Relative fluorescence was measured using a fluorescent plate reader. As shown in Fig. 1A, plates with MultiSorp and MaxiSorp surfaces demonstrated a higher capacity for binding phosphatidylserine. Because cardiolipin, a negatively charged phospholipid, is commonly used as the solid-phase antigen in most diagnostic ELISAs for aPLs, we measured the binding of cardiolipin on the modified surfaces. Fig. 1B shows the cardiolipin binding data for these surfaces. Cardiolipin binding on the solid surfaces was detected by the fluorescent dye 10-N-nonyl acridine orange as described above. 10-N-Nonyl acridine orange has been reported to interact specifically with cardiolipin isolated from mitochondria and whole cells with a stoichiometry of 2:1. Upon contact with cardiolipin, the dye forms dimers, producing a shift in emission fluorescence from 525 nm to 640 nm (11, 12 ). In Fig. 1B, the amount of cardiolipin bound to the solid surfaces was quantified by the red emission wavelength (640 nm) of the dye after unbound dye was removed by washing. A slightly increased binding of cardiolipin on both modified surfaces was observed, as shown in Fig. 1B. To demonstrate that cardiolipin immobilized on the MultiSorp surface retained an active functional configuration, we performed an anticardiolipin ELISA using a plate with the modified surface. The ELISA results are shown in Table 1. These results indicate that cardiolipin directly bound on the modified surface contains an epitope that is recognized by aCLs in human serum. The bound cardiolipin is not recognized by IgG in rabbit serum. It is also interesting to note that nonspecific proteins such as BSA and other proteins in adult bovine serum did not interact with cardiolipin on this modified surface, thus providing very low background in the ELISA. In summary, we learned that a surface modification process can both enhance the binding of anionic phospholipids and reduce the binding of IgG on a polystyrene

Table 1. Results of an anticardiolipin ELISA on MultiSorp surface.a Test group

Mean A405 ⴞ SD (n ⴝ 8 wells)

PBS BSAb ABSc Bufferd R-IgGe H-IgG-Hf H-IgG-Mg H-IgG-Lh

0.1273 ⫾ 0.0775 0.0981 ⫾ 0.0212 0.1165 ⫾ 0.0653 0.0890 ⫾ 0.0136 0.0180 ⫾ 0.0666 1.4010 ⫾ 0.2973 0.5962 ⫾ 0.1166 0.2310 ⫾ 0.0236

a

All wells were coated with cardiolipin as described in the text. BSA (20 g/L) in PBS used as blocking agent. c Adult bovine serum (100 mL/L) in PBS used as the diluent for all antibodies in assay. d Diethanolamine buffer used as buffer for substrate in colorimetric detection. e Rabbit anti-sheep serum (18 g/L) diluted 1:100 with adult bovine serum. f–h Human sera positive for IgG aCLs: f positivity ⬎80 GPL units (high); g 20 ⬍ positivity ⬍ 80 GPL units (medium); h 10 ⬍ positivity ⬍ 20 GPL units (low). All sera were diluted 1:100 with adult bovine serum (3 ). b

surface. The resulting surface is ideal for phospholipidbased immunoassays and may also be useful for other assays requiring low protein binding.

We thank Erika Tracy for technical assistance and Tony Chin for assistance in graphics.

References 1. Butler JE. The behavior of antigens and antibodies immobilized on a solid phase. In: Van Regenmortel MHV, ed. Structure of antigens, Vol. 1. Boca Raton, FL: CRC Press, 1992:209 –59. 2. Voller A, Bidwell DE, Bartlett A. Enzyme immunoassays in diagnostic medicine: theory and practice. Bull World Health Organ 1976;53:55– 65. 3. Harris EN, Gharavi AE, Patel SP, Hughes GRV. Evaluation of the anticardiolipin antibody test: report of an international workshop held 4 April 1986. Clin Exp Immunol 1987;68:215–9. 4. McNiel HP, Chesterman CN, Krilis SA. Immunology and clinical importance of antiphospholipid antibodies. Adv Immunol 1991;49:193–280. 5. Harris EN, Gharavi AE, Hughes GRV. Anti-phospholipid antibodies. Clin Rheum Dis 1985;11:591– 609. 6. Lopez LR, Hites MJ, Loker J. Association of anti-cardiolipin antibodies with thrombosis, thrombocytopenia and recurrent abortion in patients with systemic lupus erythematosus (SLE). Thromb Haemost 1991;65:1004 –9. 7. Mahmood T, Racis SP, Krey PR. IgA anti-cardiolipin antibody (aCL) in systemic lupus erythematosus (SLE). Arthritis Rheum 1990;33:R45–52. 8. Kilpatrick DC. Factors affecting cardiolipin antibody assays: modification with polyethylene glycol compound. Br J Haematol 1998;100:52–7. 9. Firer MA, Spivak T, Shoenfeld Y, Slor H. The effects of incubation temperature and coating procedure on the measurement of antibodies to cardiolipin. J Immunol Methods 1991;14:31–9. 10. Hughes JR, Davies JA, Prentice CRM. The effect of different microtitre plates on the enzyme-linked immunosorbent assay (ELISA) for anticardiolipin antibodies (ACAs). Thromb Res 1996;84:217–22. 11. Petit JM, Huet O, Gallet PF, Maftah A, Ratinaud MH, Julien R. Direct analysis and significance of cardiolipin transverse distribution in mitochondrial inner membranes. Eur J Biochem 1994;220:871–9. 12. Gallet PF, Maftah A, Petit JM, Denis-Gay M, Julien R. Direct cardiolipin assay in yeast using the red fluorescence emission of 10N-nonyl acridine orange. Eur J Biochem 1995;228:113–9.


Toward a Rapid, Integrated, and Fully Automated DNA Diagnostic Assay for Chlamydia trachomatis and Neisseria gonorrhoeae, Farzad Pourahmadi,* Mike Taylor, Greg Kovacs, Kristen Lloyd, Stan Sakai, Tamlyn Schafer, Bret Helton, Linda Western, Sandy Zaner, Jesus Ching, Bill McMillan, Phil Belgrader, and M. Allen Northrup (Cepheid, 1190 Borregas Ave., Sunnyvale, CA 94089; * author for correspondence: fax 406-541-4192, e-mail pourahmadi@ cepheid.com) Molecular testing for the diagnosis of bacterial or viral infections in raw clinical specimens requires complex, multistep procedures to release and isolate nucleic acid before PCR amplification (1, 2 ). Laboratory bench-top sample preparation procedures are very labor- and equipment-intensive, which reflects on total assay cost and increased susceptibility to sample and reagent carryover (3, 4 ). The need for automation has led to the development and introduction of robotics-based laboratory instruments with discrete operations that simulate the basic functions of a laboratory technician. These systems are typically developed for high-throughput applications and usually require intermittent operator involvement. In addition, because they directly emulate the operator’s manual functions for processing boluses of sample and reagents, the issue of sample or reagent carryover and carryin remains unresolved, and the complexity of the robotic mechanisms themselves contributes to high capital costs and poor reliability. Several factors currently contribute to and facilitate the development of the next generation of automated and integrated diagnostics instruments. One factor includes the recent advances in the microfluidics field of miniaturized integrated platforms and supporting technologies, which potentially enable the seamless execution of a sequence of protocols (5 ). Miniaturized processor-controlled microfluidic platforms contain chambers and valves, and encapsulate entities such as filter membranes or solid-phase particles. They are able to execute important processes such as cell capture, lysis, and fluid mixing, which currently require discrete laboratory instruments (6 ). Another factor is the development of a miniature, rapid PCR thermocycler device with real-time optical detection capable of multiplex analysis (7–9 ). Its modularity and adaptability allows for easy integration into an instrument platform downstream of a sample preparation module. It is possible to envision a one-step operation covering sample preparation and analysis, from raw sample to results reporting. Finally, the utilization of inexpensive materials, such as molded plastics, allows for the design of devices capable of simultaneously handling large sample volumes, up to many milliliters (required for clinical sensitivity), and microfluidic volumes, down to a few microliters (desirable for expensive analytical reagents) (8, 10 ). This report describes progress in the development of a fully integrated and automated instrument for which all sample preparation and PCR functions are integrated in a single platform. The platform is a plastic cartridge (Fig.

1511

1A) mounted in an instrument (Fig. 1B). The instrument contains a miniature, single-site thermally controlled fluorometer consisting of an assembly of two heater plates with embedded heaters and thermal sensors, a cooling fan, an optical subsystem with LEDs and detectors, and an analog printed circuit board. Heating and cooling rates of 7– 8 °C and 3– 4 °C, respectively, have been achieved for a 100-␮L reaction volume (data not shown). The cartridge itself comprises three flat plastic acrylic parts sandwiched together with laser-cut thin [0.762-mm (0.030-inch)] silicone gasket layers that provide fluid sealing and surfaces for the valves and an ultrasonic horn. It contains chambers for the sample, wash buffer, lysis buffer, neutralization buffer, PCR mastermix, and waste. In addition, there is a chamber for holding filters for cell capture and lysis and an integrated 100-␮L PCR reaction tube. Details of the filter stack assembly are shown in Fig. 1C. It consists of layers of 1.27-cm (0.5-inch) filters and glass beads: a 5 ␮m filter (LoProDyne; Pall Corp), a 1.2 ␮m filter (Versapor, Pall Corp), glass beads (ⱕ106 ␮m; Sigma), and a 0.2 ␮m filter (Durapore; Millipore). The tip of an ultrasonic horn (Selfridge and Associates) is interfaced with the cartridge by pressing against the 0.762-mm (0.030-inch) silicone gasket membrane that forms one wall of the filter stack chamber. All of the cartridge chambers are interconnected by D-shaped microfluidic lines [diameter, 0.813 mm (0.032 inches)]. Fluid flow is driven pneumatically through one or more of six ports on top of the cartridge by a minipump in the instrument, and the flow rate is controlled by digital regulators (Omega). The direction of the fluid flow is controlled by nine membrane valves individually actuated by external solenoid valves pinching against the silicone membrane. A Visual Basic software program is used to control the system, and the interface is shown schematically in Fig. 1D. The area within the dashed red square in Fig. 1D represents the physical part of the cartridge, and the area outside of the square is the instrument. The specimen containing the bacterial cells is pumped from the sample chamber through the filter stack chamber and into waste. After cell capture and washing, the fluid is pressurized, and cells are lysed by sonication and DNA is released. Chamber pressurization is required to ensure adequate coupling between the horn and the silicone gasket membrane. In a study to optimize the fluidic protocol, type E Chlamydia trachomatis (elementary and reticulate bodies; Intracel Bartels) and gonococcal cells from an overnight culture were diluted into 5.5 mL of pooled normal urine to create a mock clinical specimen. The final dilution of the Chlamydia was 1:250 000, and the concentration of gonococcal cells was 20 000 colony forming units/mL of urine. To load the cartridge, 5.0 mL of the urine specimen was added to the “sample” chamber. In addition, 2.0 mL of wash solution (10 mmol/L Tris, 1 mmol/L EDTA, 0.2 g/L sodium azide, pH 8.3), 1.5 mL of lysis solution (Roche Amplicor Chlamydia Resuspension Diluent), 0.5 mL of neutralizer solution (Roche Amplicor Chlamydia Urine Diluent), and 100 ␮L of PCR mastermix was also added to

1512


Fig. 1. Description (A–D) and performance (E–G) of the prototype GeneXpert DNA diagnostic platform. (A), plastic sample preparation cartridge with integrated PCR tube; (B), instrument; (C), exploded view of cartridge filter stack assembly; (D), software fluidic interface. Area within the red dashed square in D represents the physical part of the cartridge; the area outside of the square is the instrument. (E), multiplex PCR results for a sample obtained from the neutralization chamber. (F), PCR results for a sample obtained from the mastermix chamber. (G), lysis efficiency of the ultrasonic-based method vs the reference method.


other respective chambers. The cartridge was then placed in the instrument, and test protocols were automatically executed. To evaluate recovery, a PCR mastermix was developed for the multiplex detection of the Chlamydia and gonococcal DNA using TaqMan® probes. The multimix was composed of 200 nmol/L FAM/TAMRA probe and 200 nmol/L TET/TAMRA probe specific for gonococcal and Chlamydia DNA, respectively; 500 nmol/L primers; 5 U of Platinum® Taq (Life Technologies) per reaction; 200 ␮mol/L dNTPs; 50 mmol/L KCl; 8 mmol/L MgCl2; 0.5 g/L bovine serum albumin; and 1 mL/L Tween 20 in 10 mmol/L Tris, pH 8.3. The primer and probe sequences for Neisseria gonorrhoeae were designed using Primer ExpressTM software to amplify a portion (256 bp) of a 2-kb region of chromosomal DNA shown to be specific for N. gonorrhoeae with no cross-reactivity to N. meningitidis (11 ): Forward primer: 5⬘-TATTACGTTCAGGCGAAGCTGTATAACTTTG-3⬘ Reverse primer: 5⬘-TTAATTCCAACATACGGCGTGTTTTACCGCT-3⬘ Probe: 5⬘-FAM-TGGCGTACCTCAATTTCGTGAACGTGTGCT-TAMRA-3⬘ In addition, primer and probe sequences selected by Primer Express software were used to detect a fragment (311 bp) of the endogenous plasmid associated with C. trachomatis. The sequences were designed with melting temperatures suitable for multiplexing with the N. gonorrhea-specific sequences. Forward primer: 5⬘-TTGAGCGTATAAAGGGAAGGCTTGACAGTG-3⬘ Reverse primer: 5⬘-GTTGAGTAACCGCAAGATTTATCGCCATGT-3⬘ Probe: 5⬘-TET-TATATTCTCACAGTCAGAAATTGGAGTGCTGGCTCGTATA-TAMRA-3⬘ After the fluidic protocol was complete, samples were taken from the PCR tube, from the mastermix chamber, or the neutralization chamber, and amplification with realtime detection was performed in a Cepheid Smart Cycler® thermocycler. Cycling conditions were as follows: hold at 95 °C for 60 s; 45 cycles of 5 s at 95 °C and 30 s at 65 °C. Time to completion was typically ⬍30 min. Panels E and F in Fig. 1 show the PCR results for neutralized lysate and a complete reaction mixture obtained from the mastermix chamber, respectively. To separately evaluate the efficiency of the ultrasonicsbased lysis procedure, Chlamydia cells were also lysed in a PCR reaction tube, containing glass beads, pressed against an externally mounted ultrasonic horn. Recovery was measured using an FDA-approved kit (Roche Amplicor® Chlamydia Test kit). The efficiency of lysis was comparable to that obtained with the reference kit (Fig. 1G). We have shown significant progress in the development of a cartridge-based instrument for automated sample preparation and subsequent DNA detection of bacte-

1513

ria. Once the reagents and samples are placed in the cartridge, the entire process of sample preparation from 5 mL of urine sample was completed within 2 min. Rapid lysis of bacteria and release of DNA was achieved by sonication in the presence of glass beads, using an ultrasonic horn separated from the liquid sample by a flexible membrane. To date, we have demonstrated feasibility for the individual sample preparation steps, a fluidic management scheme, a complete fluidic protocol, and a rapid PCR-based amplification. We currently are developing a fully integrated system with improved sample handling, more optimal fluidic protocol, and a more sensitive, four-color fluorometer module. References 1. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn HA, Arnheim N. Enzymatic amplification of ␤-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 1985;230:1350 – 4. 2. Vet JAM, Majithia AR, Marras, SAE, Tyagi S, Dube S, Poiesz BJ, Kramer FR. Multiplex detection of four pathogenic retroviruses using molecular beacons. Proc Natl Acad Sci U S A 1999;96:6394 –9. 3. Vahey MT, Wong MT, Michael NL. A standard PCR protocol: rapid isolation of DNA and PCR assay for globin. In: Dieffebach CW, Dveksler GS, eds. PCR primer: a laboratory manual. New York: Cold Spring Harbor Laboratory Press, 1995:17–22. 4. McCormick RM. A solid-phase extraction procedure for DNA purification. Anal Biochem 1989;181:66 –74. 5. Wilding P, Kricka LJ, Cheng J, Hvichia G, Shoffner MA, Fortina P. Integrated cell isolation and polymerase chain reaction analysis using silicon microfilter chambers. Anal Biochem 1998;257:95–100. 6. Waters LC, Jacobson SC, Kroutchinia H, Khandurina J, Foote RS, Ramsey JM. Microchip device for cell lysis, multiplex PCR amplification, and electrophoretic sizing. Anal Chem 1998;70:158 – 62. 7. Wittwer CT, Fillmore GC, Garling DJ. Minimizing the time required for DNA amplification by efficient heat transfer to small samples. Anal Biochem 1990;186:328 –31. 8. Northrup MA, Christel LA, McMillan WA, Petersen K, Pourahmadi F, Western L, Young S. A new generation of PCR instruments and nucleic acid concentration systems. In: Innis MA, Gelfand DH, Sninsky JJ, eds. PCR application: protocols for functional genomics. San Diego: Academic Press, 1996:105–26. 9. Ibrahim MS, Lofts RS, Jahrling PB, Henchal EA, Weedn VW, Northrup MA, Belgrader P. Real-time microchip PCR for detecting single-base differences in viral and human DNA. Anal Chem 1998;70:2013–7. 10. Ogura M, Agata Y, Watanabe K, McCormick RM, Hamaguchi Y, Aso Y, Mitsuhashi M. RNA chip: quality assessment of RNA by microchannel linear gel electrophoresis in injection-molded plastic chips. Clin Chem 1998;44: 2249 –55. 11. Miyada, CG, Born TL. A DNA sequence for the discrimination of Neisseria gonorrhoeae from other Neisseria species. Mol Cell Probes 1991;5:327–35.

Direct Acquisition of Matrix-assisted Laser Desorption/ Ionization Time-of-Flight Mass Spectra from Laser Capture Microdissected Tissues, Darryl Erik Palmer-Toy,1* David A. Sarracino,2 Dennis Sgroi,1 Rebbecca LeVangie,1 and Peter E. Leopold2 (1 Department of Pathology, Massachusetts General Hospital, Warren 2, Boston, MA 02114; 2 ProteiGene, Inc., 44 Manning Rd., Billerica, MA 01821; * address correspondence to this author at: Department of Pathology, Johns Hopkins Medical Institutions, 600 N. Wolfe St./Meyer B-125, Baltimore, MD 21287-7065; fax 410-614-7609, e-mail [email protected]) Recent studies have demonstrated the usefulness of laser capture microdissection (LCM) in gene expression studies

1514


of pure cell populations derived from healthy and diseased tissues (1–3 ). However, mRNA expression does not necessarily correlate with protein abundance or predict posttranslational modifications (4 ). Currently, a few groups have reported details of protein expression profiling of cells captured by LCM, using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) followed by peptide mass fingerprinting of selected proteins (5, 6 ). Although such an approach represents the gold standard for global analysis of protein expression, 2D-PAGE often is impractical for diagnostic applications. We describe a rapid and sensitive method to obtain an abridged protein expression profile from microdissected cells by the direct acquisition of matrix-assisted laser desorption/ionization time-of-flight (MALDI-ToF) mass spectra from LCM transfer films. LCM uses an infrared laser to transiently and focally melt an ethylene vinyl acetate (EVA) thermoplastic film applied to a tissue section. During the infrared laser pulse, the EVA transfer film fuses with the underlying cells, which remain adherent when the film is separated from the tissue section. In most applications, the transfer film is subsequently immersed in a lysis buffer to extract soluble molecules from the captured cells and their microenvironment, and the lysate is used in further analysis. Our approach is to obtain MALDI-ToF mass spectra directly from microdissected cells of interest. MALDI-ToF allows the simultaneous molecular mass characterization of a broad variety of biological molecules up to 100 kDa or higher. Mass measurements of 1 pmol of an analyte present in a 1-␮L specimen are routine for pure analytes and simple mixtures. MALDI-ToF matrix solution is applied directly to the cell-adherent surface of LCM films. The matrix solution extracts soluble molecules from the captured cells and their microenvironment, which cocrystallize with the matrix as the solvent evaporates. Excitation with an ultraviolet laser pulse disrupts specific chemical bonds within matrix molecules, rapidly vaporizing the crystals, ionizing and liberating the incorporated biomolecules into the gas phase for mass analysis. Mass spectra are readily collected from the crystal-coated film using a standard MALDI-ToF mass spectrometer. Notably, the MALDI-ToF spectra do not require elaborate sample preparation, e.g., chromatography, affinity capture, or other purification step. Furthermore, no instrumentation beyond the LCM device and a MALDI-ToF mass spectrometer is required. The EVA substrate is not ionized, nor does it suppress ionization. Matrix co-crystallized LCM samples are stable for at least 2 weeks under reduced pressure. All tissue utilized for this study was obtained from a modified radical mastectomy specimen from a single patient. Tissue in excess of what was necessary for diagnostic purposes was obtained ⬍15 min after removal from the patient, embedded in cryostat mounting medium (TISSUE-Tek O.C.T.; Sakura Finetek U.S.A.), and frozen in liquid nitrogen. The tissues were sectioned at 8 ␮m in a cryostat, mounted on uncoated glass slides, and stored immediately at ⫺80 °C. At the time of microdissection,

slides containing frozen sections were immediately fixed in 700 mL/L ethanol for 30 s and stained with hematoxylin and eosin, followed by 5-s dehydration steps in 700 mL/L, 950 mL/L, and undiluted ethanol and a final 5-min dehydration step in xylene. Once air-dried, the sections were laser microdissected with a PixCell I LCM system (Arcturus Engineering). Approximately 2500 morphologically normal breast stroma cells, normal breast epithelial cells, malignant invasive breast carcinoma cells, and malignant metastatic (to an axillary lymph node) breast carcinoma cells were laser captured, according to the standard protocol of Emmert-Buck et al. (1 ). Each population was estimated to be ⬎98% homogeneous as determined by microscopic visualization of the captured cells. After LCM, the EVA transfer film was gently peeled from the CapSureTM (Arcturus Engineering) and bisected; one-half of the film was attached to a standard Bruker mass spectrometer stainless steel target (Bruker Daltonics), with the cell-adherent surface up. Common rubber cement worked well and produced no detectable peaks above 200 Da. Approximately 1 ␮L of MALDI-ToF matrix solution, consisting of saturated sinapinic acid in 300 mL/L acetonitrile containing 1 mL/L trifluoroacetic acid, was uniformly distributed over the transfer film. Acetonitrile concentrations as high as 500 mL/L yielded similar spectra. An alternative matrix that has worked well is 10 g/L 2,5-dihydroxybenzoic acid in 300 –500 mL/L acetonitrile containing 1 mL/L trifluoroacetic acid. MALDI-ToF spectra were collected in positive-ion mode, using a Bruker Reflex III or Biflex III MALDI-ToF mass spectrometer (Bruker Daltonics) in linear mode. A deflecting voltage was applied to filter out peaks ⬍2000 m/z. A spectrum consisted of the sum of 10 subspectra from different locations on the film, each collected with 50 laser shots. Raw spectra were smoothed using the ninepoint Savitzky-Golay algorithm and baseline-subtracted using Bruker data reduction software. As a challenge in selecting and analyzing pure cell populations, normal breast stroma, normal ductal epithelium, ductal carcinoma in situ, and invasive ductal carcinoma were microdissected from a single frozen section of human breast (see Fig. 1). Distinct spectra were obtained from 1250 cells from each of the four cell types. The stromal cells revealed several prominent peaks in the 4500 –7000 Da range, which were attenuated or absent in the spectra from cells of epithelial derivation. A series of high-mass peaks from 45 to 60 kDa distinguished the invasive carcinoma spectrum from that of normal epitheFig. 1. Four cell populations microdissected from a single frozen section of human breast: normal stromal cells, normal epithelial cells, ductal carcinoma in situ, and invasive ductal carcinoma (rows 1– 4, respectively). Photomicrographs illustrate the tissue before LCM as well as the captured cells adherent to the transfer films (columns 1 and 2, respectively). In addition, a population of ductal carcinoma cells metastatic to a lymph node from the same patient were microdissected from a separate frozen section (row 5). Here column 2 illustrates the lymph node cells remaining in the frozen section after LCM. The mass spectra acquired directly from these transfer films are shown for mass/ charge (m/z) of 3500 –18 000 and 20 000 –70 000 (columns 3 and 4, respectively).


1515

1516


lium and from the stromal spectrum as well. The carcinoma in situ spectrum had features intermediate between the normal epithelium and the invasive tumor. Several common features were also evident in each spectrum. A frozen section of ductal carcinoma metastatic to a lymph node was also available from the same patient. As evidence for the reproducibility of the method, the spectra derived from the invasive and metastatic ductal carcinoma were virtually identical (see Fig. 1). Are these putative markers artifacts of LCM? The infrared-induced fusion causes local heating to 90 °C (7 ), but subsequent analysis with 2D-PAGE has revealed no noticeable impact on the concentrations of expressed proteins or their integrity (5 ). Extended case series are required to demonstrate whether these findings are consistent among different patients. Capturing cells from one frozen section requires ⬍30 min, and proceeding from transfer film to 20 –50 spectral peaks is equally fast. This compares favorably with the many hours of laser-capture time and days of electrophoresis needed to visualize several hundred spots on 2D-PAGE from LCM (5, 6 ). Moreover, the mass resolution of the 5- to 10-kDa MALDI-ToF peaks obtained from the LCM cells was 1–5%, which is better than the than the typical 5–10% resolution obtained with a 2D-PAGE system. Refinements in matrix deposition, biomolecule extraction, and desalting techniques may improve signal quality. Although the mass/charge ratio (m/z) of a mass speak is insufficient to unambiguously establish the identity of any given marker, we show that spectral patterns are reproducible to the cell type and provide a credible indicator of lineage. Specific marker identification through PAGE and peptide mass fingerprinting remains an option but is not a diagnostic prerequisite. These spectral markers are not necessarily etiologically related to the disease process but may reflect generalized changes in cell physiology. This work was funded in part by NIDDK SBIR Grant 1R43 DK54118-01. D.E.P-T. was funded in part by NIH:NCI Training Grant T32-CA09216. We thank Paul Kowalski of Bruker Daltonics for access to the mass spectrometers. References 1. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, et al. Laser capture microdissection. Science 1996;274:998 –1001. 2. Suarez-Quian CA, Goldstein SR, Pohida T, Smith PD, Peterson JI, Wellner E, et al. Laser capture microdissection of single cells from complex tissues. Biotechniques 1999;26:328 –35. 3. Sgroi DC, Teng S, Robinson G, LeVangie R, Hudson JR Jr, Elkahloun AG. In vivo gene expression profile analysis of human breast cancer progression. Cancer Res 1999;59:5656 – 61. 4. Anderson L, Seilhamer J. A comparison of selected mRNA and protein abundances in human liver. Electrophoresis 1997;18:533–7. 5. Banks RE, Dunn MJ, Forbes MA, Stanley A, Pappin D, Naven T, et al. The potential use of laser capture microdissection to selectively obtain distinct populations of cells for proteomic analysis. Electrophoresis 1999;20:689 – 700. 6. Emmert-Buck MR, Gillespie JW, Paweletz CP, Ornstein DK, Basrur V, Appella E, et al. An approach to proteomic analysis of human tumors. Mol Carcinog 2000;27:158 – 65. 7. Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S, Liotta LA. Laser capture microdissection: molecular analysis of tissue. Science 1997;278:1481–3.

A Specific Artificial Antibody toward Mycophenolic Acid Prepared by Molecular Imprinting, Pierre Morissette, Martin Beaulieu, and Bernard Vinet* (De´partement de Biochimie, Centre Hospitalier de l’Universite´ de Montreál, Montreál, Que´bec, Canada H2L 4M1; * address correspondence to this author at: De´partement de Biochimie, CHUM Hoˆpital Notre-Dame, 1560 Sherbrooke, Est Montreál, Que´bec, Canada H2L 4M1; fax 514-8964651, e-mail [email protected]) Molecular imprinting is the preparation of polymeric materials with specific binding sites for a molecule. A monomer is allowed to interact with the molecule of interest (template) to create low-energy interactions. Polymerization is then induced with a bridging agent and heat or ultraviolet irradiation. During the process of polymerization, the molecule of interest is entrapped within the polymer, which finally can be crushed, sieved, and washed with highly polar solvents to remove the template molecule. The imprint of the template is maintained in the rigid polymer, which is usually made of plastic materials such as acrylics and styrenes. The molecular imprint contains many small crypts with shapes complementary to the interest molecule and which are stabilized by chemical groups oriented during the polymerization process in the presence of the substrate. The imprinted polymer can bind the original molecule with high specificity similar to the specificities observed in enzyme-substrate or antigen-antibody interactions. The high specificity of the binding led to the concept of artificial antibodies (1 ). The use of artificial antibodies in competitive assays is the most promising application of molecular imprints in clinical chemistry. It has been shown that they can replace monoclonal antibodies from animal origin in certain immunologic assays such as those for cortisol, theophylline, diazepam, morphine, s-propanolol, and methylglucoside (1 ). Artificial antibodies show numerous advantages compared with natural antibodies: Ease of production Artificial antibodies can be prepared without laboratory animals, and the preparation protocol is relatively simple. • Chemical stability Natural antibodies are labile, whereas artificial antibodies are very robust. They can be used with strong acids, bases, and organic solvents, and are temperature- and pressure-resistant. They can be preserved for many years without loss of their specificity. • Reusable The natural antibody-antigen bond is stable and almost irreversible. The binding to an artificial antibody is reversible, and the molecular imprint can be used hundred of times, especially in HPLC and competitive assays. • Organic solvent compatibility Natural antibodies are denatured in the presence of organic solvents. Artificial antibodies can be used in •


the presence of solvents such as chloroform, acetonitrile, and toluene. Mycophenolate mofetil (MMF) is an ester derivative of mycophenolic acid (MPA) and is approved as an immunosuppressant drug in renal transplant patients. The prodrug MMF is rapidly transformed in vivo to the active immunosuppressant MPA (2 ), which inhibits inosine monophosphate dehydrogenase 2. It thus suppresses the de novo synthesis of guanosine nucleotides especially in T and B lymphocytes and stops their proliferation (3 ). Clinical trials in renal transplant patients have shown that the addition of MMF to steroids and cyclosporine immunosuppressive regimens can reduce rejection episodes observed early after transplantation as well as the incidence of acute rejection. Although treatment with MMF has been shown to have a good safety profile, it may be useful to monitor MPA concentrations in serum (4 – 6 ). HPLC methods specific for MPA have been described (7 ). An immunoenzymatic assay (Emit) is also commercially available for the determination of MPA in serum (8 ). It is more automated and requires less technical work than the HPLC methods. The Emit assay yields results that are higher than those obtained with the HPLC methods. This is believed attributable to the cross-reactivity of MPA metabolites with the monoclonal antibodies. MPA is metabolized mainly to a glucuronide derivative (MPAG), which is believed inactive; however, the MPAG concentrations in sera of patients treated with MMF are an order of magnitude higher than those of MPA, and thus its potential for interference in the various analytical methods is high. Recently, two other metabolites of MPA have been identified (9 ). One of these metabolites, the acyl glucuronide of MPA, is believed to be responsible for the higher results observed with the Emit method. The Emit method has relatively high reagent costs, and the stability of reagents is limited. With the increasing popularity of the serum determination of MPA in transplant patients (10 ), there is a need for a fast, economical, and specific assay for the determination of MPA in serum. In this study, we prepared a molecular imprint of MPA by polymerization of methacrylic acid (MAA) in the presence of MPA. The polymer is easily produced, inexpensive, and highly specific. We set up the analytical conditions for binding MPA in an aqueous medium. The preparation and characterization of the binding of MPA to this polymer are described in this report. For this study, MPA was obtained from Sigma. Acetonitrile, methanol, and toluene were obtained from Fisher Scientific. MAA, ethylene glycol dimethacrylate, and 2,2⬘azobis(isobutyronitrile) were obtained from Aldrich. MPAG was isolated in our laboratory from the urine of transplant patients. The details of the polymer synthesis have been described by Muldoon and Stanker (11 ). Briefly, the molar ratios of print molecule (MPA), functional monomer (MAA), and cross-linker (ethylene glycol dimethacrylate) were 1:4:20. To initiate the polymerization, 4.125 mg of 2,2⬘-azobis(isobutyronitrile) was added. The mixture was

1517

deoxygenated continuously with a stream of nitrogen during polymerization at 60 °C for 2 h. The polymer was then crushed and ground repeatedly. The particles were free from MPA after soaking in 100 mL of methanol for 12 h. After a final drying at 60 °C, the polymer was ready for use. The polymerization produced a rigid solid that was ground into small particles averaging 25 ␮m in diameter. An important characteristic of the polymer is its robustness, which gives it the ability to retain its specific binding sites even in harsh conditions. This property also gives the polymer the advantage of being stored at ambient temperatures for many months without loss of its recognition capabilities (12 ). We determined that by passing a constant stream of nitrogen through the mixture during the polymerization, we reduced the time of polymerization to 2 h. Although it has been shown that the number of binding sites and the affinity at which the polymer binds to the print molecule can be affected by the conditions of polymerization (13 ), the deoxygenation process did not affect the affinity of binding or the number of binding sites of our polymer. Typically, the binding of MPA to the polymer is performed in a 0.01 mol/L phosphate buffer, pH 2.5, containing 5 mL/L methanol. This concentration of methanol was shown to be optimal. Most binding studies with molecular imprinted polymers have been done in organic solvents (1 ). The advantage of working in an aqueous binding solvent is that it is closer to the natural environment of human plasma. This characteristic could be useful in the development of an heterogeneous binding assay using the molecularly imprinted polymer. The binding of MPA to the polymer is affected by the pH of the binding solvent (2 ). As shown in Fig. 1A, the imprinted polymer gains affinity for MPA with a decrease of pH. In an acidic solvent, the carboxyl group of MPA should predominately be in its anionic form. Thus, we can make the hypothesis that the binding of MPA occurs by interaction of its carboxyl group with the polymer through hydrophobic interactions. To study the binding specificity of the polymer, a comparison of the binding to imprinted and nonimprinted polymers was performed (Fig. 1B). We defined specific binding as the difference between binding to imprinted and nonimprinted polymers. It was determined that the artificial antibody could bind MPA with high specificity up to an average concentration of 125 mg/L. A Scatchard analysis was used to estimate the strength of binding of the print molecule (Fig. 1C). This analysis clearly gave a nonlinear plot, which is typical of imprinted polymers and reflects the heterogeneity of the binding sites present. A two-site model was used to calculate the binding strength (KD) and the site populations (Bmax) of the polymer. The molecularly imprinted polymer gave KDs of 0.88 and 100 ␮mol/L and Bmax values of 15.0 and 233 ␮mol/g, respectively, which is comparable to other imprinted polymers (1 ). These results show the presence of binding sites with high affinity and sites with low affinity. The low-affinity sites are

1518


present in much greater numbers, even in the polymer prepared without MPA, and represent nonspecific binding to the polymer. Interference studies were performed with several drugs that may be co-administered with MPA. The drugs tested included acetylsalicylic acid, butobarbital, caffeine, cyclobarbital, diazepam, phenobarbital, phenytoin, salicylic acid, theophylline, and thiobarbital. The metabolite MPAG was also tested at a final concentration of 1500 ␮mol/L. All of the compounds tested did not interfere with the binding of MPA to the polymer. It was determined in our laboratory (data not shown) that plasma samples from renal transplant patients treated with MPA show a MPAG:MPA molar ratio of 20.4 ⫾ 21.2 (n ⫽ 53). In our experiment, the glucuronide was tested to a ratio of 100:1 (MPAG:MPA) and was shown not to interfere. In summary, we prepared a molecularly imprinted polymer that can specifically bind MPA in an acidic aqueous medium. The polymer contains two distinct binding sites. The high affinity sites are responsible for the specific binding of MPA. The imprinted polymer has a hydrophobic interface, which is believed to interact with the protonated carboxyl group of MPA in an acidic medium. Current work is being directed toward the development of a heterogeneous binding assay for MPA in plasma with the molecularly imprinted polymers. References

Fig. 1. Analysis of MPA binding by an anti-MPA polymer. Binding to the anti-MPA polymer was determined in a 0.01 mol/L phosphate buffer containing 5 mL/L methanol. Polymer (125 mg/L) was added to the solution and incubated for 30 min at room temperature with agitation (Eberbach). The polymer was centrifuged, and MPA was measured in the supernatant by HPLC. The HPLC procedure is based on the method of Shipkova et al. (7 ). The equipment consisted of a Hewlett Packard series 1100 pump, an automatic injector, an ultraviolet detector, a Supelco LC-318 column (25 cm ⫻ 4.6 mm; 5 ␮m particles), and a Hewlett Packard Kayak XA, Pentium II interface. The mobile phase consisted of 0.01 mol/L phosphoric acid buffer, pH 5.0 (adjusted with NaOH), containing 375 mL/L acetonitrile. The flow rate was 1.0 mL/min, and MPA was detected at 215 nm with a retention time of 7 min. The temperature was kept at 40 °C. (A), pH of the binding medium was adjusted from 2.5 to 7.5 with NaOH. (B), 50 to 1000 mg/L of polymer was added to the binding medium, which contained 15 ␮mol/L MPA. The dashed line and arrow indicate the region of specific binding. (C), Scatchard analysis. Dissociation constants (KD) were determined as 0.88 and 100 ␮mol/L, corresponding to high- and low-affinity sites, respectively. Site populations were 15 and 233 ␮mol/g polymer, respectively.

1. Ansell RJ, Ramstro¨m O, Mosbach K. Towards artificial antibodies prepared by molecular imprinting. Clin Chem 1996;42:1506 –12. 2. Bardsley-Elliot A, Noble R, Foster H. Mycophenolate mofetil. A review of its use in the management of solid organ transplantation. Biodrugs 1999;12: 363– 410. 3. Ransom JT. Mechanism of action of mycophenolate mofetil [Review]. Ther Drug Monit 1995;17:681– 4. 4. Halloran P, Mathew T, Tomlanovich S, Groth C, Hooftman L, Barker C. Mycophenolate mofetil in renal allograft recipients: a pooled efficacy analysis of three randomized, double blind, clinical studies in prevention of rejection. Transplantation 1997;63:39 – 47. 5. Groth CG. The European experience with mycophenolate mofetil. European Mycophenolate Mofetil Cooperative Study Group. Transplant Proc 1996; 28(Suppl 1):30 –3. 6. Shaw LM, Sollinger HW, Halloran P, Morris RE, Yatscoff RW, Ransom J, et al. Mycophenolate mofetil: a report of the Consensus Panel. Ther Drug Monit 1995;17:690 –9. 7. Shipkova M, Niedmann PD, Armstrong VW, Schu¨tz E, Wieland E, Shaw LM, et al. Simultaneous determination of mycophenolic acid and its glucuronide in human plasma using a simple high performance liquid chromatography procedure. Clin Chem 1998;44:1481– 8. 8. Yeung JS, Wang W, Chan L. Determination of mycophenolic acid level: comparison of high-performance liquid chromatography with homogeneous enzyme-immunoassay. Transplant Proc 1999;31:1214 –5. 9. Schu¨tz E, Shipkova M, Armstrong VW, Wieland E, Oellerich M. Identification of a pharmacologically active metabolite of mycophenolic acid in plasma of transplant recipients treated with mycophenolate mofetil. Clin Chem 1999; 45:419 –22. 10. Van Gelder T, Hillbrands LB, Vanrenterghem Y, Weimar W, de Fitjter JW, Squifflet JP, et al. A randomized double-blind, multicenter plasma concentration controlled study of the safety and efficacy of oral mycophenolate mofetil for the prevention of acute rejection after kidney transplantation. Transplantation 1999;68:261– 6. 11. Muldoon MT, Stanker LH. Molecularly imprinted solid phase extraction of atrazine from beef liver extracts. Anal Chem 1997;69:803– 8. 12. Vlatakis G, Andersson LI, Mu¨ller, Mosbach K. Drug assay using antibody mimics made by molecular imprinting [Letter]. Nature 1993;361:645–7. 13. Chen Y, Kele M, Sajonz P, Sellergren B, Guiochon G. Influence of thermal annealing on the thermodynamic and mass-transfer kinetic properties of Dand L-phenylalanine anilide on imprinted polymeric stationary phases. Anal Chem 1999;71:928 –38.


Hepatocyte Asialoglycoprotein Receptor Assay Using Stable Isotopes and Neutron Activation Analysis, Ernest V. Groman* and Christopher P. Reinhardt (BioPAL, 10 New Bond St., Worcester, MA 01606; * author for correspondence: fax 508-852-8118, e-mail [email protected]) Since the initial studies of receptor-mediated endocytosis on LDL (1, 2 ) and asialoglycoprotein (ASGP) (3, 4 ), many internalized ligands and their receptors have been identified and characterized (5 ). It was soon recognized that these receptors could be exploited for pharmaceutical and medical imaging applications (6 ). In the case of the ASGP receptor (ASGP-R), delivery of diagnostic or therapeutic agents to hepatocytes was achieved by attachment of the agent to a variety of macromolecules possessing the terminal galactose signal necessary to bind the ASGP-R (6 ). These moieties include radioactive isotopes (7 ), drugs (8 ), and DNA (9 ). Considering the effort devoted to developing drug delivery systems utilizing the ASGP-R, surprisingly little effort has been devoted to developing simple, easy-to-use assays to measure this receptor activity in vivo. Much of the technology associated with receptor-based diagnostic agents has arisen in fields associated with magnetic resonance and single photon emission computed tomography imaging. One class of magnetic resonance compounds includes iron chelates, which appear to be absorbed by hepatocytes by an unknown biochemical mechanism (10 ). Saini et al. (11 ) and others have demonstrated the promise of ferrite particles as a magnetic resonance contrast agent for liver imaging. These commercially available micron-sized particles are taken up by the Kupffer cells of liver and do not measure ASGP-R activity. Sato et al. (12 ) reported on the ability of the ASGP-R to take up galactose-terminated, protein-coated ferrite particles in vivo. These particles are taken up preferentially by hepatocytes (70%) and largely avoid uptake by the reticular endothelial system. For a crystal-based material to bind the ASGP-R of hepatocytes, it must penetrate pores of the 100-nm fenestrae characteristic of endothelial cells that lie between the hepatocyte and the circulatory compartment (13 ). Particles smaller than 100 nm are needed to serve as hepatocyte-directed diagnostic agents. Josephson et al. (14 ) developed the findings of Sato et al. (12 ) by synthesizing a superparamagnetic iron oxide colloid (⬃50 nm) coated with the polygalactosylated polysaccharide arabinogalactan. This colloid cleared rapidly from the vascular system; its clearance was inhibited by asialofetuin but not by fetuin, indicative of ASGP-R interaction. More than 90% of the particles were taken up in hepatocytes. The most thoroughly studied ASGP-R-directed agents are those based on galactosylated albumin labeled with radioactive technetium (99mTc) (15–19 ). Although the use of Tc-galactosylated albumin offers a unique and elegant approach to the study of liver biology, its application in the research laboratory and the clinic is limited. Many investigators have limited access to 99Tc generators and

1519

would prefer to avoid the use of radioactivity. In the clinic, radioactivity presents a potential hazard to the patient and medical personnel. 99Tc scintigraphy also relies on expensive instrumentation. Stable-isotope labeling offers a cost-effective technology to deal with measuring the ASGP-R activity and biodistribution in vivo (20 ). This report explores the feasibility of using stable-isotope labeling of ASGP-R-directed agents for measuring ASGP-R activity in vivo. Two examples of such labels are reported—arabinogalactan labeled with samarium and gold colloids coated with lactosylated albumin—and their specificity for the ASGP-R and sensitivity are explored. Unless stated otherwise, all reagents were obtained from Sigma-Aldrich. The diethylenetriaminepentaacetic acid (DTPA) conjugate of arabinogalactan was prepared as described previously (21 ). The final product contained 4.7 moles of DTPA per mole of arabinogalactan. The samarium chelate of arabinogalactan (Sm-arabinogalactan) was prepared from a solution of samarium chloride (0.3 mmol), adjusted to pH 2, that was added to DTPAarabinogalactan (0.1 mmol), also adjusted to pH 2, in saline. The solution was adjusted to pH 5, mixed for 30 min, and adjusted to pH 7.4. PolyGalactoseGoldTM colloidal gold (20 nm) and lactosylated albumin (1 g/L) were incubated for 2 h at room temperature in 0.01 mol/L bicarbonate buffer, pH 8.5. Colloidal gold was isolated by centrifugation and suspended in phosphate-buffered saline. Albumin-coated gold colloid was prepared by substituting bovine albumin for lactosylated albumin. The coating and stability of each colloid was evidenced by the suspension remaining red in the presence of saline (a light-scattering color consistent with a 20-nm particle size). In contrast, uncoated colloidal gold suspended in saline turned purple and a precipitate formed. The stability of Sm-arabinogalactan was evaluated in rat serum by incubation for 60 min at 37 °C followed by analysis using size-exclusion chromatography (Sephadex G-50). All of the samarium was retained in the fractions corresponding to arabinogalactan. Rats (300 –375 g) used in all studies were male hooded/ BBZDR (BioMedical Research Models). For each set of experiments, three rats were anesthetized with Nembutal. The carotid artery and jugular vein were exposed, and the animal received via the jugular vein an injection of Sm-arabinogalactan (4 mg/kg), PolyGalactoseGold (50 ␮g/kg), or albumin-coated gold (50 ␮g/kg). A second set of animals received injections of asialofetuin (100 mg/kg) followed by Sm-arabinogalactan or PolyGalactoseGold. At various times, blood samples were removed from the carotid artery. After 60 min, the animals were sacrificed, and ⬃1-g amounts of various tissues were recovered for quantification of samarium or gold. Tissues were blotted to remove blood and clots. No additional processing of tissue such as perfusion was necessary to perform neutron activation analysis. The samarium and gold content in all samples was quantified by neutron activation by BioPAL (Worcester, MA; www.biopal.com). Samples were placed in 2-mL

1520


polypropylene tubes free of trace element contaminants and dried at 70 °C for a minimum of 12 h. An internal standard of tungsten, to correct for variations in neutron flux, was added to each sample. Samples were activated for 15 min in a neutron field created by a 2-MW nuclear reactor. Short-lived activated products, principally resulting from sodium and chloride, were allowed to decay for 2 days, and the remaining radioactivity from activated samarium or gold was counted using a high-resolution gamma spectrometer. The fate of Sm-arabinogalactan after intravenous injection into rats was examined by obtaining the biodistribution of the conjugate 60 min after injection. Biodistribution of the conjugate was inferred by the presence of samarium in tissues after the injection of Sm-arabinogalactan. Smarabinogalactan was present in the liver (51% of injected dose) and urine (35% of injected dose). Less than 1% of the injected dose was present in spleen or blood. When asialofetuin, a ligand for the ASGP-R of hepatocytes, was co-injected with Sm-arabinogalactan, hepatic uptake decreased to 3%, whereas the urine fraction increased to 52%. Asialofetuin also increased blood concentrations of Sm-arabinogalactan from 1% to 17%, which increased urinary elimination. Biodistribution data and inhibition of hepatic uptake by asialofetuin indicated that the hepatic clearance of Sm-arabinogalactan was mediated by the ASGP-R. Renal clearance of Sm-arabinogalactan is consistent with the conjugate’s low molecular mass (⬃40 kDa) and compact globular shape allowing excretion by glomerular filtration. The fate of PolyGalactoseGold after intravenous injection into rats and mice was examined by obtaining the biodistribution of the conjugate 60 min after injection. Biodistribution of the conjugate was inferred by the presence of gold in tissues after the injection of PolyGalactoseGold. PolyGalactoseGold was present in the liver (⬃100% of injected dose) but not in the spleen, adrenals, lung, kidney, heart, marrow, brain, muscle, or urine. When asialofetuin, a ligand for the ASGP-R of hepatocytes, was co-injected with PolyGalactoseGold, blood clearance was substantially slower (Fig. 1). When albumin-coated colloidal gold was substituted for PolyGalactoseGold, blood clearance for albumin-coated gold was also substantially slower than that seen with PolyGalactoseGold, demonstrating the specific role of galactose for clearing PolyGalactoseGold. These results are consistent with reports in the literature that lactosylated albumin is a ligand of the ASGP-R and indicate that the hepatic clearance of PolyGalactoseGold was mediated by the ASGP-R. The lack of renal clearance of PolyGalactoseGold is consistent with its size (20 nm), which corresponds to a molecule approximately the size of IgM. With the activation procedure described above, 1 ng of gold corresponding to ⬃800 dpm could be easily distinguished from background (⬃5 dpm). We hypothesize that a compound consisting of a stable isotope bound to a receptor-directed reagent can be assayed with neutron activation technology to study and evaluate hepatic function. As a first step we characterized

Fig. 1. Kinetics of removal of 20-nm gold colloid coated with lactosylated albumin or albumin from blood. Rats received injections of gold colloid coated with lactosylated albumin in the absence (f) or presence of asialofetuin (F) or with gold colloid coated with albumin (Œ). Blood samples were drawn at the indicated times, and the percentage of gold remaining in the blood (normalized for gold concentration at time zero) was determined using neutron activation analysis. Values presented represent the mean value obtained with three rats. Mean (SD) values for samples were calculated using the percentages of gold remaining in the blood. Standard deviations are therefore not calculated for values of 100%. Mean (SD) values at 5, 10, and 15 min were as follows: lactosylated albumin-coated gold colloid, 33% (6.1%) and 8% (1.0%); lactosylated albumin-coated gold colloid in the presence of asialofetuin, 95% (1.7%), 43% (3.6%), and 9% (2.0%); albumincoated gold colloid, 100%, 75% (5.2%), and 42% (6.2%).

two ASGP-R-directed reagents labeled with a stable isotope of either samarium or gold, each possessing excellent properties for neutron activation. This approach presents an alternative to existing methods used to measure ASGP-R activity using radioactive, magnetic, or dyelabeled agents. References 1. Brown MS, Goldstein JL. Receptor-mediated endocytosis: insights from the lipoprotein receptor system. Proc Natl Acad Sci U S A 1979;76:3330 –7. 2. Brown MS, Goldstein JL. A receptor-mediated pathway for cholesterol homeostasis. Science 1986;232:34 – 47. 3. Ashwell G, Harford J. Carbohydrate-specific receptors of the liver. Annu Rev Biochem 1982;51:531–54. 4. Schwartz AL, Geuze HJ, Strous GJ. The asialoglycoprotein receptor: intracellular fate of ligand and receptor. In: Steer CJ, Hanover JA, eds. Intracellular trafficking of proteins. Cambridge: Cambridge University Press, 1991:279 – 301. 5. Pastan IH, Willingham MC. Receptor-mediated endocytosis of hormones in cultured cells. Annu Rev Physiol 1981;43:239 –50. 6. Meijer DKF, Molema G, Jansen RW, Mollenaar JF. Design of cell-specific drug targeting preparations for the liver: where cell biology and medicinal chemistry meet. In: Claassen V, ed. Trends in drug research, Vol. 13. Amsterdam: Elsevier, 1990:303–32. 7. Stadalnik RC, Vera DR, Woodle ES, Trudeau WL, Porter BA, Ward RE, et al. Technetium-99 m NGA functional hepatic imaging: preliminary clinical experience. J Nucl Med 1985;26:1233– 42. 8. Wu GY, Wu CH, Rubin MI. Acetaminophen hepatotoxicity and targeted rescue: a model for specific chemotherapy of hepatocellular carcinoma. Hepatology 1985;5:709 –13. 9. Wu GY, Wu CH. Receptor-mediated in vitro gene transformation by a soluble DNA carrier system [published erratum appears in J Biol Chem 1988;263: 588]. J Biol Chem 1987;262:4429 –32. 10. Lauffer RB, Greif WL, Stark DD, Vincent AC, Saini S, Wedeen VJ, Brady TJ. Iron-EHPG as an hepatobiliary MR contrast agent: initial imaging and biodistribution studies. J Comput Assist Tomogr 1985;9:431– 8. 11. Saini S, Stark DD, Hahn PF, Wittenberg J, Brady TJ, Ferrucci JT Jr. Ferrite particles: a superparamagnetic MR contrast agent for the reticuloendothelial system. Radiology 1987;162:211– 6. 12. Sato SB, Sako Y, Yamashina S, Ohnishi S. A novel method for isolating specific endocytic vesicles using very fine ferrite particles coated with biological ligands and the high-gradient magnetic separation technique. J Biochem 1986;100:1481–92.


13. McCuskey RS. The hepatic microvascular system. In: Arias IM, Boyer JL, Fausto M., eds. The liver: biology and pathobiology, 3rd ed. New York: Raven Press, 1994:1089 –106. 14. Josephson L, Groman EV, Menz E, Lewis JM, Bengele H. A functionalized superparamagnetic iron oxide colloid as a receptor directed MR contrast agent. Magn Reson Imaging 1990;8:637– 46. 15. Hiraguchi E. [Evaluation of hepatic blood flow using 99 mTc-GSA in rats with hepatic blood flow manipulation]. Kaku Igaku 1995;32:453– 63. 16. Fujioka H, Kawashita Y, Kamohara Y, Yamashita A, Mizoe A, Yamaguchi J, et al. Utility of technetium-99m-labeled-galactosyl human serum albumin scintigraphy for estimating the hepatic functional reserve. J Clin Gastroenterol 1999;28:329 –33. 17. Ha-Kawa SK, Tanaka Y. A quantitative model of technetium-99m-DTPAgalactosyl-HSA for the assessment of hepatic blood flow and hepatic binding receptor. J Nucl Med 1991;32:2233– 40. 18. Kaltwasser JP, Gottschalk R, Schalk KP, Hartl W. Non-invasive quantitation of liver iron-overload by magnetic resonance imaging. Br J Haematol 1990;74:360 –3. 19. Kokudo N, Vera DR, Koizumi M, Seki M, Sato T, Stadalnik RC, Takahashi T. Recovery of hepatic asialoglycoprotein receptors after major hepatic resection. J Nucl Med 1999;40:137– 41. 20. BioPALTM. BioPhysics Assay Laboratory, Inc. homepage. www.biopal.com. 21. Groman EV, Enriquez PM, Jung C, Josephson L. Arabinogalactan for hepatic drug delivery. Bioconjug Chem 1994;5:547–56.

Development of an Automated Quantitative Latex Immunoassay for Cardiac Troponin I in Serum, Judy Ash,* George Baxevanakis, Lela Bilandzic, Howard Shin, and Lilly Kadijevic (Spectral Diagnostics Inc., 135-2 The West Mall, Toronto, Ontario, M9C1C2 Canada; * author for correspondence: fax 416-626-3651, e-mail jash@spectral diagnostics.com) Currently, the measurement of troponin I (TnI) can only be accomplished through the use of heterogeneous assays on closed-system automated analyzers. The development of this new and innovative latex technology will allow the measurement of TnI on a variety of turbidimetry-based open-system instruments, greatly enhancing clinical applicability of this test. Determining the presence of TnI in the serum of patients is an important aid in the diagnosis of myocardial infarction. An advantage of TnI is its improved specificity for myocardial damage compared with creatine kinase-MB (1 ). In addition, there is strong evidence that future utilization of TnI will be for risk stratification to assist in the decision process for therapeutic intervention with glycoprotein II/IIIa inhibitors or low-molecular weight heparin (2, 3 ). In fact, the GUSTO trial, which should be completed soon, included TnI as one of the cardiac markers to be considered for risk stratification. The cardiac troponin complex is part of the contractile apparatus of the thin filament in striated muscle and consists of subunits C, T, and I. Different isoforms of TnI exist in the skeletal and cardiac muscles (fast skeletal, slow skeletal, and cardiac TnI). The distinct structural heterogeneity between these isoforms allows production of specific antibodies (4 ), which can be utilized by the latex assay to detect serum TnI in clinical conditions that involve myocardial damage. After acute myocardial infarction, damaged myocytes lose these proteins, and various forms of troponin (complexed, free, or fragments)

1521

appear in the blood (5 ). TnI concentrations become abnormal 4 – 8 h after the onset of chest pain, peak at 12–16 h, and remain increased for 5–9 days following an infarction. We have developed an automated latex immunoassay for the detection of TnI in serum with excellent sensitivity, precision, and stability. The assay utilizes two monoclonal antibodies and one polyclonal antibody to TnI. Antibodies were selected using Biacore analysis, epitope mapping, and pairing studies. Each antibody was separately covalently bound to 200-nm supercarboxyl polystyrene latex particles using 1-ethyl-3(3-dimethylaminopropyl)carbodiimide hydrochloride in a two-step coupling process. The various coupled antibodies were then combined into one solution in a 1:1:1 ratio. Unbound antibody was removed with a microgon system, which utilizes tangential flow. The beads were then sonicated to redistribute the particles. When TnI is present in the serum, the addition of a reaction buffer causes adjacent beads to cross-link, increasing the turbidity of the solution. The results are calculated in ⬍9 min by the Cobas Mira from a stored calibration curve generated with recombinant human cardiac TnI calibrators using a logit/log4 calculation mode. Reaction buffer (140 ␮L), water (10 ␮L), and latex (45 ␮L of a 1.25 g/L solution) are mixed and incubated at 37 °C for 125 s. Sample (30 ␮L) and water (5 ␮L) are then added, and the absorbance change of the reaction mixture is measured from ⬃1 min 25 s to ⬃8 min after the addition of sample. The rate of increase in turbidity, measured at 600 nm, is proportional to the concentration of TnI present in the serum. The assay range was 0 –25 ␮g/L. The detection limit of the assay was 0.3 ␮g/L. This value was calculated as 2 SD above the mean of 21 replicates of the zero calibrator. Using NCCLS guidelines, intra- and interassay imprecision was determined using both a high and low control over a course of 18 days with two analytical runs per day. The intraassay imprecision (CV) for the 5 ␮g/L control was 6.8%, and its interassay CV was 7.5%. The 15 ␮g/L control gave an intraassay CV of 1.6% and interassay CV of 4.2%. Stability studies were carried out on the latex particles, the assay buffer, and the calibrators by evaluating the performance of high and low controls with the test reagent at scheduled intervals with approved lots of reagents. The test reagent was considered unstable when the control results demonstrated a downward or upward shift of ⬎10% from the day 0 value for the control and this variation was consistent on 2 or more consecutive days. Calibrators were found to be stable for 2 months when stored at room temperature. Latex reagent was stable for 21 days at 37 °C, and the reaction buffer was stable for 56 days at room temperature. The assay was evaluated to determine whether antigenexcess phenomenon (prozone effect) would cause false negatives. A false negative occurs when a sample that has a high troponin concentration produces a negative result instead of the high value. This phenomenon was evaluated by preparing a high concentration of recombinant TnI in storage buffer, making serial dilutions of this

1522


material, and testing each sample. TnI concentrations up to 450 ␮g/L yielded a signal that was above the limit defined by the measuring range, and concentrations up to 4000 ␮g/L did not produce a false negative. Rheumatoid factor added to a final concentration of 160 kIU/L to control and serum samples positive for TnI did not interfere with the assay, nor did hemoglobin (up to 5 g/L) or bilirubin (up to 400 mg/L). In addition to the baseline concentrations of triglycerides and albumin already present in the serum pools, additional amounts of triglycerides (up to 250 mg/L) and albumin (up to 300 mg/L) when added to the pools did not interfere with the assay. A comparison of assay values obtained from testing serum samples using the latex TnI test and the Stratus TnI immunoassay gave the following linear regression data: Latex TnI ⫽ 0.18 (Stratus) ⫹ 0.0; n ⫽ 281, r ⫽ 0.87; range, 0 –37.2 ␮g/L (Fig. 1). At cutoffs for the diagnosis of acute myocardial infarction of 0.8 ␮g/L for the latex immunoassay as performed on the Cobas Mira and 1.5 ␮g/L for the Stratus TnI fluorometric enzyme immunoassay, the performance of the two assays was compared. There were 108 samples positive by the Stratus and 106 positive by the latex assay, which yielded a sensitivity of 98.1%. Of the 173 samples negative by the Stratus, 164 (95%) were negative by the latex assay. The overall agreement of 96% (270 of 280) demonstrates that the two products have equivalent performance for the determination of TnI in specimens from chest pain patients. This represents the first report of a rapid, fully automated homogeneous latex immunoassay for TnI with good sensitivity and precision at low concentrations, limited prozone effect, and reagent stability comparable to the existing heterogeneous immunoassays.

Fig. 1. Correlation of Cobas Mira TnI latex assay with Stratus enzyme immunoassay. Serum samples (n ⫽ 281) from chest pain patients and healthy adults were evaluated on both the Stratus and the Cobas Mira. The linear regression results were as follows: Latex ⫽ 0.18 (Stratus) ⫹ 0.0; n ⫽ 281; r ⫽ 0.87; range, 0.0 –37.2 ␮g/L.

References 1. Keffer JH. Myocardial markers of injury. Evolution and insights [Review]. Am J Clin Pathol 1996;105:305–20. 2. Antman EM, Tanasijevic MH. Cardiac-specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med 1996;335:1342–9. 3. Christenson RH, Duh SH, Newby K, Ohman EM, Califf RM, Granger CK, et al. Cardiac troponin T and I: relative values in short term risk stratification of patients with acute coronary syndromes. Clin Chem 1998;44:494 –501. 4. Larue C, Defacque-Lacquement H, Calzolari CA, Nguyen D, Pau B, New monoclonal antibodies as probes for human cardiac troponin I: epitopic analysis with synthetic peptides. Mol Immunol 1992;29:271– 8. 5. Wu A, Feng Y, Moore R, Apple F, McPherson P, Buechler K, Bodor G. Characterization of cardiac troponin subunit in serum after acute myocardial infraction and comparison of assays for troponin T and I. Clin Chem 1998;44:1198 –208.