Integrated Poly(dimethysiloxane) with an Intrinsic Nonfouling Property

Jul 7, 2010 - Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, Suzhou, 215125, P. R. China, Academy for Advanced Interdisc...
0 downloads 3 Views 214KB Size
Download PDF
Anal. Chem. 2010, 82, 6338–6342

Letters to Analytical Chemistry Integrated Poly(dimethysiloxane) with an Intrinsic Nonfouling Property Approaching “Absolute” Zero Background in Immunoassays Hongwei Ma,*,†,‡ Yuanzi Wu,‡ Xiaoli Yang,§,| Xing Liu,† Jianan He,‡ Long Fu,†,‡ Jie Wang,† Hongke Xu,§ Yi Shi,§ and Renqian Zhong| Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, Suzhou, 215125, P. R. China, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, 100871, P. R. China, HealthDigit Co. Ltd., Shanghai, 201403, P. R. China, and Changzheng Hospital, Shanghai, 200002, P. R. China The key to achieve a highly sensitive and specific protein microarray assay is to prevent nonspecific protein adsorption to an “absolute” zero level because any signal amplification method will simultaneously amplify signal and noise. Here, we develop a novel solid supporting material, namely, polymer coated initiator integrated poly(dimethysiloxane) (iPDMS), which was able to achieve such “absolute” zero (i.e., below the detection limit of instrument). The implementation of this iPDMS enables practical and high-quality multiplexed enzyme-linked immunosorbent assay (ELISA) of 11 tumor markers. This iPDMS does not need any blocking steps and only require mild washing conditions. It also uses on an average 8-fold less capture antibodies compared with the mainstream nitrocellulose (NC) film. Besides saving time and materials, iPDMS achieved a limit-of-detection (LOD) as low as 19 pg mL-1, which is sufficiently low for most current clinical diagnostic applications. We expect to see an immediate impact of this iPDMS on the realization of the great potential of protein microarray in research and practical uses such as large scale and highthroughput screening, clinical diagnosis, inspection, and quarantine. Nonspecific protein adsorption (NPA) is the main source of noise in immunoassays and has become the major obstacle in developing more sensitive and higher throughput immunoassays. NPA must be suppressed because sensitivity greatly depends on the power of signal amplification methods,1 which unfortunately only have limited selectivity in amplifying signal and noise. Highthroughput assays typically use high concentration solutions of enzyme conjugated detection antibodies and complex samples * To whom correspondence should be addressed. E-mail: hwma2008@ sinano.ac.cn. † Chinese Academy of Sciences. ‡ Peking University. § HealthDigit Co. Ltd. | Changzheng Hospital. (1) Srivastava, S.; Labaer, J. Nat. Biotechnol. 2008, 26, 1244–1246.

6338

Analytical Chemistry, Vol. 82, No. 15, August 1, 2010

such as physiological fluid, which poses severe challenge to sensitivity because NPA is protein concentration dependent.2 We report herein the preparation of a novel solid supporting material, namely, polymer coated initiator integrated poly(dimethysiloxane) (iPDMS) that is able to prevent NPA to an “absolute” zero level thus to achieve high sensitivity by using higher amplification power. We further applied iPDMS in a multiplexed enzyme-linked immunosorbent assay (ELISA) of 11 tumor markers and confirmed that both high sensitivity and high throughput could be achieved. These results encourage us to explore new combinations of iPDMS with other more powerful signal amplification methods,3-9 thus to increase both sensitivity and throughput without delicate sample purification.10 To overcome the indiscriminative amplification problem, many efforts were dedicated to develop novel polymer substrates, including nitrocellulose (NC),11 polystyrene (PS),12 poly(methyl methacrylate) (PMMA),13 nylon,14 poly(vinylidene fluoride),15,16 (2) Ma, H. Oligo(ethylene glycol) Based Nonfouling Surfaces; VDM Verlag Dr. Mu ¨ ller Aktiengesellschaft & Co. KG: Saarbru ¨ cken, Germany, 2009; pp 36. (3) Sano, T.; Smith, C. L.; Cantor, C. R. Science 1992, 258, 120–122. (4) Lizardi, P. M.; Huang, X. H.; Zhu, Z. R.; Bray-Ward, P.; Thomas, D. C.; Ward, D. C. Nat. Genet. 1998, 19, 225–232. (5) Zhang, H. T.; Cheng, X.; Richter, M.; Greene, M. I. Nat. Med. 2006, 12, 473–477. (6) Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757–1760. (7) Nam, J. M.; Thaxton, C. S.; Mirkin, C. A. Science 2003, 301, 1884–1886. (8) Sikes, H. D.; Hansen, R. R.; Johnson, L. M.; Jenison, R.; Birks, J. W.; Rowlen, K. L.; Bowman, C. N. Nat. Mater. 2008, 7, 52–56. (9) Chen, Z.; Tabakman, S. M.; Goodwin, A. P.; Kattah, M. G.; Daranciang, D.; Wang, X. R.; Zhang, G. Y.; Li, X. L.; Liu, Z.; Utz, P. J.; Jiang, K. L.; Fan, S. S.; Dai, H. J. Nat. Biotechnol. 2008, 26, 1285–1292. (10) Stern, E.; Vacic, A.; Rajan, N. K.; Criscione, J. M.; Park, J.; Ilic, B. R.; Mooney, D. J.; Reed, M. A.; Fahmy, T. M. Nat. Nanotechnol. 2010, 5, 138–142. (11) De Wildt, R. M. T.; Mundy, C. R.; Gorick, B. D.; Tomlinson, I. M. Nat. Biotechnol. 2000, 18, 989. (12) Ekins, R. P. Clin. Chem. 1998, 44, 2015. (13) Fixe, F.; Dufva, M.; Telleman, P.; Christensen, C. B. V. Nucleic Acids Res. 2004, 32, e9. (14) Bussow, K.; Cahill, D.; Nietfeld, W.; Bancroft, D.; Scherzinger, E.; Lehrach, H.; Walter, G. Nucleic Acids Res. 1998, 26, 5007. (15) Lueking, A.; Horn, M.; Eickhoff, H.; Bussow, K.; Lehrach, H.; Walter, G. Anal. Biochem. 1999, 270, 103. (16) Afanassiev, V.; Hanemann, V.; Wolfl, S. Nucleic Acids Res. 2000, 28, e66. 10.1021/ac101277e  2010 American Chemical Society Published on Web 07/07/2010

Figure 1. Multiplexed ELISAs on iPDMS and NC. (a) Protein microarray was printed on the surface of polymer coated iPDMS, followed by routine processing as sandwich type ELISA. (b) After jet-printing of capture antibodies to predetermined locations, the iPDMS sheet was sandwiched by a plastic base and a plastic cover, transforming the flat surface to a 48-well plate. The lower right part of the figure showed a chemiluminescence image taken by a CCD camera. See Figure S2 in the Supporting Information for the complete view. (c) Protein microarray was similarly printed on the NC surface. Note there is one extra BSA blocking step compared with iPDMS. (d) Possible sources of signal and noise: (1) specific interaction that gives signal; (2) noises due to cross interaction; noises due to nonspecific adsorption of (3) antigens (i.e., tumor markers) and (4) horseradish peroxidase conjugated detection antibodies (2nd Ab-HRP).

poly-L-lysine (PLL),17 and polyelectrolyte thin films.18 Many types of surface chemistries were also developed to reduce NPA but only with limited success, including dextran,19 (ethylene glycol) alkane thiol self-assembled monolayers (EG-SAMs),20 1-pyrenebutanoic acid, succinimidyl ester via π-π stacking,21 physisorption of PLL-poly(ethylene glycol) (PEG) copolymers,22 and the most common, bovine serum albumin (i.e., bovine serum albumin (BSA) blocking). We reported the first case that applied surface initiated polymerization (SIP) to prepare oligo(ethylene glycol) methacrylate (OEGMA) based polymer brushes, which could reduce NPA below the detection limit of surface plasmon resonance.23 This strategy was applied here to demonstrate that an absolute zero background is critical in increasing sensitivity and throughput of immunoassays. RESULTS AND DISCUSSION Preparation of Microarray. The polymer coated iPDMS was prepared as we recently reported.24 Briefly, iPDMS was first prepared from a simple thermo curing procedure (see the (17) Haab, B. B.; Dunham, M.; Brown, J.; Brown, P. O. Genome Biol. 2001, 2, research0004.1. (18) Zhou, X. C.; Zhou, J. Z. Proteomics 2006, 6, 1415. (19) Lo ¨fa˚s, S.; Johnsson, B. J. Chem. Soc., Chem. Commun. 1990, 1526, 1528. (20) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714–10721. (21) Chen, R. J.; Zhang, Y. G.; Wang, D. W.; Dai, H. J. J. Am. Chem. Soc. 2001, 123, 3838–3839. (22) Pasche, S.; De Paul, S. M.; Voros, J.; Spencer, N. D.; Textor, M. Langmuir 2003, 19, 9216–9225. (23) Ma, H. W.; Hyun, J. H.; Stiller, P.; Chilkoti, A. Adv. Mater. 2004, 16, 338– 341. (24) Wu, Y. Z.; Huang, Y. Y.; Ma, H. W. J. Am. Chem. Soc. 2007, 129, 7226– 7227.

Supporting Information, Scheme S1). The resulting elastomer was then subjected to surface modification via SIP of OEGMA. The resulting poly(OEGMA) coating was further functionalized with carboxyl groups (Scheme 1 in Figure 1a). This poly(OEGMA) coated iPDMS (shortened as iPDMS thereafter) has typically a water contact angle of 60° and a poly(OEGMA) coating with an estimated thickness of 1-2 µm. The NC sheet was tested as a reference solid supporting material because it is one of the most popular solid supporting material, especially in commercial products (i.e., Whatman FAST slides, R&D Systems Proteome Profiler Antibody Arrays, and GENTEL Symphony Array services). All experiments were conducted in a microarray format,25 and details can be found in the Supporting Information. To a 7.7 × 6.6 cm2 sheet of iPDMS or NC, capture antibodies were first jetprinted to 48 predetermined locations (Figure S1 in the Supporting Information). The sheet was then clamped by two plastic plates to form a 48-well plate; each well contained one 4 × 6 array (Figure 1 and Figure S2 in the Supporting Information). Probing solutions containing target antigens (i.e., tumor markers) were then added to those wells, followed by (1) addition of detection solutions containing horseradish peroxidase conjugated detection antibodies (referred to as second Ab-HRP thereafter) and (2) chemiluminescence substrates. The performance of microarray was finally evaluated by a cooled CCD camera and analyzed as following. “Absolute” Zero Background on iPDMS. The intrinsic nonfouling property of iPDMS gives a near “absolute” zero (25) Sun, Z. H.; Fu, X. L.; Zang, L.; Yang, X. L.; Liu, F. Z.; Hu, G. X. Anticancer Res. 2004, 24, 1159–1165.

Analytical Chemistry, Vol. 82, No. 15, August 1, 2010

6339

Figure 2. Analysis of LOD and noise. (a) Averaged background intensities under different conditions. Each value was averaged from 72 different spots. The gray bar was the noise due to the cooled CCD itself. The red bars were the background intensities due to nonspecific adsorption of a cocktail of 2nd Ab-HRP on iPDMS at concentrations of 0, 2.56, 5.12, and 10.24 µg/mL. The blue bars were the background intensities on NC at concentrations of 0 and 2.15 µg/mL. (b) Phosphate buffer was printed on both NC and iPDMS, followed by probing with a cocktail of 10 antigens of various concentrations, from 0 to 8 Cstd. See Table S1 in the Supporting Information for Cstd. The detection solution was a cocktail of 10 detection antibodies with a total concentration at 1.95 µg/mL. The signal was averaged from 4 spots.(c) Signal vs marker CEA concentration plot from a multiplexed ELISA of 11 tumor markers. (d) Signal vs marker CEA concentration plot from a single indexed ELISA. (e) Signal vs marker CEA concentration plot from a multiplexed ELISA of 11 tumor markers of 2nd Ab-HRP on iPDMS at concentrations of 2.56, 5.12, and 10.24 µg/mL.

background (i.e., below the detection limit of instrument) even under high amplification power. We first determined background luminescence readings by taking images of blank iPDMS and NC sheets (Figure 2a). The values were 58 for iPDMS and 65 for NC, which were very close to the 58 reading of the instrument noise (i.e., noise due to the cooled CCD camera itself). Next, we omitted the printing and probing solutions, directly exposed iPDMS and NC to detection solutions, followed by addition of chemiluminescence substrates and CCD camera imaging. NC showed increased chemiluminescence intensity from 65 to 156 even after BSA blocking when the concentration of second Ab-HRPs increased from 0 to 2.0 µg/mL (Figure 2a). This increased reading was attributed to the nonspecific adsorption of second Ab-HRPs from detection solutions. However, iPDMS did not show a significant increase in intensity. The readings were 62, 67, and 70 for elevated concentrations of second Ab-HRPs from 2.56, 5.12 to 10.24 µg/ mL, indicating iPDMS prevented NPA even at high protein concentration without any blocking procedure. This feature gives 6340

Analytical Chemistry, Vol. 82, No. 15, August 1, 2010

iPDMS the advantage of using a high concentration of second Ab-HRP to gain a higher power of signal amplification. We further challenged iPDMS with probing solutions of high antigen concentrations to prove its intrinsic nonfouling property. Capture antibodies from Table 1 were printed on iPDMS and NC except that β-HCG was replaced with phosphate buffer. The array of capture antibodies for 10 tumor markers was then challenged with a cocktail of 10 antigens of various concentrations, from 0 to 8 Cstd. We used Cstd to represent a mixture of the 10 tumor markers at individual concentrations as listed in Table S1 in the Supporting Information because tumor markers had different concentration units. It is well-known that NPA is concentration dependent.1 Thus, with analysis of the intensity of the dots of phosphate buffer, one could know the noise contribution of the nonspecific adsorption of target antigens. The signal (see the definition in the Supporting Information) from dots of phosphate buffer on iPDMS were almost constant (Figure 2b): negligible for C/Cstd < 0.1 (the clinical relevant region) and the highest

Table 1. Performance Analysis of Multiplexed ELISAs on iPDMS and NC Copta(mg/mL)

LOD

a

marker

Cref

iPDMS

NC

RNC/iPDMS

iPDMS

NC

AFP CEA CK19 C-PSA NSE SCC CA15-3 CA125 CA19-9 CA242 β-HCG

20 ng/mL 5 ng/mL 3.3 ng/mL 4 ng/mL 13 ng/mL 2.5 ng/mL 35 KU/L 35 KU/L 37 KU/L 20 KU/L 3 mIU/mL

705 pg/mL 109 pg/mL 66 pg/mL 39 pg/mL 185 pg/mL 19 pg/mL 140 U/L 501 U/L 175 U/L 105 U/L 130 mIU/L

3917 pg/mL 552 pg/mL 468 pg/mL 826 pg/mL 3972 pg/mL 176 pg/mL 426 U/L 601 U/L 582 U/L 2234 U/L 498 mIU/L

5.6 5.1 7.1 21.2 21.5 9.3 3.0 1.2 3.3 21.3 3.8

0.08 0.08 0.1 0.1 0.1 0.1 0.1 0.1 0.04 0.04 0.05

0.5 0.6 0.8 1 0.6 0.5 0.6 0.6 0.2 1 0.6

Optimized concentration of capture antibodies on NC and iPDMS, respectively.

noise was only ∼32 at C/Cstd ) 8. This constant blank signal was also attributed to the intrinsic nonfouling property of iPDMS. The ability of iPDMS to suppress NPA even under high concentrations of antigen and second Ab-HRP leads to the near “absolute” zero background. However, NC showed an antigen concentration dependent noise level: the highest noise was 1770, which indicated the blocking procedure failed at high protein concentration. After proving iPDMS has intrinsic nonfouling property, we now show evidence that a near “absolute” zero background is the key for achieving high sensitivity and high throughput. Performance of Multiplexed ELISAs. Multiplexed ELISAs of 11 tumor markers were performed on both iPDMS and NC at their own optimized conditions (Table 1). We first evaluated the performance of multiplexed ELISAs in terms of limit of detection (LOD, Table 1, Figure 2c and Figure S5 in the Supporting Information). In the case of marker CEA, iPDMS and NC showed similar performance at the high concentration range. However, iPDMS outperformed NC in the low concentration range (LOD 109 vs 552 pg/mL, Table 1), which was the critical range for early diagnosis. The ratios of LOD (RNC/iPDMS) were from 1.2 to 21.5, suggesting iPDMS is more sensitive than NC. We picked three markers that exhibited the largest differences to run single indexed ELISA, namely, CEA (5-fold, Figure 2c), CA242 (21fold), and c-PSA (21-fold, Figure S6 in the Supporting Information). In the case of marker CEA, each well contains 24 dots of capture antibody of CEA, the probing solutions are serially diluted CEA solutions, and the detection solution contains only the second Ab-HRP for CEA. Surprisingly, the difference in LOD between iPDMS and NC was reduced or even diminished (Figure 2d and Figure S6 in the Supporting Information). For example, under material specific optimized conditions (Table S5 in the Supporting Information), single indexed ELISA of marker CEA on both iPDMS and NC gave similar results in terms of the LOD (∼130). This result was explained by the following analysis. We identified four sources of intensity (Figure 1d): (1) signal from a specific antigen-antibody interaction, which was desired and existed in both iPDMS and NC; (2) noise due to cross interactions between capture antibodies and antigens, antigens and second Ab-HRP; and noise due to nonspecific adsorption of (3) antigens and (4) the second Ab-HRP. Note that all sources of noise were finally connected to the amount of unwanted immobilization of second Ab-HRP, which could be very small but

its impact as noise will be amplified by its enzymatic activity. Noise source (2) could be eliminated by using high-quality monoclonal antibodies and limiting the number of antigen-antibody pairs under 50,26 which was the case in this study. This was equally applicable to both iPDMS and NC. In previous sections, we proved that iPDMS with an intrinsic nonfouling property approached a near “absolute” zero background even under high protein concentration while NC showed a concentration dependent NPA. From Figure 2b and Table 1, we knew that the concentration of target antigens was low and would not cause significant noise, which eliminated sources (3) to be the main source of noise. Therefore, source (4) was the main contributor of noise on NC. Since single indexed ELISA used a 20-fold lower concentration of second Ab-HRP (0.11 vs 2.15 µg/mL), noise from source (4) was reduced and BSA blocking was successful at this low concentration, which led to an improved performance of NC in terms of the LOD. For iPDMS, however, the LOD was kept almost the same (109 for multiplexed and 137 for single indexed ELISA, see Table S6 in the Supporting Information) since iPDMS had a near “absolute” zero background for the whole range of protein concentration tested. Different dilution of second Ab-HRP from a multiplexed ELISA at concentrations of 2.56, 5.12, and 10.24 µg/mL were tested on iPDMS (Figure 2e). The sensitivity and dynamic range was increasced obviously when the signal amplification power (i.e., concentration of second Ab-HRP) increased as well as the noisy signal (blank) remained the same. Since multiplexed ELISA intrinsically requires the use of a high total concentration of detection solution, a solid supporting material that provides a near “absolute” zero background will be invaluable by making the assay more practical and sensitive. iPDMS also produced a near “absolute” zero background in complex physiological fluid, such as the serum of patients (Figure S11 and Table S9 in the Supporting Information). As we demonstrated in previous sections, only unwanted immobilization of second Ab-HRP led to noise, e.g., direct nonspecific adsorption of second Ab-HRP or indirect immobilization of second Ab-HRP via specific interaction with nonspecifically adsorbed target antigens. Although serum typically has an estimated total protein concentration at 60-85 mg mL-1, the nonspecific adsorption of irrelevant proteins will not necessary lead to the nonspecific (26) Saviranta, P.; Okon, R.; Brinker, A.; Warashina, M.; Eppinger, J.; Geierstanger, B. H. Clin. Chem. 2004, 50, 1907–1920.

Analytical Chemistry, Vol. 82, No. 15, August 1, 2010

6341

adsorption of second Ab-HRP. Therefore, the use of negative serum would not cause significant noise in ELISA. Rather than challenge the surface with serum or whole blood, it is more relevant to challenge the surface with a higher total concentration of second Ab-HRP (or the use of a higher power amplification method). Furthermore, we want to emphasize that LOD is determined bymultifactors:thedissociationconstant(KD)oftheantigen-antibody pair tested and the dynamic range of the detection method. For example, the dynamic range of the CCD camera used in this study is only 60-4000, which limited the LOD. As pointed out by LaBaer et al.,1 performance test should also be carried out with real-life, weak interaction protein pairs. We determined the KD of 22 pairs tested by the surface plasma resonance method (Figure S12 and Table S10 in the Supporting Information). The KD values of 11 pairs of capture antibodies and target antigens were very close, mostly at 10-8 to ∼10-9 M, which agreed with the fact that the LOD values were mostly at 1 pM. Although the KD values of 11 pairs of target antigens and second Ab-HRPs were different, from 10-8 to ∼10-10 M, this step was expected to have less impact on the LOD.26 Many factors determine the performance of an immunoassay and whether this assay will be useful in real practice, including time, cost, accuracy, sensitivity and throughput. In multiplexed

6342

Analytical Chemistry, Vol. 82, No. 15, August 1, 2010

ELISA, the increase of throughput causes the decrease of sensitivity. This contradiction was solved by using a novel solid supporting material, namely iPDMS that has intrinsic nonfouling property and could achieve near absolute zero background. We believe that suppressing NPA is a general solution to achieve higher sensitivity and throughput in immunoassays. ACKNOWLEDGMENT Y.W. and X.Y. contributed equally to this work. This work was supported by 100 Talents Programmer of CAS (Grant 08BM031001), the NSFC (Grants 20604002 and 50773001), Fok Ying Tung Education Foundation (Grant 114013), and the National Basic Research Program of China (Grant 2009CB320300). The authors would like to thank the electroscope support of Public Center for Characterization and Test, SINANO, CAS. SUPPORTING INFORMATION AVAILABLE Text, tables, figures, detailed experimental protocols, additional results, and analyses. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review May 15, 2010. Accepted June 28, 2010. AC101277E