Quantitative lateral flow immunosensor using carbon nanotubes as label

Abstract

A novel immunosensor utilizing multi-walled carbon nanotube as label on a lateral flow system for simple and quantitative electrical detection was investigated. Multi-walled carbon nanotubes (MWCNTs) were first modified with polyvinylpyrrolidone (PVP) for uniform dispersion in aqueous solution. Then, the MWCNTs were conjugated with human immunoglobulin G (IgG) using 1-(3-(dimethylamino)-propyl)-3-ethylcarbodiimide hydrochloride (EDC) coupling chemistry. The lateral flow immunosensor was made of nitrocellulose membrane that transports sample and reagents through a porous structure by capillary action. The performance of the immunosensor was demonstrated using human IgG as a model analyte competing with the conjugate (MWCNT-labeled human IgG) at the capture zone. As a result of binding reaction between the conjugate and the immobilized Protein A at the capture zone, the conjugated MWCNTs formed a conducting network at the capture zone providing conductance measurement corresponding to the amount of captured conjugate. Quantitative immunoassay response for the target human IgG was demonstrated in the range of 25 to 200 µg ml⁻¹ without an additional amplification step. The presented immunosensing technique could be expanded for the detection of various analyte-specific biomolecules with a potential for simple and rapid tests suitable for point-of-care diagnostics.

---

1. Introduction

Lateral flow (LF) immunoassay is one of the most successful point-of-care diagnostic tools commercially available for detecting various health and environmental agents.^1–6 A lateral flow immunosensor is made of a porous membrane for both immobilizing biological receptors and transporting sample reagents. Sample reagents are transported in the lateral direction through the porous membrane by capillary force and captured by the biological receptors immobilized at the capture zone. LF immunosensors are simple to use, low-cost, rapid, and portable with relatively long shelf life. Examples of lateral flow diagnostics include home pregnancy tests, detection of drugs of abuse, diagnosis of infectious diseases, and cancer markers. However, such systems often suffer from low sensitivity, and are mostly restricted to qualitative or semi-quantitative measurements.

Generally, LF immunosensors are based on optical detection where tagged particles or enzymes provide visible signal generation or amplification. Typically, colloidal gold is used as a label for colorimetric detection traced by the visible color. Several efforts have been made towards signal quantification using enhanced colorimetric labels or conducting materials such as phosphorescent nanoparticles,⁷ polyaniline,⁸ and polyaniline with colloidal gold.⁹ Colorimetric techniques are often limited by low sensitivity while fluorescent techniques suffer from instability, photobleaching, activation step, and requirement for a complicated reader device.

This work addresses such limitations by utilizing carbon nanotubes (CNTs) as labeling material for direct electrical signal measurement. CNTs have remarkable one-dimensional electronic conduction properties and large surface area, and allow the adsorption of biomolecules for sensing.^10–12 In order to utilize these properties of carbon nanotubes for biosensing applications, surface functionalization is necessary to uniformly disperse CNTs in aqueous solution and immobilize biorecognition molecules on the surface of the nanotubes. Adsorption of surfactants and polymers on the surface of CNTs is considered as an effective method to enhance the dispersion of CNTs without affecting their properties.^13–15 Compared to CNTs, other conducting materials such as colloidal gold and polymer nanowires do not possess the unique material properties suitable for surface functionalization or maintain high electrical conduction after binding reaction.^9,16

In this work, a new immunoassay technique using carbon nanotube labels on lateral flow system is introduced. The illustration of this technique is shown in Fig. 1. MWCNTs are dispersed in water with PVP using sonication. Adsorption of PVP provides a thin and rough layer of organic coating by wrapping around the MWCNTs. Human IgG is immobilized on the bare (not covered by PVP) surface of the MWCNTs. The MWCNT–human IgG conjugate is then applied on a lateral flow immunosensor where the MWCNT–human IgG conjugate is to be captured by immobilized Protein A at the capture zone. The conjugated MWCNTs form a conducting network at the capture zone corresponding to the amount of binding. Consequently, quantitative signal can be obtained by simple resistance measurement of the capture zone. Blocking schemes were also implemented on the MWCNTs and the lateral flow strips to address non-specific interaction.

Fig. 1 Illustration of the detection principle using carbon nanotubes as label for quantitative lateral flow immunosensing in competitive format. Carbon nanotubes form a conducting network at the capture zone corresponding to the amount of binding reaction. Consequently, quantitative signal can be obtained by simple resistance measurement of the capture zone.

2. Methods and experiments

2.1 Materials and reagents

Carboxylated multi-walled carbon nanotubes (purity > 95 wt%) were obtained from Cheaptubes, Inc. (Brattleboro, VT). Hi-flow Plus nitrocellulose membranes (HF13502), glass fiber pads, cellulose membranes, and a supporting card for lateral flow assay were obtained from Millipore (Bedford, MA). Protein A suspended in solution was obtained from BioVision, Inc. (Mountain View, CA). Purified human immunoglobulin G (IgG, 6.2 mg ml⁻¹) suspended in 0.01 M sodium phosphate (0.15 M NaCl, pH 7.4) was obtained from Sigma-Aldrich (St Louis, MO). Polyvinylpyrrolidone (PVP, M_w 10 000), bovine serum albumin (BSA), Tween-20, glutaraldehyde, and 2-(N-morpholino)ethanesulfonic acid (MES) were from Sigma-Aldrich. 1-(3-(Dimethylamino)-propyl)-3-ethylcarbodiimide hydrochloride (EDC) and N-hydroxysulfosuccinimide (sulfo-NHS) were obtained from Thermo Fisher Scientific, Inc. (Rockford, IL). All immunoreagents were diluted in pH 7.4 phosphate saline buffer (PBS) unless noted otherwise.

2.2 MWCNT–antibody conjugation

In order to start the conjugation process, 1.5 mg of carboxylated MWCNTs were dispersed in 2 ml of deionized (DI) water with 0.6 mg of PVP. The solution was sonicated for 4 hours in an ultrasonic bath at 27 °C to exfoliate the MWCNTs and accelerate the dispersion. The dispersed solution was mixed with 1 ml of EDC (0.4 M) and sulfo-NHS (0.1 M) in pH 6.1 MES buffer. The solution was left to react for 30 minutes with gentle agitation at room temperature. The mixture was microcentrifuged at 13 000 rpm for 3 minutes with PBS buffer and the supernatant was discarded. The washing and re-suspending process was repeated three times to remove excess reagents. Human IgG solution was prepared at 200 µg ml⁻¹ by diluting with PBS buffer for conjugation. The solution containing functionalized MWCNTs was then mixed with 100 µl of the human IgG solution and left to react overnight at 4 °C. This mixture was centrifuged at 13 000 rpm for 3 minutes and the supernatant was discarded. The washing and re-suspending process was repeated three times to remove unbound human IgG, and the resulting MWCNT–antibody conjugate solution was kept at 4 °C until use.

2.3 Preparation of lateral flow immunosensors

A nitrocellulose membrane was used as a detection pad with designated capture zone for immobilizing Protein A. Fiber glass and cellulose membranes were used for sample application and for absorption of excess reagents by placing them at the two ends of the nitrocellulose membrane, respectively. The LF immunostrips were prepared following general procedure reported elsewhere.^7,9 First, a nitrocellulose membrane was saturated with 10% (v/v) methanol in deionized (DI) water for 30 minutes to remove any residue and dried in air at room temperature. The membrane was then treated with 0.5% (v/v) glutaraldehyde solution to strengthen the immobilization of Protein A on the nitrocellulose membrane and rinsed with DI water. After drying, Protein A diluted in PBS buffer to 1 mg ml⁻¹ was immobilized at a designated capture zone by dragging a dispenser with manual control. The membrane was then incubated with 2% BSA and 0.05% (v/v) Tween-20 in PBS for 1 hour in order to block the membrane from making unwanted binding. The membrane was rinsed with PBS containing 0.05% (v/v) Tween-20 to remove the excess blocking reagent and left to dry at room temperature.

3. Results and discussion

A lateral flow immunosensor was developed by assembling a lateral flow strip with sample application, detection, and absorbent pads. The specific binding reaction was based on immobilized Protein A and human IgG that was detected by the conjugated MWCNT labels. Since CNTs have inherent affinity to biological elements causing non-specific binding, control tests were conducted at different functionalization steps. First, the PVP treated MWCNTs were applied to the LF immunosensors to bind with immobilized Protein A. As a result, unwanted binding of the MWCNTs with Protein A at the capture zone was observed as shown in Fig. 2(a). This non-specific binding could have been caused by the affinity of the exposed MWCNTs surface to the Protein A. Next, the crosslinking agents, EDC and sulfo-NHS, were added to the PVP treated carbon nanotubes and the same test was repeated. In this case, no binding reaction was observed between the MWCNTs and Protein A at the capture zone as shown in Fig. 2(b). This result showed that the surfaces of the MWCNTs were completely covered by the PVP and the amide coupling agents limiting unoccupied sites from non-specific binding. Finally, MWCNT–IgG conjugate solution was applied, where binding between human IgG and Protein A was evident at the capture zone as shown in Fig. 2(c). Scanning electron microscopy (SEM) was used to examine the capture zone, where the MWCNTs formed a conducting network as a result of binding. Fig. 3 shows SEM image of the nitrocellulose membrane without MWCNT labels and with varying concentration of MWCNT labels corresponding to the amount of binding.

Fig. 2 Non-specific adsorption on the nitrocellulose membrane: (a) MWCNTs partially coated with PVP were captured by Protein A due to non-specific adsorption which must be addressed to avoid false signal; (b) almost no detection signal was observed at the capture zone indicating that the surface of the MWCNTs were completely covered eliminating non-specific adsorption, and (c) selective binding between the MWCNT–human IgG conjugate and immobilized Protein A.

Fig. 3 SEM images of the MWCNT network at the capture zone: (a) before sample application; (b) lower MWCNT coverage due to a smaller number of binding; and (c) higher MWCNT coverage due to a larger number of binding. The scale bar is 1 µm.

Prior to the immunoassay test, varying concentration of the conjugate solutions ranging from 0 to 200 µg ml⁻¹ was prepared by dilution in PBS buffer in order to determine the detection limit of the conjugate solution. 50 µl of each solution were applied to the LF immunosensors where the conjugated human IgG was captured by the Protein A at the capture zone resulting in colorimetric signal as shown in Fig. 4. As depicted in Fig. 4, the color intensity of the signal was stronger for higher concentration of the conjugated human IgG. For quantitative analysis of the binding reaction, the resistance across the capture line was measured after the complete development of the visible signal (∼15 minutes). The measured electrical resistance decreased as the concentration of the MWCNT–human IgG conjugate increased due to the increased amount of MWCNTs forming a conducting network. Three replications of the immunosensors with same conjugate concentration were tested for consistency of the measurements. The electrical resistance was measured using electrodes with 3 mm spacing and while maintaining same contact resistance.

Fig. 4 Calibration results using conjugate solutions in different concentrations. MWCNTs as label provided a quantitative detection capability.

The developed immunosensor was used with increasing concentration of human IgG as target analyte in solution. Concentration of the human IgG target analyte ranged from 25 to 200 µg ml⁻¹ and the concentration of the conjugate solution was fixed at 200 µg ml⁻¹. Each target analyte solution was briefly mixed with the conjugate and then the mixture was applied to the LF immunosensor as a one-step immunoassay process. As the sample solution moved across the capture zone, a competitive immunoassay reaction between the conjugated human IgG and those in the solution occurred. This produced a detection signal by the MWCNTs across the capture line varying in color intensity that was inversely proportional to the amount of human IgG target analyte captured as shown in Fig. 5(a). The detection signals were visible with the application of only 20 µl, and saturation occurred with the application of 50 µl of total sample solutions within 15 minutes. The visibility of the capture zone decreased with increasing target human IgG concentration. As the concentration of the target analyte increased, it occupied more binding sites limiting the binding reaction between the conjugated human IgG and the Protein A. The corresponding measured results in Fig. 5(b) clearly show that the electrical resistance increased with increasing concentration of the target human IgG. Although the visibility of capture lines decreased with increasing target IgG concentration (100, 150, and 200 µg ml⁻¹), the electrical measurements clearly showed differences corresponding to the various concentrations at all levels.

Fig. 5 Competitive immunoassay results: (a) lateral flow immunosensors with varying target concentrations and (b) the measured electrical resistance at the capture zones. The width of the lateral flow immunostrips was 7 mm.

4. Conclusions

A lateral flow immunosensing technique using carbon nanotubes as a label was effectively demonstrated for quantitative antibody detection based on simple electrical measurements. The surface of the MWCNTs was partially modified with PVP for uniform dispersion in aqueous solutions. Further functionalization steps were conducted to attach antibodies on the exposed surface of MWCNTs utilizing EDC/sulfo-NHS chemistry. The modified MWCNTs successfully provided both direct colorimetric and conductimetric measurements for antibody binding without addition amplification step. SEM images showed the MWCNT network formed as a result of antibody binding at the capture zone. The lateral flow system facilitated the binding reaction and reagent handling in a simple and convenient manner. This immunosensing technique is easy for integration and requires ordinary tools for signal measurement. Therefore, it holds the potential for quantitative determination of various antibodies and antigen in a point-of-care setup. The detection limit and time could be improved further with proper optimization of the conjugating process and the lateral flow system.

Acknowledgements

This work was supported in part by Fund for Innovation in Engineering Research from the College of Engineering of Louisiana State University, National Science Foundation EPSCoR through Louisiana Board of Regents under contract NSF(2008)-PFUND-90, and National Science Foundation Integrative Graduate Education and Research Traineeship (IGERT) Program on Multi-Scale Computations of Fluid Dynamics.

References

1. R. E. Biagini, D. L. Sammons, J. P. Smith, B. A. MacKenzie, C. A. F. Striley, J. E. Snawder, S. A. Robertson and C. P. Quinn, Clin. Vaccine Immunol., 2006, 13, 541–546.
2. S. Buhrer-Sekula, H. L. Smits, G. C. Gussenhoven, J. van Leeuwen, S. Amador, T. Fujiwara, P. R. Klatser and L. Oskam, J. Clin. Microbiol., 2003, 41, 1991–1995.
3. T. J. Clark, P. H. McPherson and K. F. Buechler, Point Care, 2002, 1, 42–46.
4. P. Corstjens, M. Zuiderwijk, A. Brink, S. Li, H. Feindt, R. S. Niedbala and H. Tanke, Clin. Chem. (Washington, DC, U. S.), 2001, 47, 1885–1893.
5. C. Fernandez-Sanchez, A. M. Gallardo-Soto, K. Rawson, O. Nilsson and C. J. McNeil, Electrochem. Commun., 2004, 6, 138–143.
6. J. Ryan, K. Dave, E. Emmerich, B. Fernandez, M. Turell, J. Johnson, K. Gottfried, K. Burkhalter, A. Kerst, A. Hunt, R. Wirtz and R. Nasci, J. Med. Entomol., 2003, 40, 95–99.
7. X. Song and M. Knotts, Anal. Chim. Acta, 2008, 626, 186–192.
8. S. Pal, E. C. Alocilja and F. P. Downes, Biosens. Bioelectron., 2007, 22, 2329–2336.
9. J.-H. Kim, J.-H. Cho, G. S. Cha, C.-W. Lee, H.-B. Kim and S.-H. Paek, Biosens. Bioelectron., 2000, 14, 907–915.
10. P. M. Ajayan, Chem. Rev., 1999, 99, 1787–1799.
11. P. Avouris, Acc. Chem. Res., 2002, 35, 1026–1034.
12. M. Bockrath, D. H. Cobden, P. L. McEuen, N. G. Chopra, A. Zettl, A. Thess and R. E. Smalley, Science, 1997, 275, 1922–1925.
13. R. J. Chen, Y. Zhang, D. Wang and H. Dai, J. Am. Chem. Soc., 2001, 123, 3838–3839.
14. A. Hirsch, Angew. Chem., Int. Ed., 2002, 41, 1853–1859.
15. Y.-L. Zhao and J. F. Stoddart, Acc. Chem. Res., 2009, 42, 1161–1171.
16. J. Wang, Electroanalysis, 2005, 17, 7–14.

---