Immunoassay Arrays Fabricated by Dip-Pen Nanolithography with

May 22, 2013 - Here, we report the first use of resonance Raman scattering for the detection of miniaturized microscale arrays fabricated by dip-pen ...
0 downloads 0 Views 231KB Size
Download PDF
Letter pubs.acs.org/ac

Immunoassay Arrays Fabricated by Dip-Pen Nanolithography with Resonance Raman Detection Stacey Laing, Eleanore J. Irvine, Aaron Hernandez-Santana, W. Ewen Smith, Karen Faulds, and Duncan Graham* Centre for Molecular Nanometrology, WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, U.K. ABSTRACT: Here, we report the first use of resonance Raman scattering for the detection of miniaturized microscale arrays fabricated by dip-pen nanolithography. Antibody arrays for prostate-specific antigen (PSA) were printed, and a sandwich immunoassay was carried out. An enzyme-linked detection antibody was used to provide an insoluble and stable colored microdot in the recommended size range for microarray readers, which could be read with resonance Raman scattering. This gives quantitative detection as well as an improved detection limit and a larger dynamic range than that previously achieved by direct fluorescent detection methods. By Raman mapping across the arrayed area, the microdots were easily detected with very little background signal from surrounding areas. Levels of PSA as low as 25 pg/mL were detected using this method, which could be extended to a large number of useful biomarkers.

T

times, reagent compatibility, and detection capabilities of conventional scanners has led to the suggestion that the ideal range of the feature size is 1−10 μm.17,18 The fabrication of protein arrays can be achieved by the use of many techniques, such as inkjet printing,19 photolithography,20 microcontact printing (μ-CP),21 electron-beam lithography,22 and dip-pen nanolithography (DPN).23 Of these methods, DPN is desirable, as it is relatively cheap to buy and run, has extremely high accuracy and resolution, is flexible and highly scalable, and can function under ambient conditions.24 DPN has been utilized for the printing of biomolecules both directly25−27 and indirectly8,28 and is suitable for both microand nanoscale printing.27,29,30 Although protein array detection is most often carried out using fluorescence, alternative methods such as chemiluminescence have been investigated in order to improve sensitivity.10 While lower detection limits can be achieved using this technique, the multiplexing capabilities of chemiluminescence are limited.31 Rolling-circle amplification (RCA) has also been utilized to improve the sensitivity of protein immunoassay arrays.12 This is achieved by attaching a DNA sequence to the final antibody in the immunoassay and using RCA to replicate the sequence, thus allowing many fluorescent probes to be incorporated, which will enhance the observed signal. However, increased sensitivity without the need for DNA labeling and amplification is clearly more desirable.

he discovery, detection and quantification of key biomarkers can lead to a better understanding of disease, processes, detection, and diagnosis of disease and improvement of drug development through the monitoring of the therapeutic response.1 This, however, requires the development and use of sensitive and high-throughput screening techniques for the detection of proteins. The enzyme-linked immunosorbent assay (ELISA) is a widely utilized method for the detection and quantification of biomarkers. Although simple and specific, ELISAs suffer from high sample consumption, lack of multiplexing capability, and relatively low throughput.2,3 Protein array-based technology is a powerful emerging tool which seeks to overcome these limitations. By printing proteins on a surface in an array format, rather than bulk coating the entire surface, immunoassays can be built in a similar way to a standard ELISA, while introducing the possibility of multiplexing, reduced reagent volumes, and improved throughput.4−9 First exploited by Silzel et al.,4 immunoassay arrays have been investigated for the detection of many useful analytes creating highly multiplexed assays10−12 with detection limits comparable4,13 and superior7,12 to those of a standard well-based ELISA. Many of these assays have been applied to biological samples, demonstrating their capabilities for detection and quantification of biomarkers in blood serum as well as monitoring production rates by cells.7,10,12 Although microarrays typically include feature sizes of around 100−200 μm, further miniaturization would result in a reduction in reagent volumes while producing arrays of higher density; thus allowing for potential analysis of more information in a single assay. Miniaturization on the nanoscale has been achieved;14−16 however, issues with long printing © XXXX American Chemical Society

Received: March 27, 2013 Accepted: May 22, 2013

A

dx.doi.org/10.1021/ac4009012 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Letter

three times with wash buffer. Two-fold dilutions of PSA in reagent diluent (60−0.94 ng/mL) were prepared, and a sample of each was added to a separate well of the incubation chamber. Reagent diluent only was added to one of the wells as a blank control. The slide was then left to incubate for two hours on a plate shaker, followed by three washes with wash buffer. A biotinylated detection antibody in reagent diluent (200 ng/mL) was then added to each well and left to incubate for a further two hours on a plate shaker. Following another set of three washes, streptavidin-HRP was added to each well and left to incubate for 30 min. Once again, the wells were washed three times before the addition of the TMB-blotting solution. The TMB was left for 30 min before being removed from the wells. After removal of TMB, each well was rinsed with doubly distilled deionized water (dddH2O). The slide was then removed from the incubation chamber and dried under nitrogen prior to analysis by Raman spectroscopy. Detection by Raman Spectroscopy. Raman scans were collected using a WITec alpha 300 with a 633 nm laser excitation wavelength. The spectra were centered at 1300 cm−1 and collected using a 100× objective, with an integration time of 0.1 s at a 1 μm spatial resolution. In order to avoid burning of the nitrocellulose surface, the laser power had to be reduced to around 10%, which corresponds to 350 μW. Raman maps were generated based on the peak at 1609 cm−1, as this gave the most intense signal and the WITec Project software was used to calculate the total intensity of spots. This was achieved by selecting each of the spots and using the image statistics function to calculate the total signal intensity in the selected area.

In a previous study, we showed that resonance Raman scattering could be utilized as the detection method in a conventional ELISA to improve sensitivity and introduce the possibility of multiplexed analysis.32 The most commonly utilized enzyme/substrate system is the horseradish peroxidase (HRP)-catalyzed oxidation of 3,3′,5,5′-tetramethylbenzidine (TMB) by hydrogen peroxide (H2O2). We demonstrated that by using a 633 nm laser excitation wavelength, Raman spectra of the blue oxidation product, the charge transfer complex (CTC), were resonantly enhanced. We used this resonance Raman scattering to selectively detect human tumor necrosis factor alpha (TNF-α), with higher sensitivity than the conventional colorimetric detection method. This paper presents an adaptation of this improved detection technique combined with the advantages of DPN for the detection and quantification of prostate-specific antigen (PSA). We report the first use of this novel detection method with DPN, which improves throughput and provides the flexibility for a multiple target assay to be developed in the future.

■

EXPERIMENTAL SECTION Chemicals and Instrumentation. Bovine serine albumin (BSA), phosphate buffered saline (PBS), and Tween 20 were purchased from Sigma Aldrich. A human PSA DuoSet was purchased from R&D Systems Inc.. This contained monoclonal mouse antihuman PSA capture antibody, recombinant human PSA, biotinylated goat antihuman PSA detection antibody, and streptavidin-HRP. A TMB-blotting solution was obtained from Thermo Scientific. Nitrocellulose PATH slides were purchased from Gentel Biosciences and the Nexterion 16-well incubation chamber from Schott. Array printing by DPN was performed on a NLP 2000 nanolithography platform (NanoInk Inc.), using Inkwell arrays (M-6MW) and 12-probe 1D probe arrays (type M-ED). A WITec alpha300 R confocal Raman microscope with 633 nm laser excitation was used for all Raman measurements. Fabrication of PSA Antibody Arrays. Prior to printing the arrays, Type M-ED DPN probe array pens were cleaned using an oxygen plasma at 50% power and 72 cm3/minute for 40 s. The pens were then mounted on the NLP 2000 chip holder in a position that allowed the use of the M-2 end of the probe array. Capture antibody print solution was prepared by reconstituting the lyophilized protein with 5 parts PBS and 3 parts protein carrier buffer (Nanoink Inc.) to give a final antibody concentration of 2 mg/mL. The nitrocellulose slide was placed on the NLP 2000 stage alongside the inkwells containing capture antibody print solution, and the pens were leveled with the surface. Plane calculations were then performed before programming the NLP 2000 software to print the required arrays. The first step of the program was for the pens to contact the perfectly aligned inkwells to ensure that all of the cantilever tips were coated with the PSA capture antibody. Excess ink was subsequently removed by “bleeding” in a position out with the desired printing area. The arrays were then printed, using a 0.1 s dwell time. Once the arrays were fabricated as desired, the slide was incubated overnight at 4 °C. After overnight incubation the slides were washed by shaking in wash buffer (0.05% Tween 20 in PBS), before being placed in the Nexterion IC-16 incubation chamber. PSA Immunoassay. After the washing step described above, the slide was blocked by adding reagent diluent (1% BSA in PBS) to each well and leaving it to incubate for 1 h. Reagent diluent was then removed, and the arrays were washed

■

RESULTS AND DISCUSSION Resonance Raman Detection of PSA Immunoassay Arrays. The schematic diagram shown in Figure 1 represents

Figure 1. Schematic representation of the PSA assay development. The capture antibody is printed using DPN, and the remaining nitrocellulose surface is blocked with BSA followed by the addition of PSA, biotinylated detection antibody, streptavidin-HRP, and TMB to complete the assay. The charge transfer complex (CTC), as shown in here, forms as a blue precipitate around the spots.

the assay format adopted in this work. Arrays of PSA capture antibody were first printed onto the nitrocellulose surface, using the NLP 2000 with the printing conditions previously optimized by Irvine et al.27 Arrays were printed with a 0.1 s dwell time and 22 μm pitch, and each pen was programmed to print 3 × 8 spots. Since the pitch between the pens is 66 μm, this resulted in evenly spaced arrays with a total of 36 × 8 spots in each set. Following the printing process, the remaining surface was blocked with BSA before the addition of PSA, B

dx.doi.org/10.1021/ac4009012 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Letter

and from an area between the spots (red). There is very little background signal from the areas between the spots, and a clear signal for the CTC can be observed from the areas where the arrays were printed. Figure 2 (inset) is a Raman map generated with respect to the band at 1609 cm−1, which clearly illustrates that appreciable Raman scattering was observed from the protein array spots with very little background signal from the surrounding areas. The clear visibility of the arrays is due to the nature of the nitrocellulose surface. The spin-coated layer of nitrocellulose becomes porous upon hydration so that the protein can be absorbed into the matrix while retaining its function.27 This means that when the oxidized TMB subsequently precipitates around the spots, the colored product is held within this region and the Raman spectra can be obtained from the area where the spots were printed. The three main peaks in the spectrum are those arising from the blue CTC, which are resonance enhanced using 633 nm laser excitation, while three other bands, which were also observed in our previous study (1103, 1413, and 1436 cm−1), are much less prominent. Since the formulation of TMB used here causes the precipitation of the blue CTC, this oxidation product is isolated before being analyzed by resonance Raman scattering. This differs from the formulation used in an ELISA, where the reaction takes place in solution and there is a possibility of starting material and other products being leftover. Since the isolation of this blue CTC results in the relative intensity of these bands being reduced, it supports our previous conclusion that they could be attributed to the parent diamine leftover unoxidized in solution.32 Varying Concentration of PSA for Limit of Detection Study. To determine whether or not quantification of the analyte is possible using this method, we studied the assay over a concentration range (0−60 ng/mL). With the use of a Nexterion incubation chamber, the nitrocellulose slide could be split into 16 wells. In each of the wells, the PSA capture antibody arrays of 36 × 8 spots were printed, and the immunoassay was carried out on a standard dilution series of the PSA standard. The assay was carried out in duplicate and

biotinylated detection antibody, and streptavidin-HRP. The final step was the addition of the TMB-blotting solution, a commercially available formulation of TMB, which precipitates upon reaction with HRP to form the blue oxidation product in localized areas around the spots. This differs from the TMB used in an ELISA, where the reaction occurs in solution and is therefore more suitable for immunoassays on a surface. Jenison et al. used this precipitating TMB in a biosensor for the detection of multiple proteins and nucleic acids in an array format.33 However, they used large array spots, around 600 μm in diameter, and their sensitivity was limited by the colorimetric detection. Here, on the other hand, we can print miniaturized microscale arrays, allowing for higher throughput, and we can obtain resonance Raman spectra of the blue charge-transfer complex at 633 nm, thus allowing us to selectively detect the analyte by resonance Raman mapping of the arrayed surface. Figure 2 shows the Raman spectra obtained from a spot (blue)

Figure 2. Raman spectra obtained from a point within an array spot (blue) and between the spots (red). The inset shows a Raman map from the arrays containing 7.5 ng/mL of PSA. The map was generated with respect to the intensity of the band at 1609 cm−1.

Figure 3. (a−d) Raman maps of the arrays containing 3.75, 1.88, 0.94, and 0 ng/mL, respectively. (e) Plot of concentration of PSA against intensity, where the intensity is the sum intensity of the overall spot, and the plotted value is the mean of three spots from each of three different maps, where the error bars represent the standard deviation. (f) Plot of PSA concentration against area of spot. C

dx.doi.org/10.1021/ac4009012 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Letter

nanoscale may not be achievable with standard instruments. Therefore, features of this size overcome the disadvantages of nanoscale printing, while still offering the miniaturization of typical micrometer scale arrays (100 −200 μm).18,27

the reagent diluent was added in place of PSA for the blank control. After completion of the assay and development of the TMB, the slide could be analyzed using Raman mapping and 633 nm laser excitation. For each concentration three 50 × 50 μm maps were collected from each of the duplicate assays, and three spots from each map were analyzed. The sum spot intensity was calculated by selecting each spot and using the WITec Project software to calculate the total signal intensity over the area of this spot. Since the maps were generated with respect to the band at 1609 cm−1, it is the intensity of this band that the false color image corresponds to, where yellow indicates areas of high intensity through to black, where no signal is present. The Raman maps in Figure 3 show that, as well as a decrease in the Raman signal intensity, the size of the array spots also decrease with decreasing concentration of PSA. This reduction in spot size can be attributed to the fact that, when there is less PSA present, there will be less HRP bound in the assay and, therefore, a reduction in the amount of oxidized TMB, which will precipitate around the spots. This is not something that would occur with other methods where the label is incorporated into the assay but is observed here as a result of the enzymatic reaction and the precipitating TMB. As shown in Figure 3e, the decrease in the Raman signal intensity is linear across the whole concentration range (0.94−60 ng/ mL), contrary to what was found in the fluorescent assay where the signal plateaued above 15 ng/mL.27 The authors attributed this to the saturation of the PSA binding to the capture antibody above this point, which limits the number of fluorophores available for detection. In enzyme-based assays like the type described here, the detection is less limited by the number of antibodies and, hence, the number of HRP molecules, since a single enzyme catalyzes the conversion of many substrate molecules. In addition, the use of a solid product in this assay isolates it from the solution containing other reagents, making for interference-free detection. As a result, the resonance Raman assay has a larger dynamic range. The linear relationship between the concentration of PSA and the intensity of the Raman signal enables quantification. The detection limit of the assay was calculated by multiplying the standard deviation of the blank control by three and dividing the resulting number by the slope of the curve. The calculated detection limit was found to be 25 pg/mL, which is an improvement when compared to the same assay with fluorescence detection (LDL = 98 pg/mL), and also in comparison to the standard ELISA which can be purchased from R & D Systems (LDL = 69 pg/mL). The calculated limit is well below normal levels found in blood, 4−10 ng/mL; however, it has been suggested that such ultrasensitive detection may be useful for diagnosing prostate cancer in the early stages of postsurgery recurrence and also for the detection of breast cancer in women.34 An excellent level of quantification can be observed here, which is particularly impressive for the detection of a precipitate that is produced via these miniaturized reactions. This further proves the efficiency of resonance Raman scattering as the detection method in this type of assay. The diameter of the array spots increased with increasing PSA concentration across the range tested (Figure 3f); however, the average diameter ranged from around 2 μm to around 5 μm, which is significantly smaller than arrays printed with other methods and is also considered to be in the idealized range (1−10 μm) for protein microarrays. This is of particular significance when using optical detection techniques such as Raman spectroscopy, where detection of arrays on the

■

CONCLUSIONS The sensitivity of resonance Raman spectroscopy has been successfully combined with the flexibility and accuracy of DPN to produce an immunoassay with a long dynamic range and low detection limits. The detection limit achieved is lower than those obtained with similar assay formats, which use fluorescence detection. The use of a colored immobilized microscale spot of an ideal size for protein array detection has proved very effective for assays using DPN and resonance Raman scattering. This paper focused on the detection of PSA, but the methods presented herein could be applied to a wide variety of biomarkers. Since the format of the method is adapted from a commercially available ELISA, its application is simple and straightforward, while possessing many advantages over commonly employed methods. The multipen printing method utilized allows for quick and efficient array fabrication but also means that many different “inks” can be printed simultaneously, giving rise to the possibility of effective multiplexed analysis.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; Fax: +44 (0)141 548 4787; Tel: +44 (0) 141 548 4701. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

The authors acknowledge the Analytical Chemistry Trust Fund, EPSRC for the award of an analytical studentship which supported and funded this work. D.G. acknowledges the Royal Society of Chemistry for the Wolfson Research Merit award.

(1) Banks, R. E.; Dunn, M. J.; Hochstrasser, D. F.; Sanchez, J. C.; Blackstock, W.; Pappin, D. J.; Selby, P. J. Lancet 2000, 356 (9243), 1749−1756. (2) Gul, O.; Calay, E.; Sezerman, U.; Basaga, H.; Gurbuz, Y. Sens. Actuators, B 2007, 125 (2), 581−588. (3) Elshal, M. F.; McCoy, J. P. Methods 2006, 38 (4), 317−323. (4) Silzel, J. W.; Cercek, B.; Dodson, C.; Tsay, T.; Obremski, R. J. Clin. Chem. 1998, 44 (9), 2036−2043. (5) Mendoza, L. G.; McQuary, P.; Mongan, A.; Gangadharan, R.; Brignac, S.; Eggers, M. BioTechniques 1999, 27 (4), 778−788. (6) Joos, T. O.; Schrenk, M.; Hopfl, P.; Kroger, K.; Chowdhury, U.; Stoll, D.; Schorner, D.; Durr, M.; Herick, K.; Rupp, S.; Sohn, K.; Hammerle, H. Electrophoresis 2000, 21 (13), 2641−2650. (7) Moody, M. D.; Van Arsdell, S. W.; Murphy, K. P.; Orencole, S. F.; Burns, C. BioTechniques 2001, 31 (1), 186−190. (8) Lee, S.; Ko, Y. H.; Jung, H.; Kim, J. D.; Song, J. M.; Chooe, J.; Eo, S. K.; Kang, S. H. Talanta 2009, 78 (2), 608−612. (9) Karoonuthaisiri, N.; Charlermroj, R.; Uawisetwathana, U.; Luxananil, P.; Kirtikara, K.; Gajanandana, O. Biosens. Bioelectron. 2009, 24 (6), 1641−1648. (10) Huang, R. P.; Huang, R. C.; Fan, Y.; Lin, Y. Anal. Biochem. 2001, 294 (1), 55−62. (11) Haab, B. B.; Dunham, M. J.; Brown, P. O. Genome Biol. 2001, 2 (2), RESEARCH0004.

D

dx.doi.org/10.1021/ac4009012 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Letter

(12) Schweitzer, B.; Roberts, S.; Grimwade, B.; Shao, W. P.; Wang, M. J.; Fu, Q.; Shu, Q. P.; Laroche, I.; Zhou, Z. M.; Tchernev, V. T.; Christiansen, J.; Velleca, M.; Kingsmore, S. F. Nat. Biotechnol. 2002, 20 (4), 359−365. (13) Wiese, R.; Belosludtsev, Y.; Powdrill, T.; Thompson, P.; Hogan, M. Clin. Chem. 2001, 47 (8), 1451−1457. (14) Ellmark, P.; Ghatnekar-Nilsson, S.; Meister, A.; Heinzelmann, H.; Montelius, L.; Wingren, C.; Borrebaeck, C. A. K. Proteomics 2009, 9 (24), 5406−5413. (15) Lee, K. B.; Kim, E. Y.; Mirkin, C. A.; Wolinsky, S. M. Nano Lett. 2004, 4 (10), 1869−1872. (16) Lynch, M.; Mosher, C.; Huff, J.; Nettikadan, S.; Johnson, J.; Henderson, E. Proteomics 2004, 4 (6), 1695−1702. (17) Wingren, C.; Borrebaeck, C. A. K. Drug Discovery Today 2007, 12, 813−819. (18) Borrebaeck, C. A. K.; Wingren, C. J. Proteomics 2009, 72 (6), 928−935. (19) Patwa, T. H.; Li, C.; Poisson, L. M.; Kim, H. Y.; Pal, M.; Ghosh, D.; Simeone, D. M.; Lubman, D. M. Electrophoresis 2009, 30 (12), 2215−2226. (20) Ng, H. T.; Fang, A. P.; Huang, L. Q.; Li, S. F. Y. Langmuir 2002, 18 (16), 6324−6329. (21) Jun, Y.; Cha, T.; Guo, A.; Zhu, X. Y. Biomaterials 2004, 25 (17), 3503−3509. (22) Biebricher, A.; Paul, A.; Tinnefeld, P.; Golzhauser, A.; Sauer, M. J. Biotechnol. 2004, 112 (1−2), 97−107. (23) Lee, K. B.; Park, S. J.; Mirkin, C. A.; Smith, J. C.; Mrksich, M. Science 2002, 295 (5560), 1702−1705. (24) Haaheim, J.; Eby, R.; Nelson, M.; Fragala, J.; Rosner, B.; Zhang, H.; Athas, G. Ultramicroscopy 2005, 103 (2), 117−132. (25) Lee, K. B.; Lim, J. H.; Mirkin, C. A. J. Am. Chem. Soc. 2003, 125 (19), 5588−5589. (26) Wu, C. C.; Xu, H. P.; Otto, C.; Reinhoudt, D. N.; Lammertink, R. G. H.; Huskens, J.; Subramaniam, V.; Velders, A. H. J. Am. Chem. Soc. 2009, 131 (22), 7526−7527. (27) Irvine, E. J.; Hernandez-Santana, A.; Faulds, K.; Graham, D. Analyst 2011, 136 (14), 2925−2930. (28) Lee, S. W.; Oh, B. K.; Sanedrin, R. G.; Salaita, K.; Fujigaya, T.; Mirkin, C. A. Adv. Mater. 2006, 18 (9), 1133−1136. (29) Ginger, D. S.; Zhang, H.; Mirkin, C. A. Angew. Chem., Int. Ed. 2004, 43 (1), 30−45. (30) Lim, J. H.; Ginger, D. S.; Lee, K. B.; Heo, J.; Nam, J. M.; Mirkin, C. A. Angew. Chem., Int. Ed. 2003, 42 (20), 2309−2312. (31) Nielsen, U. B.; Geierstanger, B. H. J. Immunol. Methods 2004, 290 (1−2), 107−120. (32) Laing, S.; Hernandez-Santana, A.; Sassmannshausen, J.; Asquith, D. L.; McInnes, I. B.; Faulds, K.; Graham, D. Anal. Chem. 2011, 83 (1), 297−302. (33) Jenison, R.; La, H.; Haeberli, A.; Ostroff, R.; Polisky, B. Clin. Chem. 2001, 47 (10), 1894−1900. (34) Grubisha, D. S.; Lipert, R. J.; Park, H. Y.; Driskell, J.; Porter, M. D. Anal. Chem. 2003, 75 (21), 5936−5943.

E

dx.doi.org/10.1021/ac4009012 | Anal. Chem. XXXX, XXX, XXX−XXX