Article pubs.acs.org/ac
Highly Sensitive Immunoassay for the Diagnosis of Acute Myocardial Infarction Using Silica Spheres Encapsulating a Quantum Dot Layer Hyojeong Han,†,∥ Jae-Chul Pyun,‡,∥ Hyein Yoo,† Hong Seog Seo,§ Byung Hwa Jung,† Young Sook Yoo,† Kyoungja Woo,*,† and Min-Jung Kang*,† †
Molecular Recognition Research Center, Korea Institute of Science and Technology (KIST), Seoul 136-791, Republic of Korea Department of Materials Science and Engineering, Yonsei University, Seoul 120-749, Republic of Korea § Cardiovascular Center, Korea University Guro Hospital, Korea University College of Medicine, Seoul 152-703, Republic of Korea ‡
S Supporting Information *
ABSTRACT: Commercial ELISA kits for substance P (SubP), which are helpful for the clinical diagnosis of acute myocardial infarction, are limited in efficacy because of low sensitivity. A highly sensitive immunoassay was developed using silica spheres encapsulating a quantum dot-layer (SQS) and labeling antibodies, on a Parylene A-modified plate. The high sensitivity was possible by taking advantage of the enhanced photoluminescence of the SQS and dense immobilization of SubP on a Parylene A-modified plate. Glutaraldehyde was used for cross-linking of SQS to the anti-SubP antibody and SubP to the Parylene A coating. The SQS-linked immunosorbent assay (SQSLISA) was optimized and validated. The dynamic range for the assay was 1−10000 pg/mL with a linear correlation factor of 0.9992 when the competitive SQSLISA was employed. The intra- and interday accuracies were 93−100% and 87−122%, respectively. The reproducibility was lower than 11%. The developed method was applied to clinical samples collected from healthy controls (n = 30) and acute myocardial infarction (n = 16) and it displayed a high correlation with the commercial ELISA kit, with a limit of detection that was 30-fold lower. Clinical sample analysis confirmed that SubP is a promising diagnostic marker for acute myocardial infarction. The SQSLISA is expected to be a practical and useful assay tool.
A
AMI, UA, SA, and healthy controls were screened using liquid chromatography and mass spectrometry.24 The discovered biomarker candidates were confirmed by commercial enzyme linked immunosorbent assay (ELISA). From this study, the regulatory peptides substance P (SubP) was discovered as a possible candidate to serve as an early diagnosis marker for AMI. Prior to this, the elevation of SubP in plasma has been reported in inflammation.25 Elevated concentrations of SubP has been observed in pancreatic cancer26−28 and lung cancer29 patients by other researchers. The role of SubP in heart disease was a new finding. SubP could serve as a novel biomarker for AMI. However, in the case of SubP, a majority of serum samples from healthy controls displayed concentrations lower than the limit of quantification (LOQ, 30 pg/mL). Therefore, a more sensitive detection method with a lower LOQ is necessary for the utilization of SubP as a reliable biomarker. A sensitive and accurate detection method of SubP could be utilized as an AMI diagnosis kit in the clinic. To develop a highly sensitive immunoassay method for molecules found at low levels (especially, SubP), a highly sensitive signaling element is required. A semiconductor nanocrystal, quantum dot (QD), is a representative inorganic
cute myocardial infarction (AMI) and heart failure (HF) are leading causes of death in Korea and worldwide.1−3 HF-related mortality was unchanged between 1995 and 2009 in the USA and Europe.2 The predominant type of cardiovascular disease varies between Asian and Western populations.4 Cerebrovascular disease is the most common in Asian populations,3−6 while coronary artery disease occurs more frequently in Western populations.1,2,4 In particular, HF and AMI are the third highest cause of mortality in Korean men and women in 2008.3 The symptom of AMI is very serious and the patients have a burden of hard life. Accurate diagnosis of AMI is critical. Until now, several diagnosis markers for AMI have been developed.7−22 Among the identified biomarkers such as Creactive protein (CRP), MB isoenzyme of creatin kinase (CKMB), myoglobin, and cardiac troponins, cardiac troponin I (cTn I) is regarded as the most valuable cardiac-specific marker for myocardial damage.13−20 The sensitivity and specificity were 80−100 and 67−98%.13,14 However, even in the case of cTn I, the delayed increase of its circulating level (3 h after seizure)17 is a limitation. Some peptide biomarkers such as N-terminal pro-B-type natriuretic peptide and C-type natriuretic peptide have been suggested as AMI diagnostic markers;21−23 however, they have not been validated adequately for clinical diagnosis yet. In the previous study, in order to discover a sensitive biomarker for the early stages of heart diseases like unstable angina (UA), stable angina (SA), serum from patients with © XXXX American Chemical Society
Received: June 5, 2014 Accepted: September 26, 2014
A
dx.doi.org/10.1021/ac502412x | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 1. (a) Schematic diagram of Parylene-A and SubP (P) coating process of 96-well polystyrene plates, (b) labeling of the anti-SubP antibody with SQS using glutaraldehyde as a cross-linker, and (c) experimental flow for competitive SQSLISA.
sphere, intermittent sonication of 5-fold diluted solution than the published one gave less aggregated product and, thus, highly enhanced PL. The detailed procedures are described in the Supporting Information. The synthesized SQS was characterized by transmission electron microscope (TEM) image analysis using a CM30 with 200 kV (Philips, Eindhoven, The Netherlands) and spectroscopic analysis using a fluorescence spectrophotometer F-4500 (Hitachi, Tokyo, Japan) with a QD-based concentration of 5 × 10−8 M in all the traces. To calculate the concentrations of QDs and silica particles, we assumed 100% yield at each step throughout the synthesis since little PL was observed in the supernatant following centrifugation of each product. Actually, the true concentration of the involved QD in each solution should be trivially lower than the previous one in the later steps. Contrary to the expected loss of QD, increased PL was observed with the dispersed SQS. For NH2 functionalization, 14 mL of the SQS solution was diluted to 20 mL in a round-bottom flask. DW (0.6 mL) and NH4OH (0.4 mL) were added and stirred for 10 min. APTMS (0.1 mL) was added and the solution was stirred overnight. This solution was centrifuged and the amine-functionalized SQS (SQS-NH2) was washed twice with ethanol and dispersed in 20 mL of ethanol ([QD] ∼ 5.3 × 10−7 M, [SQS-NH2] ∼8.1 × 10−9 M). The SQS-NH2 (1 mL) was separated using centrifugation (10 000g, 10 min) and dispersed in a carbonate buffer (pH 9.6, 50 mM, 5 mL) with 2.5% glutaraldehyde to activate the surface. To optimize the binding ratio between SQS and the SubP monoclonal antibody, different concentrations of antibody (75, 150, and 300 nM in 200 μL) and the activated SQS-NH2 (1.6 nM, 200 μL) were mixed and incubated with slow shaking (1 400 rpm) on a thermo mixer (Eppendorf AG, Hamburg, Germany) at 37 °C for 2 h to form SQS and antibody conjugation. After optimization of the reaction ratio, 300 nM of monoclonal anti-SubP antibody was used for the preparation of the SQS-labeled Ab. The SQS-labeled Ab was separated by centrifugation at 10 000g for 10 min (4 °C) and dispersed in PBS.
nanocrystal with very interesting optical advantages such as a broad excitation spectrum, narrow emission band, a high quantum yield, robustness, and photostability.30−32 QDs have been used as sensitive probes for several bioassays including immunoassays, Western blots, and immunohistochemistry, following conjugation to specific biomolecules such as antibodies, enzymes, and aptamers to improve sensitivity in assays.33−41 However, the conjugation reactions frequently yield aggregated QDs, resulting in diminished photoluminescence (PL)31,32 and, consequently, lower sensitivity. For a sensitive immunoassay, a silica sphere encapsulating a QD layer (SQS) is an attractive signaling element because it emits enhanced PL relative to its corresponding QDs at a constant QD concentration.30 Here, we report the preparation of highly photoluminescent SQS and their application to an immunosorbent assay after labeling the antibody for the quantification of SubP from serum samples for the diagnosis of AMI.
■
EXPERIMENTAL SECTION Materials. Serum samples were collected from patients with AMI (n = 16) and healthy controls (n = 30) during 2013. The serum was collected with the written informed consent of the patients at Korea University Medical Center (Seoul, Korea). The Institutional Review Board (IRB) of Korea University Medical Center approved the sample collection and analysis. The collected blood was centrifuged for 40 min at 10 000g at 4 °C, followed by the addition of 0.6 TIU/mL aprotinin and storage at −70 °C until testing. The ELISA kit for SubP was purchased from R&D Systems Inc. (Minneapolis, MN). Parylene A Coating. The Parylene precursor-Parylene A was purchased from Femto Science Co. (Seoul, Korea). The Parylene A film was coated on a polystyrene microplate using a parylene coater from KISCO Co. (Tokyo, Japan). The Parylene A polymerization process has been described previously.42,43 Preparation of Highly Photoluminescent SQS and SQS-Labeled Antibody. The SQS was prepared following a previously published protocol30 with some modifications. In the silica shell formation reaction on a QD layer-assembled silica B
dx.doi.org/10.1021/ac502412x | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Statistical Analysis. Diagnostic criteria, sensitivity, and specificity were statistically analyzed using MedCalc software (version 12.3, Mariakerke, Belgium). The correlation coefficients (R2) and group differences were calculated by the Student’s t-test. Passing and Bablock regression44 was used for method comparison. The reciprocal operating curves (ROC) were used for the determination of sensitivity and specificity. The method of Delong et al. was used for the calculation of the standard error of the area under the curve (AUC), ROCs, and of the difference between two AUCs.45 Statistical significance was considered when the P-value was less than 0.05. The correlation between commercial ELISA and SQSLISA were calculated by the statistical method of D. G. Altman.46
Characterization of SQS-Labeled Antibody. To estimate the binding ratio between SQS and the anti-SubP antibody, the Bradford assay (Bio-Rad, Hercules, CA) was performed according to the manufacturer’s protocol. The sizes of SQS-NH2 only and SQS-labeled antibody were measured using a particle size analyzer NanoBrook ZetaPALS (Brookhaven Instruments, Holtsville, NY).
■
ELISA We followed the proposed assay protocol from the manufacturers for the SubP commercial kits. The quality control samples and calibration standards were analyzed in duplicate. The serum samples were analyzed in triplicate after a 2-fold dilution. The optical density (OD) was measured at 450 nm using a microplate reader. Sandwich ELISA was performed using both Parylene Acoated plates and polystyrene plates for the comparison of detection sensitivity. In the case of Parylene A-coated plates, the surface was activated via incubation with 10% of glutaraldehyde for 1 h. The monoclonal antibody for SubP (100 ng/mL, 200 μL) was coated for 2 h at room temperature (RT). After washing, the plate was blocked with BSA (1 mg/ mL in PBS) at RT for 30 min. Each well was then washed three times with 300 μL of PBST (PBS including 0.1% Tween) before addition of 200 μL of the SubP (0.01−100 ng/mL). The plate was incubated at RT for 1 h. The plate was washed three times with PBST followed by 2 h incubation with 200 μL of polyclonal antibody (200 ng/mL). The plate was washed again with PBST three times to remove the unbound antibody and then incubated with horseradish peroxidase (HRP)-labeled antihuman IgG (250 ng/mL). After washing, 100 μL of TMB substrate was added and incubated for 30 min. Sulfuric acid (0.6 N) was added to stop the reaction. The absorbance was measured at 450 nm with a microplate reader. For ELISA using the SQS-labeled antibody (SQSLISA), both direct and competitive SQSLISA methods were employed using Parylene A-coated plates. Direct SQSLISA was performed by coating the plate with the sample followed by incubation with the SQS-labeled antibody. For competitive SQSLISA, a constant amount of SubP was coated on the plate while the sample and the SQS-labeled antibody were preincubated in a separate tube. Prior to coating with the SubP antigen, the Parylene A-plates were prepared by treatment with 10% glutaraldehyde in ethanol for 1 h. After washing twice with distilled water, the microplates were incubated with 1 μg/mL of SubP for 2 h. The Parylene A and antigen coating process is described in Figure 1a. To block the coated plates, 1 mg/mL of BSA was incubated for 30 min followed by a wash with PBST. The SQS-labeled antibody was incubated together with the SubP standard (1−10 000 pg/mL) or serum samples in tubes and separated by centrifugation. The separated antigen-SQSlabeled antibody or SQS-labeled antibody was incubated in antigen-coated wells for 3 h. After washing with PBST, the PL intensity of the plate was measured at 485 nm excitation and 610 nm emission using a fluorescence reader (PerkinElmer Inc., Waltham, MA). The competitive ELISA was used for intra- and interday validation of the SQSLISA method. A standard solution series of SubP (1, 10, 100, 1000, and 10 000 pg/mL) was used for calibration while 10 and 100 pg/mL concentrations were used for quality control (QC). The three different sets of standards and QCs were assayed within a day or in 3 different days. The ELISA method was the same as described above.
■
RESULTS AND DISCUSSION We have previously reported that SubP levels in serum were significantly elevated in AMI patients relative to healthy controls using liquid chromatography and mass spectrometry.24 However, the available commercial ELISA for SubP, which we tested, is not sensitive enough to detect the low concentrations of SubP that are present in the serum of both patients and healthy controls. In the case of healthy controls, SubP was not detectable in more than 50% of the samples.24 It has been known that the concentrations of SubP in healthy controls are lying between 1 and 100 pg/mL.47 Therefore, we developed a more sensitive immunoassay for the detection of SubP using highly photoluminescent SQS. The parameters of antibody coupling with SQS and the competitive SQSLISA protocol were optimized and validated. We modified 96-well plate with Parylene-A using glutaraldehyde and further attached SubP antigen as illustrated in Figure 1a. In this work, a Parylene-A coated microplate was used to improve the immobilization efficiency of antibodies. The Parylene-A has primary amine groups, and these amine groups can make covalent bonds with proteins using glutaraldehyde as a linker. Such a Parylene-A modified plate is known to immobilize peptide, small proteins, and antibodies with far higher efficiency in comparison with unmodified polystyrene surface.42,43 Higher sensitivity was expected to be achieved by the dense and strong coating of SubP or SubP antibody to the microplate, using glutaraldehyde as a cross-linking agent. To increase sensitivity of the immunoassay, highly photoluminescent SQS was conjugated to the detection antibody as illustrated in Figure 1b. The synthesis of highly photoluminescent SQS and its application to the immunoassay after conjugation to antibody was intended to result in high PL intensity and photostability as will be shown in the next section. Figure 1c illustrates the competitive SQSLISA scheme. The preincubation of serum with SQSlabeled antibodies was employed to reduce the background. We also tested direct SQSLISA to reduce assay steps, but the result was less reproducible than the competitive assay. The comparison between direct SQSLISA and competitive SQSLISA will be discussed in Validation of the SQSLISA Method. Preparation of Highly Photoluminescent SQS and the SQS-Labeled Antibody. To prepare a highly sensitive signaling SQS, we modified our previous synthetic method. Previously, roughly 80, 300, and 800 nm-sized SQS particles showed 2.4-, 3.2-, and 2.1-fold enhanced PL, respectively, relative to a QD-MPA solution at a constant QD concentration.28 The mechanism of enhancement is unknown. However, we reasoned that the ∼80 nm-sized SQS should display the lowest light scattering effects, and consequently the C
dx.doi.org/10.1021/ac502412x | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
SQSLISA for SubP using both polystyrene microplates and Parylene A-modified plates. Sandwich SQSLISA was used to analyze the sensitivity of Parylene A-coated plates. Parylene Acoated plates exhibited approximately 2-fold higher PL intensity than the polystyrene plates in the range of 0.01−100 ng/mL of SubP (p < 0.05, Figure 3). When used for competitive
highest PL enhancement, among the three samples. We hypothesized that the ∼80 nm-sized SQS cannot avoid aggregation between themselves during synthesis because of the high surface energy resulting from their small size and aggregation increases the light scattering effect. On the other hand, it is relatively easy to control the aggregation of the ∼300 and ∼800 nm-sized SQS. Therefore, we expected that the reduction of aggregation in ∼80 nm-SQS can lead to the highest PL enhancement among the three sizes of SQS. To reduce aggregation, the silica encapsulation reaction of SQ assembly was performed in a 5-fold diluted solution ([QD] = 5 × 10−8 M), relative to a previously reported protocol.30 Periodic sonication of the reaction mixture further reduced aggregation. Sonication aided in the dispersion of relatively large (submicrosized) particles whereas it promoted the aggregation of relatively small nanoparticles (
