Biosensors and Bioelectronics 21 (2006) 2106–2113

A strategy for sensitivity and specificity enhancements in prostate specific antigen-␣1-antichymotrypsin detection based on surface plasmon resonance Cuong Cao a , Jun Pyo Kim a , Byung Woo Kim a , Heeyeop Chae a , Hyun C. Yoon b , Sang Sik Yang c , Sang Jun Sim a,∗ a

Department of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea b Department of Biotechnology, Ajou University, Suwon 443-749, Republic of Korea c School of Electrical and Computer Engineering, Ajou University, Suwon 443-749, Republic of Korea Received 24 June 2005; received in revised form 5 October 2005; accepted 17 October 2005 Available online 23 November 2005

Abstract A biochip based on surface plasmon resonance was fabricated to detect prostate specific antigen-␣1 -antichymotrypsin (PSA-ACT complex) in both HBS buffer and human serum. To reduce non-specific binding and steric hindrance effect, the chemical surface of the sensor chips was constructed by using various oligo(ethylene glycol) mixtures of different molar ratios of HS(CH2 )11 (OCH2 CH2 )6 OCH2 COOH and HS(CH2 )11 (OCH2 CH2 )3 OH. The self-assembled monolayers were biotinylated to facilitate the immobilization of streptavidin. Using the chip surfaces, PSA-ACT complex in HBS buffer and human serum was detected at 20.7 and 47.5 ng/ml by primary immunoresponse, respectively. However, the limit of detection could be simply enhanced by a sandwich strategy to improve the sensitivity and specificity of the immunoassay. An intact PSA polyclonal antibody was used as an amplifying agent in the strategy. As a result, PSA-ACT complex concentrations as low as 10.2 and 18.1 ng/ml were found in the HBS buffer and human serum sample, respectively. The result indicates that this approach could satisfy our goal without modifying the secondary interactant. © 2005 Elsevier B.V. All rights reserved. Keywords: Prostate specific antigen; Surface plasmon resonance; Enhancement; Oligo(ethylene glycol); Biotinylation; Self-assembled monolayer

1. Introduction Prostate cancer is a major cause of death in the male population. In 2003, approximately 220,900 cases of prostate cancer were detected in the United States. Prostate malignancy exists in 30% of all men who are over the age of 50. This rate increases to 50% for men in their eighties. The disease is increasing rapidly, which has led to the prediction that prostate cancer will become the most common cause of cancer causing death in men by the year 2010 (Savage and Waxman, 1996). At present, no curative therapy is available once the disease metastasizes to other sites in the body. Early and accurate detection of prostate cancer offers the best hope to combat against the disease while it is still localized in the prostate gland.

∗

Corresponding author. Tel.: +82 31 290 7341; fax: +82 31 290 7272. E-mail address: [email protected] (S.J. Sim).

0956-5663/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2005.10.014

Prostate specific antigen (PSA), a 33-kDa serin protease, can be used to detect prostate cancer in the early stages. PSA has been recognized as the premier tumor marker for prostate cancer (Armbruster, 1993; Savage and Waxman, 1996). In addition, the major forms of PSA found in serum are complexes with two major extracellular serine protease inhibitors, ␣1 -antichymotrypsin (PSA-ACT, MW 90 kDa) and ␣2 -macroglobulin (PSA-AMG), and a free form (f-PSA, MW 34 kDa) (Lilja et al., 1991). PSA-ACT is the predominant form of PSA complex; it is immunoreactive, whereas PSA-AMG is not. The minor forms are constituted by a combination of PSA and protein C inhibitor (PSA-PCI), ␣1 -antitrypsin (PSA-AT), and ␣-trypsin (PSA-IT). Therefore, PSA-ACT and f-PSA are two molecules that if measured can be used to determine prostate cancer (Lilja et al., 1991; Savage and Waxman, 1996; Sarkar et al., 2002). Conventional assays for PSA detection mostly involve a monoclonal or a polyclonal antibody of PSA tagged with an

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

enzyme, a fluorophore or a radioactive isotope (Armbruster, 1993). These approaches had several disadvantages, such as being time consuming, inconvenient, and expensive. Furthermore, the immune reactions could not be observed in real time. Moreover, use of a radioactive isotope is very dangerous, and has not been approved by the Food and Drug Administration (FDA) (Armbruster, 1993). Surface plasmon resonance (SPR) can be used to overcome the disadvantages of traditional methods. SPR is an affinity optical sensor based on the detection of changes in mass concentration at a biospecific interface. The advantages of SPR-based biosensors are that biomolecular reactions can be monitored in real time, there is no chemical labeling, the technology is rapid, the chip is reusable, there is flexible experiment design and only a small sample size is required (Homola, 2003). On the other hand, the major disadvantage is that it is difficult to determine an analyte at low concentration or with a low molecular mass. The detection limit is approximately 1–10 nM for a 20-kDa molecule and is even higher for smaller molecules (Gomes and Andreu, 2002). To improve the detection limit or sensitivity of SPR, sandwich immunoassays using Au or Ag nanoparticles (Gu et al., 1998), latex spheres (Homola, 2003), liposomes (Wink et al., 1998), streptavidin-biotinylated antibody complexes (Pei et al., 2001), or an enzyme precipitation strategy (Kim et al., 2005) have been reported. These assays involve modifying the secondary interactants chemically so that they can be conjugated with the SPR signal enhancers. As a result, this increases the mass concentration at the biochip interface, leading to amplification of the signals. However, chemical modification of the secondary interactants, which are mostly antibodies, would partially or completely change their biological activities. Sometimes the activity loss is caused by physically blocking the antigen binding sites or by conformation changes during conjugation (Hermanson, 1996). The sensitivity of an immunological interaction can be considerably enhanced by a sandwich strategy using an intact polyclonal antibody if the molecular weight of the antigen is smaller than that of the polyclonal antibody. This approach does not change the physical, chemical or biological characteristics of the antibodies. Because polyclonal antibodies contain the entire antigen-specific antibody population, one antigen molecule can form a complex with several antibody molecules. Moreover, use of the sandwich strategy can enhance the specificity of an immune reaction because the overall specificity of the antibody–antigen–antibody sequence is higher than that of the antibody–antigen (Chapman et al., 2000; Cui et al., 2003). This paper describes a strategy for detection of PSA-ACT complex based on an SPR immunosensor in combination with signal enhancement by polyclonal antibodies. Since SPR-based detection depends on the refractive index of the medium close to the non-illuminated side of a gold layer (Homola, 2003), it is important to employ suitable materials in the construction of a monolayer onto the thin gold surface. Different self-assembled monolayer (SAM) surfaces have been developed to improve the sensitivity of SPR biosensors. Oligo(ethylene glycol) (OEG) has been recognized for its ability to prevent non-specific adsorption of proteins (Chapman et al., 2000; Benesch et al., 2001;

2107

Frederix et al., 2004; Chen et al., 2005). Thus, we functionalized a bare gold surface using various OEG mixtures of different molar ratios of HS(CH2 )11 (OCH2 CH2 )6 OCH2 COOH and HS(CH2 )11 (OCH2 CH2 )3 OH to optimize the surface coverage of protein and reduce non-specific binding. Biotin–streptavidin chemistry has been routinely used as a biomaterial immobilization device (Green et al., 1971; Orth et al., 2003). Their bond formation is very rapid and, once formed, is unaffected by wide extremes of pH, temperature, organic solvents, and other denaturing agents. Therefore, the SAM was biotinylated, followed by SA immobilization. The immobilized surface was used to determine PSA-ACT complex concentration in HBS buffer and human serum. The results show that this real time immunoassay is very simple, effective, and easy to implement. It gives an alternative method for detection of PSA-ACT complex. 2. Materials and methods 2.1. Instrumentation SPR measurements were performed on a BIAcore 2000 apparatus (Pharmacia Biosensor AB, Uppsala, Sweden). The instrument was operated using the BIAcore 2000 control software; data were evaluated using BIAevaluation 3.2. 2.2. Materials PSA-ACT complex, PSA-ACT complex monoclonal antibody (PSA-ACT mAb) and goat PSA polyclonal antibody (PSA pAb) were supplied by BiosPacific, Inc. (Emeryville, CA, USA). HS(CH2 )11 (OCH2 CH2 )6 OCH2 COOH (EG6 -COOH) and HS(CH2 )11 (OCH2 CH2 )3 OH (EG3 -OH) were purchased from Cos Biotech (Korea). Human serum, bovine serum albumin (BSA), human immunoglobulin G (IgG), fibrinogen from human plasma, streptavidin (SA), 2-morpholinoethane sulfonic acid (MES), dimethyl sulfoxide (DMSO) were purchased from Sigma–Aldrich. Biotin hydrazide, EZ-link sulfoNHS-LC-Biotinylation Kit (including a D-SaltTM dextran desalting column with a molecular weight cut-off 5000 and 2-Hydroxyazobenzen-4 -Carboxylic Acid (HABA) assay reagents) were supplied by Pierce (Rockford, IL, USA). HBS buffer (0.01 M HEPES pH 7.4, 0.15 M NaCl, 3 mM EDTA, 0.005% Surfactant P20), ethanolamine–HCl solution, 50 mM NaOH solution, N-hydroxysuccinimide (NHS), N-ethyl-N -(3diethylaminopropyl) carbodiimide (EDC), and the bare gold surface (SIA Kit Au® ) were obtained from BIAcore AB (Uppsala, Sweden). 2.3. Formation of oligo(ethylene glycol)-terminated SAMs on the bare gold surface (Fig. 1, step (a)) The gold-coated chips were first modified with a mixture of EG6 -COOH and EG3 -OH to form different mixed SAM surfaces. The clean bare gold chips were separately immersed into total 0.5 mM absolute ethanol solutions containing 1:2, 1:9 and 1:18 molar ratio of EG6 -COOH/EG3 -OH for 24 h to form a

2108

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

Fig. 1. Schematic diagram illustrating the steps in sensor chip fabrication (e.g. molar ratio of EG6 -COOH:EG3 -OH = 1:2) and detection of PSA-ACT complex based on SPR.

mixed SAM on the gold surface. The chips were then rinsed in sequence with ultra pure water (18.2 m
cm) and absolute ethanol. 2.4. Biotinylation of the carboxyl-terminated groups of the SAMs (Fig. 1, step (b)) Biotinylation of carboxyl-terminated groups of the SAMs on a solid phase is quite similar to that for an aqueous phase, which has been described elsewhere (O’Shannessey and Quarles, 1987). We modified the procedure as follows: The surface chips prepared above were rinsed in 0.1 M MES buffer (pH 4.7–5.5). They were then soaked into a clean bottle containing 5 ml of 0.1 M MES buffer with the gold-coated layers upward. The carboxyl groups were activated by adding 65 ␮l of 100 mg/ml EDC, and then conjugated with 130 ␮l of biotin hydrazide (50 mM). After 12 h at room temperature with gentle shaking, the chips were cleaned several times with ultrapure water and HBS buffer (pH 7.0). Finally, the gold chips were cleaned and dried under a pure N2 gas stream. 2.5. Biotinylation of PSA-ACT mAb (Fig. 1, step (b )) The experimental steps for biotinylating PSA-ACT mAb, purifying the biotinylated mAb, and measuring the level of biotin incorporation were carried out using the EZ-link sulfo-NHS-LCBiotinylation Kit according to the manufacturer’s protocol.

2.6. Immobilization of SA and biotinylated PSA-ACT mAb (Fig. 1, step (c) and (c )) The biotinylated chip was mounted onto the BIAcore 2000 system. The flow rate of all solutions was maintained at 5 ␮l/min and the working temperature was kept at 25 ◦ C. SA was immobilized by injecting SA solution (20 ␮g/ml in HBS buffer pH 7.4) for 7 min. Next, biotinylated PSA-ACT mAb (20 ␮g/ml in HBS buffer) was flowed over the flow cell 2 (Fc2) for 7 min. Flow cell 1 (Fc1) was considered as a reference cell for correction of the signal responses. The unbound biotinylated PSA-ACT mAbs were washed away using a mixed solution of 25 mM NaOH/0.2 M NaCl for 2 min. To investigate protein non-specific binding of the chip before determination of PSA-ACT complex, control experiments were performed by injecting a 20 ␮l volume of different concentrations of BSA, IgG and fibrinogen (10, 20 and 40 ␮g/ml) over the immobilized surface for 2 min. 2.7. Detection of PSA-ACT complex by immune reaction (Fig. 1, step (d)) In this experiment, the flow rate of all solutions was maintained at 20 ␮l/min. To reduce high viscosity and protein non-specific adsorption, human serum (with total protein concentration of approximately 90 mg/ml, pH 7–9) was diluted into 1:10 with HBS buffer prior to being used in the analyte

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

2109

preparation. PSA-ACT complex was diluted with HBS buffer or human serum to yield concentrations of 0, 1, 10, 100, 500, 1000, and 5000 ng/ml. The analyte was injected over the immobilized surface for 5 min. After measuring the immune response, the sensor chip was regenerated by injecting 50 mM NaOH/1 M NaCl for 2 min. 2.8. Detection of PSA-ACT complex by signal amplification using intact PSA pAb (Fig. 1, step (e)) Amplification of the SPR signal was investigated using a sandwich immunoassay. After the immune response between PSA-ACT and biotinylated mAb occurred, PSA pAb (20 ␮g/ml in HBS buffer) was subsequently injected over the surface for 5 min at 20 ␮l/min for the enhancement step. Regeneration of the surface was carried out as above. The relative resonance units (RU) were corrected by subtracting the reference RU values, which were obtained by flowing PSA pAb over the surface without PSA-ACT, from the enhanced RU values. The limit of detection (LOD) of the PSA-ACT complex is defined as three times the standard deviation for the average measurements of blank samples (zero concentration of PSA-ACT complex). 3. Results and discussion 3.1. Immobilization of SA and biotinylated PSA-ACT mAb onto the oligo(ethylene glycol) SAM surfaces When molecules in a sample bind to the SPR sensor surface, the concentration and thus the refractive index at the surface is changed and the SPR response can be detected. Therefore, the sensitivity of the SPR assay depends on the density of the surface ligands. To achieve a high amount of immobilized ligands that are available for binding of an analyte, it is necessary to optimize the accessibility between the terminal groups of SAMs and the protein (Patel et al., 1998). As mentioned above, the biotin–SA interaction is well understood, and its application in the development of prefunctionalized biotin–SA surface has been investigated in a number of studies (Jung et al., 1999). Moreover, biotin-functionalized SAMs yield good accessibility and lead to very fast immobilization of SA to the surface (H¨aussling et al., 1991). Based on different molar ratios of EG6 -COOH/EG3 -OH, several sensor chips were fabricated to maximize the immobilization of SA and biotinylated PSA-ACT mAb. Biotinylation of the SAMs was a prerequisite step so that SA could be immobilized onto the SAM surfaces with high coverage, specificity, and activity. The biotinylation of PSA-ACT mAb was achieved with a yield of 72% for a molar ratio of biotin/antibody of approx˚ imately 2.26. Using the extended spacer arm length (22.4 A) of sulfo-NHS-LC-biotin, the interaction between SA and the biotinylated antibody could be improved significantly, thus minimizing the steric hindrance present at the biotin binding sites of SA (Green et al., 1971).

Fig. 2. SA and biotinylated PSA-ACT mAb immobilization.

As shown in Fig. 2, SA immobilization was highest using the 1:2 surface (2242 RU), and continuously decreased with decreasing EG6 -COOH/EG3 -OH molar ratios (1:9 and 1:18 surface). However, the best biotinylated PSA-ACT mAb coverage was observed on the 1:9 surface (3326 RU). The results could be rationalized due to the fact that steric hindrance was reduced, exposing more active sites on the SA surface for biotinylated antibody binding. On the 1:2 surface, the density of biotinylated groups was significantly increased, resulting in the best SA immobilization. Nevertheless, the other free active sites of SA could be spatially overlapped by the neighbouring SA molecules. Therefore, the amount of biotinylated PSA-ACT mAb bound to the SA layer on the 1:2 surface was reduced. On the 1:18 surface, the lowest SA surface coverage was achieved, but biotinylated PSA-ACT mAb immobilization did not show the highest loading amount on the SA layer. In fact, the number of carboxyl terminal groups was reduced and hence the biotin ligands were insufficient for generation of the most suitable SA surface. The suitability mentioned could be considered as a balance between the reduction of steric hindrance and the maximization of capture protein quantity. In conclusion, the best EG6 -COOH/EG3 -OH molar ratio for the antibody binding was observed for the 1:9 surface. Prefunctionalized biotin–SA surfaces have been shown to be advantageous for the development of biosensors. Normally, SA is immobilized on a SAM via carboxyl terminal groups in the presence of NHS/EDC as activators. Then the nonreacted surface is blocked by ethanolamine–HCl or cysteine (Frederix et al., 2004). In our experiment, no blocking reagents were required during SA immobilization because SA strongly bound to the biotinylated layer, on which biotin penetrated to react with the carboxyl groups of the SAM as a result of

2110

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

the immobilization method used (O’Shannessey and Quarles, 1987). Moreover, effective docking of the bulky SA molecules is prevented by the close-packed biotin layer. SA layer formed in this way is quite homogenous, and exposes two of its binding sites away from the surface (Jung et al., 1999). Accordingly, secondary antibodies chemically modified with biotin can be rapidly immobilized on these SA-activated surfaces (Jung et al., 1999), leading to accessibility being enhanced. Jung and co-workers also showed that the stability and orientation of the SA within the adlayer can be improved if two biotin linkages are formed. Indeed, after being maintained for 3 months in HBS buffer at 4 ◦ C, the immobilized SA layers of the sensor chips were tested again in terms of biotinylated PSA-ACT mAb immobilization and no deterioration was observed. For the 1:2 surface, the SPR response was approximately 1880 RU, even slightly higher than the previous value (1816 RU). Similar results were also observed with the other sensor chips (data not shown). 3.2. Investigation of non-specific binding on the 1:9 surface In protein chip fabrication, it is essential to minimize the interaction between human fluid proteins and the sensor surfaces, especially for clinical applications. As mentioned earlier, the best biotinylated PSA-ACT mAb coverage was observed for

the 1:9 surface. With this result in mind, we decided to investigate non-specific binding between the surface materials and several serum proteins, such as albumin (BSA), IgG and fibrinogen (Fig. 3). These are the most common proteins present in human serum at high concentration. Their adsorption characteristics have been well studied on a variety of surfaces (Green et al., 1999). Furthermore, BSA and IgG have opposite net charges (Silin et al., 1997), which is a very effective factor for the investigation of non-specific binding caused by electrostatic interactions. As shown in Fig. 3, the adsorption of BSA, IgG and fibrinogen was negligible. When the concentrations of BSA, IgG and fibrinogen were increased from 10 to 40 ␮g/ml, the RU responses increased non-significantly from 11.6 to 18.7 (Fig. 3(a)), 10.8 to 15.8 (Fig. 3(b)), and 14 to 25 (Fig. 3(c)), respectively. These results indicate that the OEG—terminated SAM on the bare gold surface was very effective in reducing non-specific binding. This could be explained by the fact that the OEG has the ability to render the surface biocompatibility. The SAM surface of OEG provides a template for water nucleation, and protein resistance of OEG—terminated SAM is a consequence of the stability of the interfacial water layer, which prevents a direct contact between the surface and proteins (Silin et al., 1997; Wang et al., 1997).

Fig. 3. Non-specific binding of different proteins on the 1:9 surface. The adsorption of (a) BSA; (b) IgG; (c) fibrinogen was negligible. Dash arrow shows time when the investigated protein solutions were injected; solid arrows show time when the injections were stopped, and washing was begun by means of HBS buffer flow. Numbers 1, 2, and 3 indicate the sensorgrams generated by injecting different concentrations of 10, 20, and 40 ␮g/ml, respectively.

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

2111

Fig. 4. Representative overlaid sensorgrams showing the real time immune reactions of PSA-ACT complex (primary responses) and antibody enhancement detections (enhanced responses) on the 1:9 surface. The captions describe: (a) different concentrations of PSA-ACT complex were flowed over the biotinylated PSA-ACT mAb layer; (b) injection of PSA-ACT complex was stopped; (c) then 100 ␮l of 20 ␮g/ml PSA pAb was injected to yield the enhanced response; (d) injection of PSA pAb was stopped. The resonance units were corrected by subtracting the reference RU values from the sample RU values. A continuous HBS buffer flow was maintained during the SPR measurements.

3.3. Immune reaction between PSA-ACT complex and biotinylated PSA-ACT mAb and enhancement assay After the PSA-ACT mAb was immobilized onto the SA layer, PSA-ACT complex in HBS buffer or human serum was flowed over the 1:9 surface. Response of the antigen–antibody reaction was measured in real time using the BIAcore 2000 apparatus. Subsequently, PSA pAb was injected over the sensor surface to achieve better sensitivity and specificity. 3.3.1. Immune reaction HBS buffer As shown in Fig. 4, some sensorgrams, especially at low concentration of the analyte (e.g. 1 ng/ml), dropped below the baseline during the association phase. This was because nonspecific binding on the reference flow cell was higher than the amount of analyte bound to the antibody surface. Therefore, the corrected response momentarily gave a negative value. Overall, the dissociation phase showed stable plateaus under conditions of continuous HBS buffer flow, and therefore net binding of both

the analyte and the enhancing agent seemed to be derived from high affinity binding of the antigen–antibody reaction. Fig. 5(a) shows calibration curves generated by plotting the relative RU against PSA-ACT complex concentrations ranging from 0 to 5000 ng/ml. The error bars illustrate the relative standard deviation (R.S.D.) for three replicates in the concentrations range 1–5000 ng/ml, and ten replicates of the blank sample. The plots indicate that the PSA-ACT complex concentration could be determined by the SPR biosensor over a wide range, and the SPR signals were found to be significantly amplified because of the enhanced response. However, the analyte range was only linear for concentrations ranging from 0 to 1000 ng/ml. Fig. 5(b) shows the linear scale of the primary immune response and the enhanced response. A linear regression equation for the primary response was calculated as y = 0.2765x + 6.9604 (R2 = 0.9969, n = 6), where y and x are the relative RU response and analyte concentration, respectively. The R.S.D. of the zero concentration was 1.91. Therefore, the LOD was determined to be as low as 20.7 ng/ml. Similarly,

Fig. 5. (a) The calibration curves of PSA-ACT detection and (b) the linear range of the immunoassay in HBS buffer.

2112

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

Fig. 6. (a) The calibration curves of PSA-ACT detection and (b) the linear range of the immunoassay in human serum.

the linear regression equation for the enhanced response was y = 1.0049x + 108.76 (R2 = 0.9995, n = 6) and the R.S.D. was 3.45; the LOD derived from this equation was 10.2 ng/ml. From the slopes of the equations, it is evident that the sensitivity was improved by a factor of 3.63 (from 0.2765 to 1.0049 RU) per concentration unit (ng/ml) by the sandwich enhancement strategy. However, the LOD practically decreased by a factor of 2.03 (from 20.7 to 10.2 ng/ml), probably due to non-specific binding of the amplified antibody to the chip surface. The noise can be seen from the intercepts of the regression equations (in Fig. 5(b)), where the primary and enhanced response of the blank sample increased from 6.9604 to 108.76 RU, respectively. 3.3.2. Immune reaction in human serum The PSA-ACT complex in human serum solution was also detected by using the 1:9 surface. Fig. 6(a) shows the typical calibration curves obtained over the concentrations range 0–5000 ng/ml. Fig. 6(b) shows the linear scale of the immunoassay. Linear regression equations were also obtained by the same calculation as the previous section. In the primary response, the RU values were greatly higher than that in HBS buffer. This was because the human serum contains numerous proteins at high concentration (total protein concentration of the examined sample was 9 mg/ml), leading to a high non-specific adsorption of protein. Thus, no difference was recorded in low concentrations of PSA-ACT complex (0–10 ng/ml), and the analyte range was only linear for concentrations ranging from 10 to 1000 ng/ml in the primary response. As seen in Fig. 6(b), a linear regression equation for the primary response was y = 0.2592x + 252.1 (R2 = 0.9983, n = 4). The R.S.D. of the blank sample was 4.11, the LOD was determined to be 47.5 ng/ml. By applying the intact PSA pAb for the enhanced response, the linear range was over 0–1000 ng/ml. The linear regression equation for the enhanced response was y = 0.6295x + 100.86 (R2 = 0.9985, n = 6), and the R.S.D. was 3.8. Consequently, the LOD was found to be as low as 18.1 ng/ml PSA-ACT complex. The slopes of the equations show that the sensitivity was improved by a factor of 2.43 (from 0.2592 to 0.6295 RU) per concentration unit (ng/ml). However, to compare with that in HBS buffer, the sensitivity enhancement of the

immunoassay was deteriorated from 3.63 to 2.43. Obviously, this is attributed to interference of the high non-specific adsorption of protein in human serum during the primary immunoassay, thus it leads to a high noise background indicated by the intercept of the equation (252.1 RU). Interestingly, it was observed that the noise background in the enhanced response (100.86 RU) did not greatly change to that in HBS buffer (108.76 RU). In this case, the intact PSA pAb shows its function as an effective agent to enhance the specificity of the assay. In spite of the high concentration of protein in the crude sample, PSA-ACT complex was also detected at a quite low concentration (18.1 ng/ml) specifically. In immune reaction based on SPR, it is generally accepted that the specificity of the assay is enhanced by a sandwich technique, because the amplified antibody can mostly interact with its antigen (F¨agerstam et al., 1992; Cui et al., 2003). On the other hand, the sensitivity seems to depend on the class of antibodies added and their molecular weight, as well as the antigenic valence for specific binding of antibodies. PSA levels are closely related to tumor volume and clinical stage, in that over 90% of patients with PSA levels above 20 ng/ml are in the metastatic stage of the disease (Savage and Waxman, 1996). In this experiment, the detection limit of the assay in a real human serum solution was found to be as low as 18.1 ng/ml PSA-ACT complex after sandwich enhancement. Moreover, it took less than 20 min for each entire measurement (Fig. 4). Therefore, the result indicates that sandwich amplification can be used for fast detection of prostate specific antigen, and it could meet the requirements for clinical application. 4. Conclusions A simple strategy for real time detection based on an SPR system was used for the first time to determine the concentration of PSA-ACT complex in both HBS buffer and human serum. An immunosensor chip was functionalized using a mixture of EG6 COOH and EG3 -OH to reduce steric hindrance. The best result was observed for a 1:9 molar ratio of EG6 -COOH/EG3 -OH. Biotinylation of the SAM was found to facilitate SA immobilization as well as to stabilize the SA layer. Using this chip, the

C. Cao et al. / Biosensors and Bioelectronics 21 (2006) 2106–2113

PSA-ACT complex in HBS buffer was detected in the primary response at a concentration of 20.7 ng/ml. Moreover, the detection limit was decreased to 10.2 ng/ml by simply amplifying the primary signal with intact PSA pAb without any chemical modification. In human serum, the LODs were 47.5 and 18.1 ng/ml PSA-ACT complex for the primary and enhanced response, respectively. The results show that the sandwich strategy significantly improved the sensitivity and the specificity of the immunoassay. Although simple in design and concept, the sensor chip fabrication and enhancement strategy could provide a fast and useful tool for detection of the PSA-ACT complex. Acknowledgement This work was supported by grant no. RTI04-03-05 from the Regional Technology Innovation Program of the Ministry of Commerce, Industry and Energy (MOCIE). References Armbruster, D.A., 1993. Prostate specific antigen: biochemistry, analytical methods, and clinical application. Clin. Chem. 39, 181–195. Benesch, J., Svedhem, S., Svensson, S.C.T., Valiokas, R., Liedberg, B., Tengvall, P., 2001. Protein adsorption to oligo(ethylene glycol) self-assembled monolayers: experiments with fibrinogen, heparinized plasma, and serum. J. Biol. Sci. 12 (6), 581–597. Chapman, R.G., Ostuni, E., Yan, L., Whitesides, G.M., 2000. Preparation of mixed self-assembled monolayers (SAMs) that resist adsorption of proteins using the reaction of amines with a SAM that presents interchain carboxylic anhydride groups. Langmuir 16, 6927–6936. Chen, H., Zhang, Z., Chen, Y., Brook, M.A., Sheardown, H., 2005. Protein repellant silicone surfaces by covalent immobilization of poly(ethylene oxide). Biomaterials 26, 2391–2399. Cui, X., Yang, F., Sha, Y., Yang, X., 2003. Real time immunoassay of ferritin using surface plasmon resonance biosensor. Talanta 60, 53–61. ˚ F¨agerstam, L.G., Karlsson, A.F., Karlsson, R., Persson, B., R¨onnberg, I., 1992. Biospecific interaction analysis using surface plasmon resonance detection applied to kinetic, binding site and concentration analysis. J. Chromatogr. A 597, 397–410. Frederix, F., Bonroy, K., Reekmans, G., Laureyn, W., Campitelli, A., Abramov, M.A., Dehaen, W., Maes, G., 2004. Reduced non-specific adsorption on covalently immobilized protein surfaces using poly(ethylene oxide) containing blocking agents. J. Biochem. Biophys. Methods 58, 67–74. Gomes, P., Andreu, D., 2002. Direct kinetic assay of interactions between small peptides and immobilized antibodies using a surface plasmon resonance biosensor. J. Immunol. Methods 259, 217–230. Green, N.M., Konieczny, L., Toms, E.J., Valentine, R.C., 1971. The use of bifunctional biotinyl compounds to determine the arrangement of subunits in avidin. Biochem. J. 125, 781–791.

2113

Green, R.J., Davies, M.C., Roberts, C.J., Tendler, S.J.B., 1999. Competitive protein adsorption as observed by surface plasmon resonance. Biomaterials 20, 385–391. Gu, J.H., L¨u, H., Chen, Y.W., Liu, L.Y., Wang, P., Ma, J.M., Lu, Z.H., 1998. Enhancement of the sensitivity of surface plasmon resonance biosensor with colloidal gold labeling technique. Supramol. Sci. 5, 695– 698. H¨aussling, L., Ringsdorf, H., Schmitt, F.J., Knoll, W., 1991. Biotinfunctionalized self-assembled monolayers on gold: surface plasmon optical studies of specific recognition reactions. Langmuir 7 (9), 1837–1840. Hermanson, G.T., 1996. Bioconjugate Techniques. Academic press, California, pp. 456–493. Homola, J., 2003. Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem. 377, 528–539. Jung, L.S., Nelson, K.E., Campbell, C.T., Stayton, P.S., Yee, S.S., P´erezLuna, V., L´opez, G.P., 1999. Surface plasmon resonance measurement of binding and dissociation of wild-type and mutant streptavidin on mixed biotin-containing alkylthilate monolayers. Sens. Actuators B 54, 137– 144. Kim, M.G., Shin, Y.B., Jung, J.M., Ro, H.S., Chung, B.H., 2005. Enhanced sensitivity of surface plasmon resonance (SPR) immunoassays using a peroxidase–catalyzed precipitation reaction and its application to a protein microarray. J. Immunol. Methods 297, 125–132. Lilja, H., Christtensson, A., Dahl´en, U., Matikainen, M., Nilsson, O., Petterson, K., L¨ovgren, T., 1991. Prostate specific antigen in serum occurs predominantly in complex with ␣1 -antichymotrypsin. Clin. Chem. 37, 1618–1625. Orth, R.N., Clark, T.G., Craighead, H.G., 2003. Avidin–biotin micropatterning methods for biosensor applications. Biomed. Microdevices 5, 29–34. O’Shannessey, D.J., Quarles, R.H., 1987. Labeling of the oligosaccharide moieties of immunoglobulins. J. Immunol. Methods 99, 153–161. Patel, N., Davies, M.C., Heaton, R.J., Roberts, C.J., Tendler, S.J.B., Williams, P.M., 1998. A scanning probe microscopy study of the physisorption and chemisorption of protein molecules onto carboxylate terminated selfassembled monolayers. Appl. Phys. A 66, S569–S574. Pei, R., Yang, X., Wang, E., 2001. Enhanced surface plasmon resonance immunosensing using a streptavidin-biotinylated protein complex. Analyst 126, 4–6. Sarkar, P., Pal, P.S., Ghosh, D., Setford, S.J., Tothill, I.E., 2002. Amperometric biosensors for detection of the prostate cancer marker (PSA). Int. J. Pharm. 238, 1–9. Savage, P., Waxman, J., 1996. PSA and prostate cancer diagnosis. Eur. J. Cancer 32A, 1097–1099. Silin, V., Weetall, H., Vanderah, D.J., 1997. SPR studies of the non-specific adsorption kinetics of human IgG and BSA on gold surfaces modified by self-assembled monolayers (SAMs). J. Colloid Interface Sci. 185, 94– 103. Wang, R.L.C., Kreuzer, H.J., Grunze, M., 1997. Molecular conformation and solvation of oligo(ethylene glycol)-terminated self-assembled monolayers and their resistance to protein adsorption. J. Phys. Chem. B 101, 9767–9773. Wink, T., Zuilen, S.J., Bult, A., Bennekom, W.P., 1998. Liposome-mediated enhancement of the sensitivity in immunoassays of proteins and peptides in surface plasmon resonance spectrometry. Anal. Chem. 70, 827–832.