Research highlights

Fabrication across length scales

An advantage to using lab-on-a-chip devices in the study of biological systems is the access they provide to a wider set of experimental parameters than would be possible at larger length scales. This is particularly important in studying human cells and tissue, where complex micro- and nanoscale spatial cues are known to dictate cell fate.¹

To overcome the spatial limitations of a single fabrication scheme, Doh and colleagues utilized a dual fabrication scheme to design surfaces with micro- and nanopatterned protein arrangements. Kwon et al.² first coated a substrate with biotin and then with a layer of photosensitive poly(2,2-dimethoxy nitrobenzyl methacrylate-r-methyl methacrylate-r-poly(ethylene glycol) methacrylate), or PDMP. They then molded PDMP from a polyurethane acrylate master by applying pressure (Fig. 1, top), a process referred to as capillary force lithography. The resulting nanoscale PDMP features were 700 nm wide ridges and 350 nm wide grooves. The biotin surface underneath the grooves was exposed in this process and subsequently coated with isothiocyanate-conjugated streptavidin.

Fig. 1 Two step fabrication scheme for multiscale patterning via capillary force lithography for nanoscale patterns and microscope projection photolithography for microscale features (top), and the final structures patterned with different proteins (bottom). Reprinted and adapted with permission from Kwon et al.² Copyright 2011 American Chemical Society.

To generate microscale features, the remaining PDMP ridges were immersed in PBS and exposed to UV light (365 nm) through a mask, via microscope projection photo-lithography. The exposed PDMP ridges dissolved, thereby exposing more areas coated with biotin. Now these newly available regions could be coated with different molecules, e.g. FITC-conjugated streptavidin. By applying this two-step technique in which different biotin-coated regions were sequentially made available for functionalizing, the authors were able to create both multiscale topographical features and multiple regions coated with different proteins (Fig. 1, bottom).

The ability of such multi-component patterns to regulate cell behavior was demonstrated on human colon cancer cells. When seeded onto such complex topographical surfaces, the cells were observed to preferentially align along the features, rather than across them, and more importantly, to move towards nanoscale structures. Namely, when cells were seeded onto microscale features, they were evenly distributed across the surface. However, when nanoscale patterns were incorporated into microscale features, the cells preferred the nanogrooves and nanoridges to the larger features.

Spatial cues such as extracellular matrix arrangement and tissue topography exist across multiple length scales. The ability to understand and model these multiscale interactions, offered by this two-step fabrication process, may result in engineered devices that better mimic and interface with biological systems.

Microfiber coding

Nature has been able to generate complex fibers with enhanced functions. These fibers range from the structure of hair strands down to smaller networks such as collagen nanofibrils. Our ability to replicate these processes at small length scales may potentially lead to a range of applications from the textile industry to biomedicine. For example, the fibrous extracellular matrix of connective tissue is structurally encoded with precise mechanical and biochemical cues. A means to encode similar cues within a synthetic fibrous scaffold may lead to better methods to regenerate lost tissue.³

Lee and co-workers have recently developed a microfluidic platform to fabricate spatially and biochemically coded hydrogel microfibers by mimicking the spider silk production. Kang et al.⁴ describe a simple PDMS device which utilizes flow focusing to generate fibers with controlled spatial features. In this process, several coaxial sample channels were connected in a flow focusing arrangement. The sample channels were controlled by on-chip pneumatic valves, activated via a digital interface. This allowed the authors to regulate the flow of each sample towards the flow focusing region at the device outlet. To generate alginate fibers, CaCl₂ was used to focus the flow and prevent the wetting of the sample on the channel walls. It also initialized sample solidification by chelation of the calcium ions. The generated fibers were wound and twisted via a motorized spool.

Controlled activation of the pneumatic valves could be used to regulate serial or sequential coding of the resulting fibers with different channel contents. By controlling the valve operation, coding regions along the fiber that were as small as 800 μm could be generated. Furthermore, the fiber thickness could be adjusted by controlling the number of streams.

These co-flow and alternating flow mechanisms were used to produce topographically and chemically diverse fiber regions. For example, alginic acid mixtures with and without salt were employed to generate porous structures or knots connected via thin non-porous links. The high-salt and porous knots had higher surface energy than the links, giving rise to Laplace pressure, which in turn enabled water collection at the knots. The authors were also able to introduce gas into the fibers. In this case, gas pressure was controlled to determine the bubble size. In another experiment, microfluidic channels with grooved walls were used to produce grooved fibers for studies of neuron adhesion and elongation. It was observed that the neurons were collectively more aligned on grooved fibers than on smooth ones. Furthermore, cell experiments showed that encapsulated fibroblasts and hepatocytes remained viable for several days and that controlled co-cultures of these cells survived longer than the individually cultured regions.

Given the capability of the platform to fabricate topographically and biochemically diverse alginate fibers with great levels of accuracy and precision, this device may be useful for a range of applications. The current system has so far only been tested for ionically gelling materials such as alginate, thus the extension of this process for other types of materials may be an important step in enabling other applications.

Polymers with face memory

Flexible materials that can remember and return to their designated shape after deformation are valuable tools in engineering. Such materials may be useful in applications that range from delivery of cardiac stents through narrow blood vessels to deployment of origami based space antennas.⁶ Much of the work in shape memory polymers (SMP) has focused on programming bulk geometry into the materials. Recently, Ding and colleagues fabricated polymeric surfaces with memorized nanoscale topography that could be recovered after thermal processing. These types of pattern-programmed devices may be useful in applications that require switchable surface features such as optical sensors or controllable interfaces with biological tissue.

To achieve dynamic shape memory nanopatterns, Wang et al.⁵ programmed a permanent structure into the equilibrium shape of the polymer film via room-temperature step-and-flash nanoimprint lithography. In this method, the polymer was crosslinked via UV light onto a mold with the desired shape. The programmed structure contained a dense grating with equally wide ridges and grooves (∼800 nm) and a height of ∼180 nm. The surface generated colorful diffraction patterns under white light. Then, the polymer was molded into a flat surface at 180 °C, a temperature well above its glass transition temperature of 95 °C, by applying pressure (4 MPa) for 30 min with a flat Si-wafer. Cooling the polymer down to 40 °C before removing the pressure source ensured that this new shape was temporarily programmed into the material. AFM measurements showed that the surface height varied by a maximum of 12 nm after this molding step. The newly transparent surface was also indicative of the change in the topography of the polymer. To regain the original equilibrium shape of the surface, the device was briefly heated to 120 °C, when it again acquired its colorful appearance (Fig. 2). AFM measurements of recovered feature height and the cross-sectional profile indicated that the original pattern was completely recovered. This finding was independent of feature width and height, as concluded from further experiments.

Fig. 2 Shape-memory polymer: changes in the shape in response to temperature—schematic, photographs and structural profile (a–i). Structural height as a function of permanent feature height in programmed (j) and recovered original (k) surfaces. Figure reprinted with permission from Wang et al.⁵

The authors showed that they could not only temporarily flatten the permanent pattern, but modify it in arbitrary ways. Instead of using a smooth Si-wafer to flatten the permanent features on the polymer surface, they used a patterned wafer, whose linear ridges were oriented normally to the polymer grating. Feature fidelity in the temporary and permanent surface shapes was comparable to the simpler case described above.

It was also shown that multiple patterns could be selectively programmed into and deleted from a polymer by taking advantage of the material's wide glass transition range. Here, Nafion, a sulfonated fluoropolymer, was used. Its glass transition temperature ranges from roughly 55 °C to 130 °C. Three patterns, oriented at different angles, were successively programmed into a smooth Nafion film, the first pattern at 180 °C, the second at 90 °C, and the third at 60 °C. Lowering the temperature for programming each additional pattern ensured that the previous patterns were not modified. To sequentially delete the programmed patterns and regain the flat surface, the material was first heated to 90 °C and then to 190 °C.

There are multiple potential applications for these SMP. For example, the controlled programming of permanent and temporary shapes allows the polymers to be used in optical instruments, privacy coatings, and for geometric regulation of liquid crystal phases. Nanoimprint lithography is capable of producing features below 1 μm in a cost-effective and simple manner, and in combination with programmable materials could enable bioengineering applications, such as studies of cellular response to novel topographies.

Dry blood microdiagnostics

Blood is an important fluid in clinical diagnostics. But, its perishability and potentially hazardous nature require controlled storage and transport conditions. Dried blood spots (DBS) are emerging as a convenient alternative due to easier storage, transport, analysis and disposal. Yet challenges remain in simplifying and streamlining the analysis of DBS.⁷

Wheeler and colleagues have recently developed a microfluidic device that directly accepts dry samples and automates their analysis. This powerful platform was capable of resuspending blood spot constituents and analyzing their amino acid composition with tandem mass spectrometry. This approach has the potential to broaden the range of analytical techniques available for DBS.

The device by Jebrail et al.⁸ utilizes droplet actuation on a surface open to the atmosphere to move the droplets on the surface (Fig. 3). In this process, sequential actuation of several electrodes enables the user to move the droplet on the surface. In the case of DBS, the dry sample—either stemming from a directly deposited drop of blood or from filter paper containing the dried sample—was first mixed on-chip with an extractant solution and incubated. Then the extract was allowed to dry and subsequently dissolved in a derivatization solvent. In this latter step the amino acids in the sample were modified into their respective butyl esters, which could then be detected by tandem mass spectroscopy. The derivatized amino acids were evaporated on chip and stored in a centrifuge tube until off-chip spectroscopic analysis. Alternatively, they were resuspended on-chip with a mass spectroscopy solvent, directed into a microchannel for in situ analysis by nanoelectrospray ionization mass spectrometry.

Fig. 3 Digital microfluidics for analysis of fresh and dried blood (top). Extraction of analytes from dried blood samples (bottom). Figure reprinted with permission from the Royal Society of Chemistry from Jebrail et al.⁸

The efficiency and accuracy of the device was determined through the DBS analysis of three amino acids—methionine, phenylalanine and tyrosine—serving as markers for certain diseases in newborns. Amino acid samples prepared on-chip were analyzed both off- and on-chip. Several samples were artificially spiked with high concentrations of the amino acids, while other samples came from healthy as well as sick patients. In all cases the accuracy and precision of the results and hence the accuracy and precision of the microfluidic sample preparation and analysis proved highly competitive with standard tests conducted at a dedicated laboratory in Ontario, Canada.

Digital microfluidics in general are simple to automate and the quantification of disease markers described here can be completed by a single technician rather than by several laboratory staff using conventional DBS analysis methods. The processing time is also much shorter (∼1 h) compared to other techniques (>3 h). Also, the authors showed that a sample could be analyzed on-chip via tandem mass spectroscopy, which further simplifies and shortens the process and hence the cost of testing. Given its numerous merits, this device has great potential to revolutionize the point-of-care and diagnostic applications of DBS and to alter the screening paradigm in public health.

References

1. A. Khademhosseini, et al., Microscale technologies for tissue engineering and biology, Proc. Natl. Acad. Sci. U. S. A., 2006, 103(8), 2480–2487.
2. K. W. Kwon, et al., Multiscale Fabrication of Multiple Proteins and Topographical Structures by Combining Capillary Force Lithography and Microscope Projection Photolithography, Langmuir, 2011, 27(7), 3238–3243.
3. F. T. Moutos, L. Freed and F. Guilak, A biomimetic three-dimensional woven composite scaffold for functional tissue engineering of cartilage, Nat. Mater., 2007, 6, 162–167.
4. E. Kang, et al., Digitally tunable physicochemical coding of material composition and topography in continuous microfibres, Nat. Mater., 2011 DOI:10.1038/nmat3108.
5. Z. Wang, et al., Programmable, pattern-memorizing polymer surface, Adv. Mater., 2011, 23(32), 3669–3673.
6. Ed. K. Otsuka and C. M. Wayman, Shape Memory Materials, Cambridge University Press, Cambridge, 1998.
7. P. M. Edelbroek, J.v.d. Heijden and L. M. Stolk, Dried blood spot methods in therapeutic drug monitoring: methods, assays, and pitfalls, Ther. Drug Monit., 2009, 31(3), 327–336.
8. M. J. Jebrail, et al., A digital microfluidic method for dried blood spot analysis, Lab Chip, 2011, 11, 3218–3224.

---