Reusable, polyethylene glycol-structured microfluidic channel for particle immunoassays
- First Online:
A microfluidic channel made entirely out of polyethylene glycol (PEG), not PEG coating to silicon or polydimethylsiloxane (PDMS) surface, was fabricated and tested for its reusability in particle immunoassays and passive protein fouling, at relatively high target concentrations (1 mg ml-1). The PEG devices were reusable up to ten times while the oxygen-plasma-treated polydimethyl siloxane (PDMS) device could be reused up to four times and plain PDMS were not reusable. Liquid was delivered spontaneously via capillary action and complicated bonding procedure was not necessary. The contact angle analysis revealed that the water contact angle on microchannel surface should be lower than ~60°, which are comparable to those on dried protein films, to be reusable for particle immunoassays and passive protein fouling.
Polyethylene glycol (PEG) surfaces have been recognized to resist protein fouling due to their hydrophilic nature (water contact angle = ~20°). The existence of oxygen in their backbone -(CH2CH2O-)n and a high degree of H2O/PEG structural organization  enable the reversal of binding before the adsorbed protein "flattens out" and denatures through forming multiple attachments to a surface . This protein fouling is a key problem in performing biological assays in a microfluidic device [3, 4]. Therefore, there have been several attempts to modify their surfaces, including silicon and polydimethylsiloxane (PDMS), with PEG. These modifications include passive adsorption , chemical vapour deposition (CVD) [6, 7]. These efforts have proved unsuccessful due to their fabrication complexity  or poor long-term stability (some PEG coatings may eventually come off from microchannel surface upon rinsing ). Alternatively, PEG has been covalently added to microchannel surface through self-assembling PEG-terminated alkyl silane on silicon-based surfaces. This coating is also known as PEG-SAM (the latter represents self-assembled monolayer) [9, 10]. The potential problems of PEG-SAM include: (1) difficulty of immobilizing certain bioreceptors (e.g. antibodies) within a microchannel (PEG repels those bioreceptors), (2) uneven coating to complicated structures such as cross junctions, view cells, and microvalves, and (3) time-consuming, complex process of making PEG-SAM on the covered microchannel [8, 11]. Chemical grafting of PEG onto silicon or PDMS has been attempted [12, 13], which provided better quality final film. However, this method require multiple, difficult-to-control processing steps to achieve a high quality [8, 14]. In addition, the organic solvents required for PEG coating would swell the PDMS network . All these complications originate from the fact that the PEG layer is added to the existing silicon or PDMS surface.
A better alternative is to fabricate a microfluidic channel made solely out of PEG. Kim et al.  has recently fabricated a microchannel comprised entirely of PEG by cross-linking it through exposure to ultraviolet (UV) radiation. Indeed, their device was successful in resisting protein fouling, but a few complications could be found in their work. They discussed how to prevent PEG swelling and how to make better bonding between mold-replica or microchannel-cover slip, indicating potential fabrication complications. We have actually duplicated their technology and found that PEG device was not suitable for repeated uses. Bonding of a PEG microchannel to either a glass cover slide or another PEG substrate was found to be difficult due to their surface roughness. Leaking was observed from the very first use and became worse upon repeated use. In addition, they did not expose their device to repeated washing conditions that is common in practical biological assays. As expected, they did not performed actual biological assays with their device.
In this work, we expanded the work of Kim et al.  by (1) eliminating the bonding procedure between the PEG substrate and a glass cover slide, (2) demonstrating simpler liquid delivery via capillary action (Kim et al. used a micropump), and (3) performing actual biological assays in a repeatable manner at relatively high protein concentrations (Kim et al. tested 20–50 μg ml-1; we tested ~1 mg ml-1). Particle immunoassays for mouse immunoglobulin G (mIgG) were repeated at very high concentrations such that the reusability times could be estimated for the PEG microfluidic channel. Fluorescein-labelled bovine serum albumin (BSA) was also tested to further evaluate passive protein fouling.
Results and discussion
Simplified assembly/use of a PEG microchannel
Reusability as observed with microscopic images
This threshold angle of 60° can be correlated to those on protein films. Water contact angles on salt-free, dried protein films were measured as: 75 ± 1° for bovine serum albumin, 67 ± 1° for bovine hemoglobin, and 47 ± 1° for hen egg white lysozyme. These contact angles are comparable to our threshold angle of 60°. Once the water contact angle of microchannel surface exceeds those of proteins, hydrophobic-interactions-induced protein adsorption will be preferred on microchannel surface that will permanently foul the surface [19, 20].
Through this work, we reported that PEG-structured microchannel was not only protein fouling-resistant but also reused repeatedly. We hope that this device can be installed at a permanent location to perform unmanned bio-assays, eliminating the need for device replacement. More detailed biocompatibility studies should be followed for the other types of bio-assays and with various target biomolecules.
Particles and target proteins
In order to perform particle immunoassays, antibodies were conjugated to microparticles by physical adsorption as described previously . Briefly, 1 ml of 0.02% w/v, 0.92 μm highly carboxylated polystyrene particles (parking area = 10.3 Å2 per carboxyl surface group; Bangs Laboratories, Fishers, Indiana, USA) were mixed with 1 ml of 1.023 μg ml-1, anti-mouse immunoglobulin G (anti-mIgG; catalog number M8642, Sigma-Aldrich Co, St. Louis, Missouri, USA) solution, followed by centrifuging and resuspension (the whole cycle was repeated twice) to eliminate the free antibodies. The surface coverage of the antibodies on particle surface is approximately 33%, which is appropriate in maximizing particle immunoagglutination . Target protein was mouse immunoglobulin G (mIgG; catalog number I5381, Sigma-Aldrich). For a comparison purpose, 1 mg ml-1 fluorescein-labelled bovine serum albumin solution (with fluorescein isothiocyanate; FITC-BSA; catalog number A9771, Sigma-Aldrich) was used to monitor the protein-fouling behaviour within the PEG microfluidic channel. All dilutions were made with 10 mM phosphate buffered saline (PBS; pH 7.4; Sigma-Aldrich).
Fabrication of PEG microfluidic channels
Image analysis of microfluidic channels
Particle suspensions and/or protein solutions were introduced to a PEG microfluidic device via capillary action from the liquid droplets (3 μl each) sitting on the inlets, as shown in Fig. 1. Five minutes after introducing the solutions, an adhesive tape was removed and the microfluidic channel was rinsed with deionized water, followed by observation of the inner surfaces of a microfluidic channel with an inverted, light or fluorescent microscope (Nikon Instruments, Tokyo, Japan) (Fig. 1). This procedure was repeated until visually identifiable particle agglutinates (i.e. triplets or larger clumps) or BSA (i.e. three or more bright fluorescent spots) could not be removed from the microfluidic channel by rinsing. Additionally, we used PDMS microfluidic channels (i.e. the most popular device) as negative controls, with the same layout and dimensions of microchannels. Both "fresh" oxygen-plasma-treated PDMS (water contact angle < 10°) and "aged" PDMS (two days of incubation at room temperature after oxygen plasma treatment; water contact angle ~90°) were tested. The fabrication procedure of PDMS microfluidic device can be found in previous publications [17, 23].
A contact angle/surface tension analyzer (FTÅ200, First Ten Ångstroms, Portmouth, Virginia, USA) was used to measure the contact angles on the PEG surfaces with 5 μl sessile drops of deionized water. To simulate water rinsing of a microfluidic channel, the PEG surfaces were rinsed with deionized water after each contact angle measurement. Sessile drops were placed on the surfaces for 2 min, the same as the liquid exposure time of PEG (and PDMS) microfluidic channels. A single data point was averaged from three different measurements of contact angle on PEG (and PDMS) surfaces. Substrates were thoroughly dried with nitrogen gas prior to contact angle measurement.
Water contact angles on salt-free, dried protein films were also measured using the same instrument. Bovine serum albumin (catalog number P-7656, essentially salt-free, Sigma, St. Louis, Missouri, USA), bovine hemoglobin (catalog number H-9891, essentially salt-free, dimethylated and primarily methemoglobin, Sigma), and hen egg white lysozyme (salt-free, catalog number 10 837 059 001, Roche, Germany) were used as model proteins. 15 μl of 1 mg ml-1 albumin, haemoglobin or lysozyme solution was deposited on a precleaned glass microscope slide (catalog number 12–552, Fisher Scientific, Pittsburgh, Philadelphia, USA), and stored in a nitrogen-purged desiccator for more than a week. 5 μl droplets of deionized water (from Millipore's Simplicity, resistivity > 18 MΩ cm) were automatically dispensed by the same contact angle analyzer, and deposited on the protein films. Contact angles were measured right after a droplet stops vibrating and forms a perfect spherical shape. This was normally achieved within 33 ms (images were captured every 33 ms). Sometimes the contact angle kept decreasing significantly over time, probably due to the absorption of solvents into the films. We simply eliminated such data from our experimental set.
The authors are grateful to Micro/Nanofabrication Center at the University of Arizona for the cleanroom facility and equipment assistance. This work was funded by National Veterinary Research and Quarantine Service (NVRQS), Republic of Korea, award no. C-AD14-2006-11-0.
- 3.Mukhopadhyay R: When microfluidic devices go bad. Anal Chem 2005, 77: 429A-432A.Google Scholar
- 5.Gingell D, Owens N, Hodge P, Nicholas CV, Odell R: Adsorption of a novel fluorescent derivative of a poly(ethylene oxide)/poly(butylene oxide) block-copolymer on octadecyl glass studied by total internal-reflection fluorescence and interferometry. J Biomed Mater Res 1994, 28: 505-513. 10.1002/jbm.820280413CrossRefGoogle Scholar
- 8.Harbers GM, Emoto K, Greef C, Metzger SW, Woodward HN, Mascali JJ, Grainger DW, Lochhead MJ: Functionalized poly(ethylene glycol)-based bioassay surface chemistry that facilitates bio-immobilization and inhibits nonspecific protein, bacterial, and mammalian cell adhesion. Chem Mater 2007, 19: 4405-4414. 10.1021/cm070509uCrossRefGoogle Scholar
- 11.Choi HG, Zhang Z, Boccazzi P, Laibinis PE, Sinskey AJ, Jensen KF: Poly(ethylene glycol) (PEG)-modified poly(dimethylsiloxane) for protein- and cell-resistant surfaces in microbioreactor. In Proceedings of μTAS 2003: 5–9 October 2003; Squaw Valley, California. Edited by: Northrup MA, Jensen KF, Harrison DJ. The Transducers Research Foundation; 2003:1105-1108.Google Scholar
- 21.Bangs Laboratories Inc: Handling-Specific Technotes No. 205: Adsorption to Microspheres. Fishers: Bangs Laboratories Inc; 2002.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.