SERS-based immunoassay using a gold array-embedded gradient microfluidic chip

Abstract

Here we report the development of a programmable and fully automatic gold array-embedded gradient microfluidic chip that integrates a gradient microfluidic device with gold-patterned microarray wells. This device provides a convenient and reproducible surface-enhanced Raman scattering (SERS)-based immunoassay platform for cancer biomarkers. We used hollow gold nanospheres (HGNs) as SERS agents because of their highly sensitive and reproducible characteristics. The utility of this platform was demonstrated by the quantitative immunoassay of alpha-fetoprotein (AFP) model protein marker. Our proposed SERS-based immunoassay platform has many advantages over other previously reported SERS immunoassay methods. The tedious manual dilution process of repetitive pipetting and inaccurate dilution is eliminated with this process because various concentrations of biomarker are automatically generated by microfluidic gradient generators with N cascade-mixing stages. The total assay time from serial dilution to SERS detection takes less than 60 min because all of the experimental conditions for the formation and detection of immunocomplexes can be automatically controlled inside the exquisitely designed microfluidic channel. Thus, this novel SERS-based microfluidic assay technique is expected to be a powerful clinical tool for fast and sensitive cancer marker detection.

---

Introduction

Immunoassays are a well-established method that can be applied to biochemical studies, clinical diagnostics, food safety tests and environmental monitoring.^1–3 The enzyme-linked immunosorbent assay (ELISA) has been most widely used as a screening tool for validating candidate protein markers.⁴ This technique allows for the simultaneous estimation of the concentrations of numerous biomarkers in a small volume. It has several significant advantages over other conventional assay methods, such as a high-throughput performance, cost-effectiveness and facile clinical applications. Nonetheless, this assay technique has some technical limitations. To identify a specific antigen–antibody interaction in an immunoassay, fluorescence detection using molecular labels is commonly used, but this has several drawbacks including a poor limit of detection, photobleaching, and a limited multiplex detection capability.^5–7

The surface-enhanced Raman scattering (SERS)-based immunoassay, using antibody-conjugated metal nanoparticles, is considered a promising alternative because of its rapid and sensitive sensing capability.^8–16 One of the most popular technologies for the SERS-based immunoassay utilizes the typical protocol of a sandwich immunocomplex platform. Here, primary antibodies are immobilized on a substrate and antigens and secondary antibody-conjugated Raman probes are sequentially added. After the formation of a sandwich immunocomplex on a substrate, SERS signals are measured and analyzed. Various kinds of biomarkers including immunoglobulin (IgG) antigens,^17–19 feline calicivirus (FCV),²⁰ prostate-specific antigen (PSA),²¹ protein A,²² hepatitis B virus,²³ and mucin protein (MUC4)²⁴ have been investigated using this SERS-based immunoassay technique.

Recently, our research group also developed a novel SERS-based immunoassay platform using a gold-patterned microarray chip and hollow gold nanospheres (HGNs).²⁵ Here, a hybrid microarray pattern, including hydrophilic gold wells and other hydrophobic areas, was fabricated and used as a detection platform for the SERS-based immunoassay. HGNs were used as SERS agents because of their capability to localize the electromagnetic fields through the pinholes in the hollow particle structures.^26–28 Consequently, HGNs are highly sensitive and reproducible SERS probes for the quantitative immunoanalysis of target markers. This approach does not require an expensive arrayer which is essential for the conventional microarray-based assay because a hydrophilic target sample and HGNs are automatically arranged on the surface of gold-patterned microwells. In addition, the limit of detection (LOD) and dynamic range for detectable concentrations are greatly improved compared with those of the conventional ELISA.

Nonetheless, this SERS-based immunoassay using a gold-patterned microarray chip has some technical problems. Firstly, it is awkward to optimize the homogeneous distribution of target biomarkers on gold-patterned wells because all the sequential antibody/antigen immobilization and washing steps are difficult to manually control using a micropipette. Secondly, all of the immunoreagents should be immobilized on the surface of the gold substrates in air. In many cases, however, the exposure of proteins to air seriously reduces their biological activity. Finally, the repetition of washing steps to remove non-specific binding proteins makes this immobilization-based assay technique inconvenient. Thus, this time-consuming and manually controlled process makes the SERS-based gold-patterned microarray platform less attractive. To resolve these problems, we implemented the gold-patterned microarray wells in a gradient microfluidic platform. Recently, SERS-based optofluidic sensors have been extensively used for the sensitive analysis of various biological targets.^29–38 Using microfluidics, it is possible to automatically control all the sample distribution and washing steps. In addition, a highly sensitive and reproducible SERS-based analysis is possible if a continuous flow and homogeneous mixing conditions are maintained in a microfluidic channel.

Here, we report the development of a programmable and fully automatic gold array-embedded gradient microfluidic chip that integrates the gradient microfluidic device with the gold-patterned microarray wells. A serial dilution of antigen marker can be achieved in a stepwise manner using microfluidic concentration gradient generators with N cascade-mixing stages.^39–41 In this channel, desired concentrations can be achieved by means of controlled volumetric mixing ratios of two merging solutions in each stage. Among three different types of microfluidic gradient generators for linear, logarithmic and Gaussian gradients, the two-fold logarithmic gradient was selected in this work in order to generate a wide dynamic concentration range for immunoassay. In addition, 30 μm gold-patterned microarray wells (5 × 5) were embedded on a glass substrate for the immobilization of antibodies on the surface. By using this novel gold microarray-embedded gradient microfluidic channel, the tedious manual dilution process can be avoided and highly accurate immunoanalysis could be achieved with automatic flow controls.

The utility of this platform was demonstrated by the quantitative immunoassay of the alpha-fetoprotein (AFP) model protein marker. AFP is a well-known indicator of tumor activity in hepatocellular carcinoma (HCC).⁴² Here, the total assay time including incubation, washing steps and detection was less than 60 min. To the best of our knowledge, this is the first report about a SERS-based optofluidic immune-sensor using a gold array-embedded gradient channel.

Experimental

Chemicals and reagents

Silicon wafers (p-type, 100 orientation) were purchased from Silicon Sense. The SU-8 photoresist and developer were purchased from Microchem. The AZ1512 photoresist and its developer were purchased from AZ Electronic Materials. Poly(dimethylsiloxane) was obtained from Dow Corning (Sylgard 184). Gold etchant was purchased from Duksan chemical. Gold chloride, sodium borohydride, 11-mercaptoundecanoic acid (11MUA), 6-mercapto-1-hexanol (6MHA), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), malachite green isothiocyanate (MGITC), dihydrolipoic acid (DHLA), Tween 20, mercaptoethanol, glycerol, phosphate buffered saline (PBS, pH 7.2) and bovine serum albumin (BSA) were purchased from Sigma (St. Louis, MO, USA). Alpha 1 fetoprotein (AFP) antigen, anti-AFP monoclonal antibody (anti-AFP McAb) and anti-AFP polyclonal antibody (anti-AFP PcAb) were purchased from Abcam (Cambridge, MA, USA). All aqueous solutions were prepared using deionized water (DIW, 18 MΩ) obtained from a Milli-Q system (Millipore S.A., Bedford, USA).

Fluorescence and SERS measurements

Fluorescence measurements were performed using an Olympus IX81 inverted research microscope system (Olympus Co., Japan). Precision syringe pumps (PHD 2000, Harvard Apparatus, USA) and 1 mL Norm-Ject plastic syringes (Henke-Sass Wolf GmbH, Germany) were used to inject the sample into the gradient microfluidic chip. SERS measurements were performed using a Renishaw 2000 confocal Raman microscope system (Renishaw, UK). A Spectra Physics He–Ne laser (Research Electro-Optics, Inc., USA) operating at λ = 633 nm was used as the excitation source with a laser power of approximately 12 mW. The Rayleigh line was removed from the collected Raman scattering using a holographic notch filter located in the collection path. Raman scattering was collected using a charge-coupled device (CCD) camera at a spectral resolution of 4 cm⁻¹.

Fabrication of a gold array-embedded gradient microfluidic chip

The device consisted of three panels as shown in Fig. 1. The top and middle PDMS panels (Fig. 1(a) and 1(b)) of the microfluidic circuits were fabricated by soft lithography. The bottom glass panel (Fig. 1(c)), patterned with gold microarray wells, was fabricated by a photolithographic technique. A master wafer, patterned with SU-8 photoresist, was used to fabricate the microfluidic channels. The dimensions of the channels were 100 μm in height and 150 μm in width.

Fig. 1 Schematic illustration of a gold array-embedded gradient chip consisting of three layers. (a) Top layer: PDMS panel for the uniform distribution of antibody-conjugated HGNs; (b) middle layer: PDMA gradient panel for the generation of various concentrations of a cancer marker; and (c) bottom layer: gold array-embedded glass substrate for the immobilization of sandwich immunocomplexes. (d) Photograph of an integrated gold array embedded gradient chip, and (e) photograph of the channel filled with different ink solutions (yellow, blue and red) injected through three inlets.

The gold array-patterned glass substrate was prepared using a method previously reported by our group.²⁵ First, the glass substrate was cleaned using surface-cleaning solutions, SPM (H₂SO₄ : H₂O₂ = 4 : 1) and SC-1 (NH₄OH : H₂O₂ : DI water = 1 : 1 : 5). Second, titanium metal was sputtered onto the clean glass slide, followed by gold deposition. Third, the photoresist (AZ1512, AZ-EM, USA) was spin-coated at 3000 rpm for 30 s. The thickness of the coated photo-resist was estimated to be 1 μm after being baked on a hot plate at 100 °C for 1 min. A UV aligner (EVG620, EVG, Austria) and a developer (CEE100FX, CEE, USA) were used to control the arrayed well patterns. The photo-resist was exposed to the broadband wavelength (365 < λ < 436 nm) at 50 mJ cm⁻² and then it was treated with developer solution (300MIF, AZ-EM, USA) for 1 min. Fourth, the gold thin film was etched to form array patterns. The pattern consisted of 5 × 5 wells, each with a diameter of 30 μm. The array-patterned gold films on the glass layer were modified with 10 mM of THF solution containing 11MUA and 6HMA (1 : 10) for 24 h. Here, the carboxylate-terminated self-assembled monolayer (SAM) was used for the specific conjugation of capture antibodies. Finally, the surface was washed with ethanol and deionized water, and then heated at 120 °C for 15 min. Irreversible bonding between the top PDMS layer and the middle PDMS layer, and between the middle PDMS layer and the bottom glass layer was achieved by plasma treatment of the interfaces of the respective layers.

Concentration validation using a fluorescence microscopy

A fluorescence microscope was used to evaluate the mixing and dilution performance of a gold array-embedded gradient microfluidic chip. Two inlets (A and B) of the chip were interconnected by a syringe pump (Harvard Apparatus Model PHD2000 Syringe Pump Series, MA, USA) using disposable plastic microbore tubes while the six outlets were interconnected to the waste reservoir using the same plastic tubes. A 3.0 × 10⁻⁴ M fluorescein sodium salt solution was used to visualize the dilution process of the two confluent streams. Aqueous fluorescein solution was pumped into one inlet and distilled water was pumped into the other two inlets, respectively. The traveling time for the generation of two-fold logarithmic distributions was estimated to be 2.6 min. Images of the fluid dilution in the microfluidic channel were acquired and analyzed using the Q-imaging software package. For a quantitative analysis of the dilution of the confluent streams, the standard deviation for the pixel intensity distribution was calculated from the cross-section of the channel.

Preparation of antibody-conjugated HGNs

HGNs were prepared using a previously reported method.^26–28 Cobalt nanoparticles were synthesized by reducing CoCl₂ with NaBH₄ under N₂ purging conditions and were used as templates for the synthesis of HGNs. The diameter and wall thickness of the HGNs were estimated to be 45 ± 12 and 15 ± 5 nm, respectively. To use these HGNs as SERS-active probes, a Raman reporter, MGITC, was adsorbed onto their surface. Then dihydrolipoic acid (DHLA) was used to conjugate the antibodies. Two –SH terminal groups of DHLA were cleaved and chemically bonded to the HGN surfaces. To activate the –COOH terminal groups, EDC and NHS were added and allowed to react. Finally, polyclonal anti-AFP antibody was added to the NHS-activated HGNs and reacted. Antibody immobilization on the surfaces of the HGNs was achieved by displacing the NHS group to the lysine residues, which contain an amine group in the antibody.

Immunoassay procedure using the gold array-embedded gradient microfluidic chip

Immunoassay was performed by a typical sandwich AFP immune-complex formation on gold array wells embedded in a gradient microfluidic chip. First, whole channels were filled with buffer solution using three micro-syringe pumps. For the activation of the carboxyl acid-immobilized gold array patterns, a mixture of 20 mM EDC and 50 mM NHS was injected through an inlet C (Fig. 2) at a flow rate of 3.1 μL min⁻¹ for 30 min. PBS solution was injected into the same inlet for 5 min to remove any non-specifically bound chemicals. Then, 10 μg mL⁻¹ anti-AFP McAbs in PBS solution was injected into inlet C for 15 min to immobilize them on the NHS-activated gold arrays in the channel, followed by incubation for 30 min. After immobilization of anti-AFP antibodies on the gold-patterned microarray wells, 1% BSA solution was injected through inlet C at a rate of 3.1 μL min⁻¹ for 30 min to block any unbound sites in the gold array wells. Then, 50 μg mL⁻¹ of anti-AFP antigen in PBS solution was injected into inlet B at a rate of 1.0 μL min⁻¹ and PBS solution was injected into inlet A at a rate of 2.1 μL min⁻¹. The chip was subsequently incubated for 30 min. PBS solution was injected into inlets A and B for 5 min to remove any unbound AFP antigens. Then, 0.7 nM of anti-AFP PcAb-conjugated MGITC-HGNs was injected into inlet C at a rate of 3.1 μL min⁻¹ for 15 min to generate sandwich immune-complexes. Raman spectra of the sandwich immunocomplexes on gold arrays were obtained using a 50 × long working distance objective lens in confocal mode. All of the Raman spectra were collected for a ten-second exposure time in the range of 800–1800 cm⁻¹. The SERS images of each chamber were decoded for the peak area of 1560–1650 cm⁻¹ using WiRE software V 1.3 (Renishaw, UK).

Fig. 2 Layout of a gold array-embedded gradient chip for the SERS-based immunoassay. The illustrations in the enlarged circles represent the formation of sandwich immunocomplexes on the surface of 5 × 5 round gold wells embedded in the gradient channel.

Results and discussion

Function of three layers in the gold array-embedded gradient chip

Fig. 1 illustrates the design of the gold array-embedded gradient chip for the SERS-based immunoassay. This SERS-based gradient chip consists of three layers. The top layer is a PDMS panel for the injection of capture antibodies, blocking solutions and antibody-conjugated nanoparticle colloids (Fig. 1(a)). Here, the injected flow split into six streams of the same concentration and they are connected to the middle layer through six holes. The middle layer is a PDMS panel for the generation of antigen gradients with six diluted concentrations (Fig. 1(b)). This layer was designed by modifying our previous serial dilution module including N cascaded-mixing stages.^39,40 Various concentrations of antigen solutions can be automatically generated using a network of gradient channels. As previously reported by Dertinger et al.⁴³ the branching structure of the channel serially dilutes one stream with a second stream. Consequently, the laminar flow of fluids inside the channel permits multiple streams of solutions containing different concentrations of flow side-by-side. By adjusting the compositions of individual streams, various concentration profiles can be generated. This technique replaces the set of microwells or microarrays used for manual dilution which is time-consuming and sometimes leads to serious experimental errors during the preparation of different concentrations of target samples. In this work, the second gradient layer automatically produces a series of different concentration gradients with a two-fold logarithmic dilution by continuously mixing and diluting the antigen with a buffer solution. The bottom layer is a glass substrate embedded with six gold-patterned microarray sets (Fig. 1(c)). The gold films were etched for the formation of array patterns on glass substrates. Three layers were aligned and bonded by oxygen plasma treatment. Fig. 1(d) shows an integrated gold array-embedded gradient chip composed of three separate panels. Fig. 1(e) shows a real photograph of the microfluidic channel filled with three different colors of ink solution.

Fig. 2 shows a detailed schematic layout of the gradient microfluidic channel integrated with six embedded gold microarrays. As shown in this figure, the gold array pattern consisted of 5 × 5 round wells, each with a diameter of 30 μm. First, gold array wells were modified with 10 mM of THF solution containing 11-mercaptoundecanoic acid and 6-mercapto-1-hexanol (1 : 10) for 24 h. This is to create carboxylate-terminated self-assembled monolayers for the formation of a hydrophilic surface. Anti-AFP capture antibodies were immobilized on the hydrophilic surface of gold array wells, and anti-AFP antigens (cancer markers) and anti-AFP PcAb conjugated MGITC-HGNs (functional nanoprobes) were sequentially injected, flowed down, and attached onto the gold patterned microarray wells for the formation of sandwich immunocomplexes as shown in Fig. 2. Three different types of hot junctions can be generated on the surface of each well. First, HGNs have strong enhancement effects from individual particles because hot junctions are localized at the pinholes in the hollow particle structures. Second hot junctions may emerge among HGN aggregates formed on the gold array-patterned plane substrate. Finally, hot spots emerge from the edges between nanoparticles and the plane well substrate.

Evaluation of mixing behaviors in a gold-embedded gradient microfluidic channel using fluorescence microscopy

To evaluate the dilution efficiency of generating accurate concentrations of a marker in the gradient channels, visualization experiments using fluorescence microscopy were performed. Fig. 3 demonstrates the result of a serial dilution test using green fluorescein solution. To find an optimal mixing condition for the fluorescence measurements, the flow rate was varied from 0.1 μL min⁻¹ to 50 μL min⁻¹ by regulating the micro-syringe pump. According to our experiments, the flow rate conditions of 10.5 μL min⁻¹ (for AFP antigen and deionized water) and 5 μL min⁻¹ (for nanoprobes) were optimized for the initial fill-up process with distilled water and fluorescence solution. The inlet for nanoprobes was closed to avoid reverse flow after the channels were filled with deionized water. Once the channels were filled up with confluent stream mixtures, the flow rates were slow down by controlling syringe pumps (2.1 μL min⁻¹ for the AFP antigen inlet; 1 μL min⁻¹ for the water inlet) to obtain the desired concentrations. When confluents travel along the channel, they are split at the nodes, combined with neighboring streams, and allowed to mix by diffusion. Two-fold logarithmic gradients were generated with the single serial dilution module.

Fig. 3 (a) Fluorescence image of six outlet channels demonstrating the serial dilution of the fluorescein sodium salt dye. (b) Calibration curve for a series of diluted solutions generated by the gradient channel. The R² value for two-fold logarithmic concentrations is 0.998.

In this experiment, the fluorescence signal was measured 10 min after the initial injection of fluids to achieve a steady state of equilibrium gradients. Fig. 3(a) displays the fluorescence images of six outlet channels. The fluorescence intensity gradually increased from left to right along with the increase in the fluorescein concentration. Fig. 3(b) shows experimental results for the fluorescence measurements. The fluorescence intensities in each channel were estimated to be 0, 6.25, 12.5, 25, 50, and 100%, and this distribution of fluorescence intensities reveals the linear gradient of precise two-fold logarithmic concentrations. Thus, it is possible to automatically generate a series of solutions containing desired concentrations of antigens using this gradient microfluidic device. The design achieves 1 : 1 mixing of antigen and buffer in each stage of dilution. Thus, each stage decreases the concentration of antigen by one-half. Thus, the tedious manual dilution process is removed because various concentrations of target marker are generated by a network gradient channel.

SERS-based immunoassay in a gold array-embedded gradient microfluidic channel

The gold array-embedded gradient chip was used for on-chip immunoassay of lung cancer marker AFP. Fig. 4 shows the sequential steps required for the formation of sandwich AFP immunocomplexes for SERS detection. All of the assay processes were performed under continuous flow conditions, and the sequential flow time was 45 min. A sandwich-type immunoassay of AFP using SERS detection was performed through the following steps - (a) Primary antibodies were immobilized on the gold surface. The gold surfaces, modified with carboxylic acids, were activated using EDC/NHS solution. Then monoclonal anti-AFP antibodies were immobilized on the activated gold surface. After immobilization of the antibodies, unreacted sites on the gold microarray well surfaces were blocked using 1% BSA solution. (b) Various concentrations of AFP antigen were generated using a microfluidic device, and they were captured by the immobilized antibodies. Six different concentrations of AFP antigen solutions (0, 0.625, 1.25, 2.5, 5, and 10 ng mL⁻¹) were generated by injecting PBS buffer and AFP into inlets A and B, respectively. (c) The gold arrays were washed with buffer to remove any unbound antigen. Then, polyclonal anti-AFP antibody-conjugated HGNs were attached to form sandwich immunocomplexes. (d) SERS signals were measured for each array well on the channel.

Fig. 4 Sequential steps for the formation of sandwich AFP immunocomplexes for SERS detection. (a) Immobilization of monoclonal antibodies, (b) capturing of AFP antigens, (c) formation of HGN-binding immunocomplexes, and (d) SERS detection of the sandwich immunocomplexes on gold-embedded wells.

Under continuous flow conditions, the SERS spectra indicated a concentration gradient trend but the signal intensities were too low to perform the quantitative analysis. In other words, continuous flow conditions are not sufficient for forming high density immunocomplexes on gold patterned arrays. To resolve this problem, we adopted an incubation step with a static flow condition to give sufficient time for immunocomplex formation. First, each solution was continuously flowed away for 15 min and stopped for 30 min to induce more efficient immuno-reactions. In this case, the intensity of the SERS spectra was eight times stronger than that of continuous flow conditions. Incubation provides enough time for an increase in the loading density of immunocomplexes. In each reaction chamber, SERS spectra were measured from the identical patterns of the inside nine spots among the 5 × 5 arrays.

The SERS signal was measured under the steady-state condition by moving the laser spot to each reaction chamber. In this experiment, a two-slit confocal arrangement was used to reduce the background Raman scattering from the unfocused laser beams. In the Raman system, the function of the pinhole was replaced by the cooperation of the entrance slit and the pixels in the CCD detector. The first confocal slit was set to a width of 15 mm. The signal was then collected from only two pixel rows on the CCD detector, creating a virtual second slit that was aligned perpendicular to the spectrometer slit. The stray background light, due to any out-of-focus regions, was effectively removed using the confocal technique. Experiments were conducted to investigate the feasibility of confocal SERS as a sensitive detection tool for the evaluation of the SERS signals of Raman reporter molecules adsorbed onto the surface of HGNs. Since the SERS signal intensities of the Raman reporter (MGITC) molecules are dominant over those measured from other chemical constituents or antibodies, their variations can be used as quantitative indicators for the immunoassay of AFP.

Fig. 5(a) shows the SERS spectra collected from sandwich immunocomplexes for six different gold patterned microarrays in each channel after employing the static flow condition. Each Raman spectrum displays the average from nine accumulated spectra. The 3 × 3 patterned images on the right side of each spectrum are color-decoded images for the range of 1560–1650 cm⁻¹. Central 3 × 3 areas out of 5 × 5 microarray wells were used for the quantitative analysis. In the absence of AFP antigen, a weak SERS signal was observed (0 ng mL⁻¹). This indicates that a small number of HGNs remained in the solution as a result of nonspecific binding, even though most of the HGNs were removed from the solution by washing. However, this factor was considered in the calibration process. The intensity of the Raman peaks increased concomitantly with the increase in the antigen concentration. The Raman peak areas in the range of 1560–1650 cm⁻¹ were used for quantitative evaluation of the AFP antigen. The calibration curve is shown in Fig. 5(b), where the error bars indicate standard deviations from nine measurements of different spots. A good linear response was achieved within the concentration range of 0 to 10 ng mL⁻¹. This means that highly sensitive and reproducible immunoassays in a microfluidic channel are possible using a gold array-embedded gradient chip and SERS detection. Tedious manual procedures such as repetitive pipetting and inaccurate dilution can be avoided by applying the gradient micro-network channels. Moreover, the assay time was greatly reduced compared with traditional ELISA-based assays.

Fig. 5 (a) SERS spectra for decreasing concentrations of AFP: (1) 0 ng mL⁻¹, (2) 0.625 ng mL⁻¹, (3) 1.25 ng mL⁻¹, (4) 2.5 ng mL⁻¹, (5) 5.0 ng mL⁻¹ and (6) 10 ng mL⁻¹. Each SERS spectrum was averaged from the signals for nine gold array wells embedded in six different channel outlets. The 3 × 3 patterned images on the right side are the color decoded Raman peak images in the 1560–1650 cm⁻¹.range. (b) Corresponding intensity change of the SERS signal for different concentrations of AFP antigen. The graph reveals a linear relationship over the whole concentration range from 0 to 10 ng mL⁻¹ (coefficient of determination, R² value = 0.978). The error bars indicate standard deviations from nine measurements for each concentration.

Conclusions

In the present study, a programmable and fully automatic microfluidic device was designed by combining a gradient microfluidic device with gold-patterned microarray wells. This gold-embedded gradient microfluidic channel provides a convenient and reproducible SERS-based immunoassay platform for cancer biomarkers. The potential immunoanalytical capability of this device was evaluated for AFP antigen. A serial dilution of AFP antigen marker was automatically generated in a stepwise manner by the gradient generators including N-cascade-mixing stages, and sandwich immunocomplexes were successfully formed on gold patterned microarray wells via sequential immobilization steps. Quantitative analysis was performed by measuring the SERS peak areas in the range of 1560–1650 cm⁻¹, and a good linear response was obtained within the concentration range of 0 to 10 ng mL⁻¹. The LOD for rabbit AFP antigen was estimated to be 0–1 ng mL⁻¹ from the standard deviations above the background.

Our proposed SERS-based immunoassay system, using a gold-embedded gradient microfluidic channel and antibody-conjugated HGNs, has several advantages over other previously reported SERS immunoassays. The tedious manual dilution process that involves repetitive pipetting and can result in inaccurate dilution can be eliminated because various concentrations of biomarker are automatically generated by a gradient channel. The assay time from serial dilution to SERS detection takes less than 60 min because all of the experimental conditions for the formation and detection of immunocomplexes can be automatically controlled inside the exquisitely-designed microfluidic channel.

In summary, the immunoassay using a SERS-based optofluidic sensor has multiple advantages including small sample consumption, an easy washing step, and a quick assay time. By using properly-designed SERS nanoprobes, the Raman signals for specific cancer targets could be quickly analyzed by on-chip detection in the microfluidic channel. The simultaneous detection of multiple cancer markers and their quantitative analysis for clinical sera are under investigation using this technology. We anticipate that this SERS-based optofluidic sensor could open up a new detection method for early cancer diagnosis.

Acknowledgements

This work was supported by the National Research Foundation of Korea (grant numbers R11-2008-0061852 and K20904000004-11A0500-00410). This work was also partially supported by the Agency for Defense Development through the Chemical and Biological Defense Research Center.

References

1. V. Kanda, J. K. Karjuki, D. J. Harrison and M. T. McDermott, Anal. Chem., 2004, 76, 7257–7262.
2. L. A. Terry, S. F. White and L. J. Tigwell, J. Agric. Food Chem., 2005, 53, 1309–1316.
3. C. A. Rowe, S. B. Scruggs, M. J. Feldstein, J. P. Golden and F. S. Ligler, Anal. Chem., 1999, 71, 433–439.
4. A. Y. Rubina, A. Kolchinsky, A. A. Makarov and A. S. Zasedatelev, Proteomics, 2008, 8, 817–831.
5. M. Bally, M. Halter, J. Voros and H. M. Grandin, Surf. Interface Anal., 2006, 38, 1442–1458.
6. N. Rosi and C. A. Mirkin, Chem. Rev., 2005, 105, 1547–1562.
7. A. W. Peterson, R. J. Heaton and R. Georgiadis, J. Am. Chem. Soc., 2000, 122, 7837–7838.
8. S. W. Bishnoi, Y. Lin, M. Tibudan, Y. Huang, M. Nakaema, V. Swarup and T. A. Keiderling, Anal. Chem., 2011, 83, 4053–4060.
9. A. Shen, L. Chen, W. Xie, J. Hu, A. Zeng, R. Richard and J. Hu, Adv. Funct. Mater., 2010, 20, 969–975.
10. Y. C. Cao, R. C. Jin and C. A. Mirkin, Science, 2002, 297, 1536–1538.
11. T. Park, S. Lee, G. H. Seong, J. Choo, E. K. Lee, Y. S. Kim, W. H. Ji, S. Y. Hwang, D. G. Gweon and S. Lee, Lab Chip, 2005, 5, 437–442.
12. L. Chen and J. Choo, Electrophoresis, 2008, 29, 1815–1828.
13. L. X. Quang, C. Lim, G. H. Seong, J. Choo, K. J. Do and S. K. Yoo, Lab Chip, 2008, 8, 2214–2219.
14. H. Chon, C. Lim, S. M. Ha, Y. Ahn, E. K. Lee, S. I. Chang, G. H. Seong and J. Choo, Anal. Chem., 2010, 82, 5290–5295.
15. 17 C. Lim, J. Hong, B. G. Chung, A. J. deMello and J. Choo, Analyst, 2010, 135, 837–844.
16. H. Chon, S. Lee, S. W. Son, C. H. Oh and J. Choo, Anal. Chem., 2009, 81, 3029–3034.
17. X. X. Han, L. J. Cai, J. Guo, C. X. Wang, W. D. Ruan, W. Y. Han, W. Q. Xu, B. Zao and Y. Ozaki, Anal. Chem., 2008, 80, 3020–3024.
18. T. Li, L. Gu and Z. Wang, Biosens. Bioelectron., 2008, 23, 1125–1130.
19. J. Ni, R. J. Lipert, G. B. Dawson and M. D. Porter, Anal. Chem., 1999, 71, 4903–4908.
20. J. D. Driskel, K. M. Kwarta, R. J. Lipert, M. D. Porter, J. D. Neill and J. F. Ridpath, Anal. Chem., 2005, 77, 6147–6154.
21. D. S. Grubusha, R. J. Lipert, H. Y. Park, J. Driskell and M. D. Porter, Anal. Chem., 2003, 77, 5936–5943.
22. C. C. Lin, Y. M. Yang, Y. F. Chen, T. S. Yang and H. C. Chang, Biosens. Bioelectron., 2008, 24, 178–183.
23. S. Xu, X. Ji, W. Xu, X. Li, L. Wang, Y. Bai, B. Zhao and Y. Ozaki, Analyst, 2004, 129, 63–68.
24. G. Wang, R. J. Lipert, M. Jain, S. Kaur, S. Chakraboty, M. P. Torres, S. K. Batra, R. E. Brand and M. D. Porter, Anal. Chem., 2011, 83, 2554–2561.
25. M. Lee, S. Lee, J. H. Lee, H. W. Lim, G. H. Seong, E. K. Lee, S. I. Chang, C. H. Oh and J. Choo, Biosens. Bioelectron., 2011, 26, 2135–2141.
26. A. M. Schwartzberg, T. Y. Oshiro, J. Z. Zhang, T. Huser and C. E. Talley, Anal. Chem., 2006, 78, 4732–4736.
27. H. Wang, G. P. Goodrich, F. Tam, C. Oubre, P. Nordlander and N. J. Hallas, J. Phys. Chem. B, 2005, 109, 11083–11087.
28. C. E. Talley, J. B. Jackson, C. Oubre, N. K. Grady, C. W. Hollars and S. M. Lane, Nano Lett., 2005, 5, 119–124.
29. C. D. Syme, C. Martino, R. Yusvana, N. M. S. Sirimuth and J. M. Cooper, Anal. Chem., 2012, 84, 1491–1495.
30. C. D. Syme, N. M. S. Sirimuth, S. L. Faley and J. M. Cooper, Chem. Commun., 2010, 46, 7921–7923.
31. R. Wilson, S. A. Bowden, J. Parnell and J. M. Cooper, Anal. Chem., 2010, 82, 2119–2123.
32. R. Wilson, S. P. Monaghan, S. A. Bowden, J. Parnell and J. M. Cooper, Anal. Chem., 2007, 79, 7036–7041.
33. A. März, T. Henkel, D. Cialla, M. Schmitt and J. Popp, Lab Chip, 2011, 11, 3584–3592.
34. A. Walter, A. Maerz, W. Schmacher, P. Roesch and J. Popp, Lab Chip, 2011, 11, 1013–1021.
35. K. R. Strehle, D. Cialla, P. Roesch, T. Henkel, M. Koehler and J. Popp, Anal. Chem., 2007, 79, 1542–1547.
36. B. Han, N. Choi, K. H. Kim, D. W. Lim and J. Choo, J. Phys. Chem. C, 2011, 115, 6290–6296.
37. L. C. Taylor, T. B. Kirchner, N. V. Lavrik and M. J. Sepaniak, Analyst, 2012, 137, 1005–1012.
38. R. M. Connaster, M. Cochran, R. Harrison and M. J. Sepaniak, Electrophoresis, 2008, 29, 1441–1450.
39. K. W. Oh, K. Lee, B. Ahn and E. P. Furlani, Lab Chip, 2012, 12, 515–545.
40. K. Lee, C. Kim, B. Ahn, R. Panchapakesan, A. R. Full, L. Nordee, J. Y. Kang and K. W. Oh, Lab Chip, 2009, 9, 709–717.
41. K. Lee, C. Kim, Y. Kim, B. Ahn, R. Panchapakesan, J. Bang, J. Kim, Y. K. Yoon, J. Y. Kang and K. W. Oh, Microfluid. Nanofluid., 2011, 11, 75–86.
42. A. Riaz, R. K. Ryu, L. M. Kulik, M. F. Mulcahy, R. J. Lewandowski, J. Minocha, S. M. Ibrahim, K. T. Sato, T. Baker, F. H. Miller, S. Newman, R. Omary, M. Abecassis, A. B. Benson and R. Salem, J. Clin. Oncol., 2009, 27, 5734–5742.
43. S. K. W. Dertinger, D. T. Chiu, N. L. Jeon and G. M. Whitesides, Anal. Chem., 2001, 73, 1240–1246.

---

Footnotes

† Published as part of a themed issue on optofluidics

‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/c2lc40353f

§ Joint first authors

---