The selective protection, or "masking", of substrates prior to a chemical etch or chemical/biological deposition is important for micropatterning substrates in a variety of applications. Traditionally, the first step in creating a mask is based on uniformly coating and selectively exposing a photoresist layer. However, standard photoresist technology is based on organic solvents (thus it is incompatible with most plastics and biomolecule-coated substrates), requires a dry substrate (making it incompatible with the presence live cells), often results in toxic remains (thus requiring extra cleaning steps), and is not easilty implemented with non-flat surfaces (the photoresist is spin-coated).

To overcome some of these limitations, we have developed elastomeric etch masks that (just like photoresist) protect the substrate during the etch step, but are applied onto the substrate (rather than patterned on it). In one form, the etch masks are molded as thin PDMS membranes containing holes, or "stencils", from a master containing posts.

Please click on "Long-term Cellular Micropatterns" above to learn how we use the elastomeric stencils to pattern non-adhesive and cell-adhesive surfaces. Briefly, the stencil is applied onto a chemically grafted glass surface (the graft is a thin film containing the non-adhesive polymer poly-ethylene glycol (PEG), and the stencil/substrate is exposed to an oxygen plasma, which etches PEG on the hole ares to expose glass. Since glass is adhesive to cells, the stencils allow for creating cellular "islands" surrounded by non-adhesive PEG.

Conveniently, oxygen plasma also penetrates into open-ended microchannels, so it is not strictly necessary for many applications to have top access to the substrate. We use this second method to create lines of cells.

This method was mainly developed by Anna Tourovskaia, a Ph.D. student in the lab, in a collaboration with Prof. Kevin Healy's lab at UC Berkeley, who developed the method for grafting PEG on glass surfaces. Prof. David Castner's lab (UW) provided the Secondary Ion Mass Spectroscopy (SIMS) surface analysis to show that, after etching, the PEG is indeed removed:

For more information, please read:

Tourovskaia, A., Barber, T., Wickes, B., Hirdes, D., Grin, B., Castner, D. G., Healy, K. E., and Folch, A. "Micropatterns of Chemisorbed Cell Adhesion-Repellent Films Using Oxygen Plasma Etching and Elastomeric Masks", Langmuir 19, 4754 (2002).

The ability to produce three-dimensional (3D) microstructures is of increasing importance in the miniaturization of mechanical or fluidic devices, optical elements, self-assembling components, and tissue-engineering scaffolds, among others. Traditional photolithography, the most widely used process for microdevice fabrication, is ill-suited for 3D fabrication because it is based on the illumination of a photosensitive layer through a "photomask" - a transparent plate that contains opaque, unalterable solid-state features -, which inevitably results in features of uniform height.

We have devised photomasks in which the light-absorbing features are made of fluids (i.e. dyes). We term them "microfluidic photomasks" (mFPMs) because the photomask features can be addressed (i.e. altered) by means of microfluidic channels. The mFPMs are sheets of the transparent elastomer polydimethylsiloxane (PDMS) that enclose microfluidic channels. Unlike in conventional photomasks, the opacity of the mFPM features can be tailored to an arbitrary number of grayscale levels and their spatial pattern can be re-configured in the time scale of seconds. Importantly, the mFPMs are advantageous in that they can be manufactured rapidly and inexpensively. For selected applications, mFPMs offer a low-cost alternative to present grayscale photolithography approaches.

The mFPMs can be used for the inexpensive fabrication of photoresist patterns that contain features of multiple and/or smoothly-varying heights. For a given mFPM, the developed photoresist pattern can be predicted as a function of the dye concentrations and photomask dimensions.

A photoresist pattern can then be used, for example, to make unique devices. Shown below is a microfluidic device featuring three channels abutting to one channel (flow is from right to left). The three channels feature "walls hanging from the roof" that gradually disappear until the three channels merge:

This work was performed by Chihchen Chen, a Ph.D. student in the lab, and Danny Hirdes, a research assistant. The paper was publised in the Proceedings of the National Academy of Sciences and was subsequently featured in the New York Times, Materials Today, Science News, Physics World, Photonics Spectra, Biophotonics International, and Physics Web.

For more information, please read:

Chen, C., Hirdes, D., and Folch, A. "Gray-Scale Photolithography Using Microfluidic Photomasks", Proceedings of National Academy of Sciences 100, 1499 (2003).

Living systems are sustained by complex, highly coupled biochemical reactions;hence, their response to manipulation with multiple chemical or biochemical compounds often can not be extrapolated from tests with single compounds. Therefore, tests of potentially coupled and non-linear responses (e.g., in drug testing or biological assays) typically include multiple combinations of multiple dilutions. As our understanding of living systems expands, so does the number of combinations to be tested, and conventional large-scale testing approaches based on robotic fluid handling and multi-well plates become increasingly expensive, impractical, or infeasible.

We have devised a laser-cut plastic microfluidic mixer with four different flow levels that produces all the mixture combinations of a given number of dilutions of the input compounds. Our proof-of-concept device generates four titrations of two dye solutions, blue (B) and yellow (Y), and combinatorially mixes the blue titrations (B0, B1, B2, and B3) with the yellow titrations (Y0, Y1, Y2, and Y3) to deliver the sixteen mixture combinations in separate outlet microchannels.

The mixer contains 51 "serpentine mixers". Each serpentine channel twists the flow lines left, right, up, and down to enhance mixing. The degree of mixing can be quantitated and compared with a fluid dynamics model. The modeling was performed by Zach Tyree in Prof. Bruce Finlayson's group at UW (ChemE dept.):

The mixing scheme can be straightforwardly generalized to larger number of titrations and of compounds, and has broad applicability in high-throughput combinatorial testing applications such as drug screening, cell-based biochemical assays, lab-on-a-chip devices, and biosensors.

This work was performed by Chris Neils, a postdoc in the Folch lab.

For more information, please read:

Neils, C. M., Tyree, Z., Finlayson, B., and Folch, A. "Combinatorial Mixing of Microfluidic Streams", Lab on a Chip 4 (4), 342 (2004).

The goal of this project was to "marry" two seemingly incompatible technologies. Present microfluidic devices suffer from an inherent limitation for a certain class of cell-based studies: the inaccessibility of cells cultured in closed channels precludes the use of micropipettes. We have demonstrated the integration of microfluidic and micropipette technologies in one same platfrom consisting of microfluidic streams that are open to air (i.e. microchannels without a roof or "microcanals") and that, as a result, are uniquely compatible with micropipette manipulation and probing of single cells.

While micropipette operation is labor-intensive and yields low throughputs, micropipettes are, on the other hand, readily available in vast numbers of biomedical laboratories and effectively constitute an inexpensive nanotechnology tool (apertures of submicron diameter can be straightforwardly produced with a commercial pipette puller) of unique versatility: they have been widely used for decades for single-cell injection, patch-clamp electrophysiology, iontophoretic stimulation, and "puffing" (pressure ejection to form gradients) of signaling factors. In addition, micropipettes can be used for mechanical manipulation of single cells, e.g. to bring a given cell into contact with other cells for secretion studies or to sever a portion of the cell membrane for cellular transport studies, among other applications.

We have been able to demonstrate two micropipette functionalities (intracellular injection and patch-clamp recordings) in one same microfluidic platform. In the images shown here, flow is from top to bottom; in the left image, the microcanal area starts at the rounded edge:

Although the micropipette/microcanal technique necessarily reduces the high-throughput testing benefits that are typical of microsystems, it still takes advantage of other salient features of microfluidic devices such as small size, minimal reagent consumption, laminar-flow regime, and automated mixing/titration of complex solutions. Below is an example of a set (two only shown) of parallel laminar-flow streams, each featuring three components (water-dye-water). The orange dye stream is hydrodynamically focused by the two water side streams (flow is from left to right). Note how the laminar flow regime is preserved after the microchannel becomes the microcanal.

This work was performed by Chia-Hsien Hsu and Chihchen Chen, Ph.D. students in the Folch lab.

For more information, please read:

Hsu, C.-H., Chen, C., and Folch, A. "'Microcanals' for Micropipette Access to Single Cells in Microfluidic Environments", Lab on a Chip 4, 420 (2004).

Our approach to build 3D (i.e. multilayer) microfluidic devices is to stack sheets of microstructured polymer sheets. When the polymer is PDMS, the sheets are micromolded by pressing a polyethylene (PE) sheet between the master wafer (covered with a few droplets of PDMS pre-polymer) and a glass plate. The goal of applying pressure is to fully exclude PDMS from the areas where the master contacts the PE sheet. Conveniently, when the PDMS is cured and the PE sheet is peeled off, the PDMS microstructures trapped between the PE sheet and the master "prefer" to stick to the PE sheet. The PDMS microstructures then are exposed to oxygen plasma, which activates the PDMS surface; when this PDMS surface is contacted with clean glass (or PDMS), the PDMS binds to the glass (or PDMS) surface. The PE can then be peeled off, leaving the PDMS layer on the surface. The PE sheet allows for precise alignment and visualization of the PDMS pattern during alignment.

The micromolding-by-exclusion process is repeated for each layer. By doing two-layer photolithography, as shown above, it is possible to produce in one molding step both a fluidic layer and the via holes that connect that layer to the next. An example of a complex device with three inlets and two fluidic layers (three PDMS layers) is shown below. The left inlet delivers blue dye and the right inlet delivers red dye at the top fluidic layer. The middle inlet delivers water at the bottom fluidic layer. At selected via holes (elliptical shape), the blue and red dye go down one level to join the water stream, producing a three-component laminar flow stream.

We have used the laser-cut Mylar laminates to build complex mixers with up to four fluidic layers (nine laminates). The work on 3D mylar microfluidic devices was performed by Chris Neils, a postdoc in the Folch lab. The work on 3D PDMS devices was performed by Chia-Hsien Hsu, a Ph.D. student in the Folch lab.

For more information, please read:

Neils, C. M., Tyree, Z., Finlayson, B., and Folch, A. "Combinatorial Mixing of Microfluidic Streams", Lab on a Chip 4 (4), 342 (2004).

PDMS Nanochannels

For more information, please read:

T.F. Kosar, C. Chen, N. Stucky, and A. Folch, "Arrays of Microfluidically-Adressable Nanoholes", Journal of Biomedical Nanotechnology (2005), in press.

N. Stucky, C. Chen, T.F. Kosar, and A. Folch, "Fabrication of Microfluidically-Accessible Planar Nanoholes on Elastomeric Substrates", Journal of Biomedical Nanotechnology (2005), in press.

The use of polymers for the fabrication of microdevices is attractive because polymer parts can usually be mass-produced at low cost by replica-molding and because their physicochemical properties can be tailored by appropriate monomer choice. However, most micromolding processes are limited by the range of features that can be produced in the master - with standard photolithography, features of uniform height and vertical sidewalls - and by the need of a different mold for each device design.

We have devised elastomeric "micro-tunable molds" whose features (cavities topped with elastomeric membranes) can be individually deformed (or "tuned") by selective pressure application. The pressure settings (see values in kPa below), and not just the original master's topography, determine the mold topography. Many replicas with dissimilar microstructures can be created from one mold:

The method is straightforwardly implemented and allows for creating microstructures with multiple heights and rounded profiles:

The micro-tunable molds provide a simple route for microfabricating structures that are difficult, very expensive, or impossible to produce with existing methods:

The deflections can be predicted using finite-element modeling tools (FEMLAB) and hyper-elasticity theory:

This work was performed as volunteer work by John Hoffman and Jian Shao, presently graduate students in the Bioengineering and Mechanical Engineering departments, respectively. The modeling work was done by Chia-Hsien Hsu, a Ph.D. student in the lab.

For more information, please read:

Hoffman, J., Shao, J., Hsu, C.-H., Folch, A. "Elastomeric Molds With Tunable Topography", Advanced Materials 16, 2201 (2004).

Microfluidic devices are critical components of many biomedical and bioanalytical instruments and allow for probing chemical and physical phenomena at the micron scale. Most current microfluidic devices, however, have the inherent functional limitation that the cross-sectional channel geometry, which determines the fluid flow patterns, is essentially constant at any given point in the channel. We have demonstrated microfluidic channels that contain topographical features whose size can be tuned by the user in real time. The topographical changes occur when a membrane at the floor of the microchannels is deflected (to a tunable height) by pressure application. The membrane deflections can be used to alter the laminar flow patterns in microfluidic mixers:

The microtopographies can also be used to physically trap small fluid volumes within microchannels. The sequence of optical images below illustrate a series of steps in which "doughnut"-shaped membranes were pressurized (as indicated by black arrows) to trap various dyes, after the microchannel was sequentially filled with each dye (color arrows):

This work was performed by Chia-Hsien Hsu, a Ph.D. student in the lab.

For more information, please read:

C.-H. Hsu and A. Folch, "Microfluidic Devices with Tunable Microtopographies", Applied Physics Letters 86, 023508 (2005).

Cellular micropatterns are important for biosensors, tissue engineering, and cell biology, to name a few applications. Typically, the cells attach to an adhesive region and are surrounded by a non-adhesive background. However, this differential adhesiveness is not easily maintained because the substrate can change (e.g. due to protein adsorption or secretion of proteases by cells), specially in the presence of protein-rich, serum-containing medium. We have published a simple technique for confining cells to micropatterns of adhesive proteins for periods of time longer than 1 week. In this method, a protein- and cell-repellent thin-film containing poly-ethylene glycol (PEG) is grafted onto a glass substrate. The PEG-derivatized substrate is covered with an elastomeric mask (i.e. microchannels or a stencil), which selectively protects the substrate during a subsequent oxygen plasma etch. The etched areas, essentially bare glass, become amenable to protein adsorption and cell attachment. An added advantage of this technique is that it allows for culturing the cells in serum-containing medium, so -- unlike methods based on alkanethiol monolayers on gold or on microstamping of proteins -- it is compatible with virtually all known protocols for cell seeding and culture used by biologists for decades.

A variation of the technique uses elastomeric stencils (essentially, rubber membranes featuring holes) and allows for making cellular islands. In the color picture below, the substrate contains three different degrees of adhesivity: a fibronectin circle (green, very adhesive -- where the cells initially attach), an albumin ring (red, temporarily non-adhesive -- over hours and days, depending on the medium, cells will eventually spread over it), and a PEG-IPN background (very non-adhesive -- we have never seen cells spreading over it). The grayscale picture below shows how the muscle cells have spread over the albumin area but have not been able to cross the PEG boundaries.

A more straightforward use of the stencils is to directly mask the deposition of proteins or cells on plain glass or polystyrene substrates without using surface chemistry; here, the substrate is a polystyrene tissue culture dish, the cells are labeled blue, fibronectin is labeled green, and albumin is labeled red:

This method was developed by Anna Tourovskaia, a Ph.D. student in the lab, in a collaboration with Prof. Kevin Healy's lab at UC Berkeley, who developed the method for grafting PEG on glass surfaces. Prof. David Castner's lab (UW) provided the Secondary Ion Mass Spectroscopy (SIMS) surface analysis to show that, after etching, the PEG is indeed removed.

For more information, please read:

Tourovskaia, A., Barber, T., Wickes, B., Hirdes, D., Grin, B., Castner, D. G., Healy, K. E., and Folch, A. "Micropatterns of Chemisorbed Cell Adhesion-Repellent Films Using Oxygen Plasma Etching and Elastomeric Masks", Langmuir 19, 4754 (2002).

Conventional cell culture techniques have changed little over the last several decades and essentially consist of growing cells on a homogeneous large surface (polystyrene or glass dishes or wells) immersed in a homogeneous bath. However, in vivo cells respond to spatially and temporally organized signals in the surrounding substrate and fluidic microenvironment. A striking example is the differentiation of muscle cells at the neuromuscular junction.

Reliable fluidic control is particularly paramount for perfusion-based cell differentiation studies because differentiation processes are biochemically delicate and they take several days to complete. We have reported a microfluidic system that allows for maintaining cell cultures for more than two weeks using gravitational flow. As a test, we demonstrate that the microfluidic culture supports C2C12 cell growth and differentiation. C2C12 cells are myogenic cells that in vitro can be induced to fuse into multinucleated myotubes, providing a test bed for cell differentiation. Furthermore, the system can be combined with cell-adhesive and cell-repellent micropatterns to produce cellular microstructures within the microfluidic environment. Over one week, the tracks of C2C12 cells transform into precisely aligned/oriented myotubes. Thus, the cellular micropatterns allow for controlling the width, spacing, and orientation of the myotubes with respect to the flow. Long-term perfusion at such low flow rates did not interfere with cell differentiation and was not observed to affect cell morphology or growth.

Long-term viability is of great interest not only for differentiation studies but also for a variety of applications encompassing cell-based biosensors, high-throughput assays, and basic cell biology research. With this setup, the target cells can be focally labeled by a fluorescent tag and/or selectively exposed to biologically-relevant molecules to study spatiotemporally-confined extracellular stimuli.

The image below demonstrates the regioselective tagging of muscle cell surface receptors/ligand-gated channels, which are nicotinic acetylcholine receptors (AChRs). During skeletal muscle development, the expression of AChRs is amplified just before the fusion of myoblasts and the AChRs are shuffled to eventually become localized to the site of muscle-nerve contact. Soon after myotube fusion, clusters of AChRs are formed, which can be used as an early muscle differentiation marker. Snake toxin a-bungarotoxin (BTX) binds irreversibly to the a-subunit of the AChR complex and is lethal to the cells at high concentrations. The image below demonstrates the live labeling of AChRs (with a fluorescent conjugate of BTX, BTX*) confined to the central zones of the myotubes. This subcellular labeling results in a sub-population of tagged receptors that could in principle be followed throughout their life span at the plasma membrane surface until subsequent internalization:

This work was performed by Anna Tourovskaia, a Ph.D. student in the lab, and Xavier Figueroa-Masot, a postdoctoral fellow.

For more information, please read:

A. Tourovskaia, X. Figueroa-Masot, and A. Folch, "Differentiation-on-a-chip: A microfluidic platform for long-term cell culture studies", Lab on a Chip 5, 14 (2005).

Microvalves are important in our microfluidic cell culture systems because they allow for isolating small cell populations and for accurately timing the release of biochemical compounds into the cells' vicinity (for chemotaxis studies, for example). All our microvalve devices consist of two PDMS microchannel circuits (both filled with fluid) separated by a PDMS membrane. When pressurized, the bottom circuit serves to deflect the PDMS membrane, which affects the routing of fluids through the top circuit (where cells are). In their passive state, the microvalves are closed and opening requires application of suction; reversing to positive pressure closes the valve again, at which point the positive pressure can be released.

The same pressure line can be used to actuate a microarray of valves, or a portion of the microarray. A set of microchambers initially isolated can be connected by opening the valves (i.e. bending the membrane down). The largest chamber below is around 4.5 nL. Unlike other elastomeric membrane designs, this design allows for controlling volumes on the range of femtoliters because the volume of the closed chamber is defined photolithographically.

Similarly, a microvalve array can be used to store sub-nanoliter-sized volumes and to automatically generate many mixtures with one mixing step:

This work was performed by Greg Boggy, Chia-Hsien Hsu, and Tom Keenan, Ph.D. students in the lab, and postdoc Nianzhen Li.

For more information, please read:

N. Li, C.-H. Hsu, and A. Folch. "Parallel mixing of Photolithographically-Defined Nanoliter Volumes Using Elastomeric Microvalve Arrays", Electrophoresis 26, 3758 (2005).

A major goal in neuroscience is to understand the formation and development of synapses, the tiny membrane specializations that enable nerve cells to communicate with each other. The sequence of molecular signals leading to synapse formation ("synaptogenesis") is qualitatively well known for the more accessible neuromuscular synapse. It is well established that, immediately after contacting the muscle cell, the nerve terminal secretes agrin to stimulate and/or stabilize the clustering of acetylcholine receptors (AChRs) at the postsynaptic site.

Very little is known of the quantities (concentration, duration, onset, etc.) of the various neurochemical signals involved in synaptogenesis. Importantly, all except for one of the axons innervating a given myotube at birth retract after a period of a week or so according to a synaptic competition process that remains, for lack of quantitative methods, poorly understood. Such quantitative description is lacking because present experimental setups for the study of the neuromuscular junction do not allow for a precise control over the many variables involved in synaptogenesis.

We culture muscle cells and let them grow until they fuse and differentiate, expressing AChRs. If no agrin is added to the culture, the distribution of AChRs is homogeneous, whereas the AChRs form clusters if agrin is added:

We use microdevices to focally expose the myotubes to agrin, in an attempt to mimic the presynaptic neuron during innervation. Focal exposure is done in one of two ways: 1) placing the cells under a heterogeneous, laminar-flow stream containing agrin at the center:

or 2) placing the cells atop a nanofabricated aperture:

Both methods have advantages and disadvantages, but equally prove that local exposure to soluble agrin entices the cells to cluster AChRs only at the site of agrin stimulation. An additional advantage of these methods is that they allow for obtaining many data points per experiment.

This work was performed by Anna Tourovskaia, a Ph.D. student, and by Xavier Figueroa-Masot, a postdoctoral fellow. This project was started in collaboration with Profs. Stan Froehner and Marv Adams (UW, Physiology and Biophysics Dept.)

For more information, please read:

Tourovskaia, A., T.F. Kosar, and Folch, A."Local Induction of Acetylcholine Receptor Clustering in Myotube Cultures Using Microfluidic Application of Agrin", Biophysical Journal 90, 2192 (2006).

T.F. Kosar, Tourovskaia, A., Figueroa-Masot, X., Adams, M., and Folch, A. "A Nanofabricated Planar Aperture as a Mimic of the Nerve-Muscle Contact During Synaptogenesis", Lab on a Chip 6, 632 (2006).

During embryonic development, axon tips (growth cones) are guided through a dynamic three-dimensional (3-D) landscape by soluble chemotropic factors ("non-contact" cues), as well as by immobilized, growth-permissive or growth-inhibiting guidance molecules present in the extracellular matrix and on the surface of surrounding cells ("contact cues"). Since most in vitro axon guidance studies have utilized two-dimensional (2-D) substrates, it has been difficult to probe the search algorithms used by the growth cone to respond to multiple contact cues during 3-D navigation.

We have undergone an in vitro study of neuronal growth in which the axons of murine embryonic cortical neurons are challenged with competing growth options, using substrates that feature variations in permissiveness and microtopography. Neurons can be guided by surface chemical micropatterns (here, poly-D-lysine or PDL) or by microtopography (here, a PDMS stamp):

The neuron on the left clearly prefers to grow on PDL, and the axon of the neuron on the right prefers to stay inside the PDMS well. On close inspection of the microtopographical preferences of neurons, one finds that axons prefer to grow as straight as possible within a distance consistent with the exploratory range of the growth cone, which suggests that prior to turning the neuron explores the step in a 3D way:

What happens if the two cues (surface chemistry and topography) compete with each other?

To address this question, we used poly-D-lysine (PDL) coatings on microfabricated steps of polydimethylsiloxane (PDMS), and complementary features of Matrigel, a basal lamina extract. We found that, as a general rule, axons display a preference for PDL over Matrigel, not only in 2D but also in 3D:

But, if the axon was forced to try to stay on a PDL-coated surface that bent too much, then the axon would rather keep going straight into Matrigel:

Taken together, these observations indicate that growth cones make 3-D navigation decisions based on a combination of permissiveness and topographical cues.

This work was performed by Nianzhen Li, a postdoctoral fellow, and Tom Keenan, a Ph.D. student in the lab.

For more information, please read:

N. Li and A. Folch, Experimental Cell Research 311, 307 (2005).

Glass micropipettes are inexpensive, they are readily available in many biomedical laboratories, and have been widely used for decades for single-cell injection, patch-clamp electrophysiology, ionophoretic stimulation, and "puffing" (pressure ejection) delivery of gradients of signaling factors, among other applications. Unfortunately, micropipette techniques generally require substantial operator expertise and suffer from low throughput and high failure rates; expensive micromanipulators are also required to precisely position the micropipette, and mechanical vibration and drift are common sources of failure in electrophysiological applications. Additionally, due to the fragility and bulky geometry of micropipettes, they are difficult to incorporate into arrays.

In contrast to the heat-and-pull simplicity of producing micropipettes, planar nanoholes can befabricated using microfabrication techniques such as ion track etching, ion-beam sculptin, electron-beam lithography (EBL), conventional silicon-based techniques or elastomeric micromolding of a point-contact in PDMS. Planar nanoholes have been used for cellular immunoisolation, biomolecular separation, chip-based patch-clamping, lipid bilayer-based single ion channel studies, and show potential for use in single-molecule DNA sequencing.

We have developed a robust, optically transparent device that incorporates a planar array of nanoholes that can potentially be used for electrical or pharmacological interrogation of large numbers of cells. The nanoholes are defined using EBL, which is commonly available in most research universities and is capable of generating features with dimensions on the order of tens of nanometers. Additionally, the EBL patterns are software-generated and not in physical form like in other lithographic techniques, allowing for easy and inexpensive scaling and design iteration. Most of the fabrication steps are appropriate for large batches, which reduce fabrication time and lower production costs.

Each nanohole in the array is microfluidically accessible and individually addressable from both sides, allowing for rapid perfusion, delivery, and exchange of fluids. Each nanohole is electrically isolated from all the others in the array as well as from the rest of the substrate - essentially behaving like the nanohole of a micropipette -, which makes them suitable for low-current measurements of ion diffusion through the nanoholes.

	Additionally, our device is biocompatible to the extent that cells (fibroblasts as well as chinese hamster ovary cells) can be cultured on it for several days:

Compared to other planar nanohole approaches, our device is inexpensive, scalable, simple to use, and straightforwardly fabricated as a reliable microarray for highly parallel measurements. Because of the optical transparency of thin LSN membranes, fluorescence and photometric techniques can also be used. Additionally, the small volume and addressability of the microwells and channels allows for patch clamping to be combined with (or replaced by) focal delivery of biomolecules or pharmacological agents. We have used the planar nanoprobles in its reverse setup to study neutrophil chemotaxis:

The neutrophils are seen to migrate towards the surface hole:

A caveat of these silicon-based devices is that they are expensive. Hence, we have developed methods to mold nanoholes in PDMS:

This work was performed by Fettah Kosar, Chihchen Chen, and Nick Stucky, Ph.D. students in the Folch lab.

For more information, please read:

T.F. Kosar, C. Chen, N. Stucky, and A. Folch, "Arrays of Microfluidically-Adressable Nanoholes", Journal of Biomedical Nanotechnology 1, 161 (2005).

N. Stucky, C. Chen, T.F. Kosar, and A. Folch, "Fabrication of Microfluidically-Accessible Planar Nanoholes on Elastomeric Substrates", Journal of Biomedical Nanotechnology 1, 384 (2005).

Leukocyte migration is essential for the innate immune response to microbial pathogens and immune system organization. The migration of leukocytes from the vasculature into tissue involves a number of different molecules including adhesion molecules and chemoattractants. Chemokines, a family of endogenous chemotactic factors made up of small basic peptides (8-14 kDa), are responsible for the directed migration of leukocytes from the bloodstream into tissue. CXCL8, a member of the CXC family of chemokines (historically known as Interleukin-8 or IL-8), is an important neutrophil chemotactic factor.

We have developed a microfluidic gradient generator whereby a chemoattractant is released to a cell-containing chamber through a microvalve. The device is fabricated using soft lithographic techniques and is formed of three layers: the top layer where the fluids and the cells are handled, a middle layer consisting of a PDMS membrane, and a bottom layer to actuate the PDMS membrane at desired locations by means of microfabricated pneumatic lines:

To measure neutrophil chemotaxis, we acquire phase-contrast and fluorescence microscopy images (to capture neutrophil movement and gradient formation, respectively) at 2.5 minute intervals for two hours. The 10 kDa fluorescent Dextran added to the CXCL8 solution is used as an indicator for CXCL8 concentration and is assumed to diffuse at roughly the same rate as CXCL8 based on the similarity of their molecular weights.

Cell migration is digitally tracked using a MATLAB script developed for this purpose. The sequence clearly shows that the gradient develops with approximately radial symmetry around the source and that the cells display a biased migratory response toward the chemoattractant source. This bias is even more apparent when the trajectory of each cell is transformed to radial coordinates (taking the radial axis for all time points as the initial cell-source axis for each cell), as shown on the image below (left). We observe that 1) neutrophil migratory response is, on average, biased towards the source, 2) significant deviations from the initial axis occur often; and 3) there is considerable heterogeneity in the migration speeds and directions amongst the neutrophil population.

As gradient formation proceeds, the concentration at any given location in the device increases with time, as shown on the image above (right) for one neutrophil's trajectory. Thus, in order to help the user find quantitative correlations between cell behavior and the local CXCL8 concentration, the software tool outputs various plots for any given cell n as a function of the local CXCL8 concentration (rather than time), such as plots (for cell n) of the speed Vn, radial velocity Vrn along the instantaneous cell-source axis, and the cosine of the migration angle.

The software tool can also output histograms of V, Vr, and turning angle for any chosen bracket of concentration values. The histograms of radial velocities were clearly shifted to non-zero positive values (migration biased towards the source) for low CXCL8 concentrations but symmetrical around 0 (random migration) for high concentrations. This behavior is confirmed in Rosetta histograms of the turning angle. Histograms of the speed reveal a considerable heterogeneity in migratory speeds, with average values around 2.5 microns/min at the lowest concentration bracket but shifting towards zero at the highest concentration brackets, suggesting that the neutrophils are switching from chemotaxis (directed migration) to chemokinesis (random migration).

Further work to address the biological implications and origin of this striking difference in migratory behavior at different concentration brackets is underway.

This work was performed by Greg Boggy and Chia-Hsien Hsu, Ph.D. students in the lab. Neutrophils were provided by our collaborator Chuck Frevert's group at UW.

For more information, please read:

Frevert, C.W., Boggy, G., Keenan, T.M., and Folch, A. "Measurement of Cell Migration in Response to an Evolving Radial Chemokine Gradient Triggered by a Microvalve", Lab on a Chip 6, 849 (2006).

The analysis of single cells at high throughput offers the possibility of obtaining statistically-rich measurements of cell behavior variability as well as the opportunity to detect abnormal cells amongst a large cell population. We are developing tools to identify and single out cells among a large set of cells. In our preliminary studies, we have produced microarrays of microwells, each of which fits a small, discrete amount of single cells. The procedure is inspired from everyday life:

The principle is deceivingly simple: once the cells fall in the wells, dislodging the cells is difficult because it requires fluid to penetrate the remaining space -- which is, by then, very tight. Since the microwells are made of PDMS, the cells can be imaged in fluorescence and phase-contrast microscopy, as shown in the overlays below:

In the image above, two populations of fibroblasts (labeled with the live dyes Cell Tracker Green and Cell Tracker Red, respectively) were seeded simultaneously. The image was accepted as the cover of a recent issue in Lab on a Chip:

Large arrays are straightforward to produce and to image in a variety of imaging modes. Shown below is a microarray of 12,500 microwells; around 90% of the microwells are occupied by live cells (green) and only a few are occupied by a dead cell (red).

The scale bar is 500 microns. The image was acquired with a large-area CCD camera (Hamamatsu-HR). Despite the large number of cells, the number of pixels devoted to each cell is sufficient to detect faint fluorescent signals at a single-cell level. This cell was monitored during the uptake of a fluogenic esterase activity marker (CellTracker Green by Molecular Probes):

When thousands of cells are monitored in parallel (using a MATLAB image recognition script), the cell responses show high heterogeneity:

We envision that these arrays will be used as "Cellomics" chips that will allow for probing a variety of cell functions using standard biomarkers.

This work was performed by Jackie Rettig, a masters student in the lab.

For more information, please read:

J.R. Rettig and A. Folch, "Large-Scale Single-Cell Trapping and Imaging Using Microwell Arrays", Analytical Chemistry 77, 5628 (2005).

The ability to control the spatiotemporal microenvironment of cultured cells, such as cells in soluble gradients of signaling molecules, is paramount for understanding the quantitative relationship between physicochemical signals and the induced biological responses. Several methods to generate gradients of biomolecules over cells exist, but they all entail potentially confounding or cell-adverse conditions and/or result in poor reproducibility.

We have developed a device containing nanofluidic "fountains" acting as "jets" that inject picoliter amounts of fluid into an open pool with negligible exposure of the surface to flow. The microjets are formed by replica-molding of a biocompatible elastomer (PDMS) and transfer-molding onto a glass cell culture surface.

In essence, the microjets act as surface-bound pipettes or "nanofountains", where fluid emanates from a nanohole (fabricated in the mold by electron beam lithography):

A stable, reproducible gradient can be formed in minutes, depending on the distance between the microjets and the number of microjets, but independently of the molecular weight. Flat-front gradients can be formed using a microarray of microjets:

By changing the pressure settings, the gradients can be steered or changed in slope. Other, more complex gradients can be formed by using more complex arrangements of microjets. Importantly, the surface does not "see" appreciable flow because it is mostly directly upwards, so we believe it will be very useful for exposing cells to gradients of soluble signaling factors.

For more information, please read:

Keenan, T.M., Hsu, C.-H., and Folch, A. "Microfluidic "Jets" for Generating Steady-State Gradients of Soluble Molecules on Open Surfaces", Applied Physics Letters 89, 114103 (2006).

In collaboration with Paolo Vicini's lab in the Bioengineering Dept. here at the University of Washington, we are developing image recognition tools to analyze cell morphology at high throughput, i.e. from time-lapse imaging movies. The image below shows a cultured embryonic cortical neuron in the process of growing an axon. The image recognition algorithm recognizes the tip of the axon (white square) and tracks it on subsequent images, allowing for generating data (e.g. axon growth speed, turning angle, etc.) from large numbers of images. The graph plots the position of the computer-recognized axon tip for each image; red triangles indicate a forward axon growth and the blue triangles indicate axon retraction.

Here is an animated GIF showing some of the frames of the time-lapse movie and its corresponding axon tip recognition plot:

The neuron culture and isolation was performed by Nianzhen Li (postdoc) and Tom Keenan (Ph.D. student). The movies typically take ~10-20 hours to acquire (these neurons grow slowly), and are taken with a phase-contrast/fluorescence microscope encased in a custom-made environmental cage with temperature (37 oC) control; a small plexiglas box is fitted over the sample and perfused with 5% CO2, with an open dish filled with water that acts as a humidifier. The cells are exposed to light only during image acquisition (a fraction of a second for each image), when the image acquisition software (Metamorph) activates the lamp shutter. The tip recognition algorithms were developed in Matlab by Ph.D. students Mary Spilker and Andy Hooker in the Vicini lab and by Tom Keenan in the Folch lab.

For more information, please read:

Keenan, T.M., Hooker, A., Spilker, M. E., Boggy, G. J., Li, N., Vicini, P., and Folch, A. "Automated Identification of Axonal Growth Cones in Time-Lapse Image Sequences", J. Neurosci. Methods 151, 232 (2005).

In a traditional cell culture setup (i.e. cells seeded in petri dishes and immersed in a homogeneous bath on the order of ~1 mL in volume), long-term cell culture studies demand large amounts of cells (and its associated animal suffering if primary cells are used), supplies (petri dishes, media), precious reagents, and painstaking human labor (to periodically exchange medium).

Microfluidic technology offers the possibility to automate cell culture procedures and, at the same time, allowing for 1) minimal consumption of costly reagents, 2) large number of experiments per unit area, 3) small number of cells required to perform the experiment, and 4) perfusion of cells with heterogeneous laminar-flow streams with subcellular resolution. With notable exceptions, most work to date has reported short-term cell cultures, which have less stringent cell maintenance requirements (e.g. pH stability, constant gas and nutrient concentration) than long-term cultures. Our long-term perfusion setup is simple and can be transferred to any biology laboratory equipped with a cell culture incubator. Long-term viability is of great interest not only for differentiation studies but also for a variety of applications encompassing cell-based biosensors, high-throughput assays, and basic cell biology research.

The microfluidic cell culture device is made by molding in poly(dimethylsiloxane) (PDMS) using soft lithography techniques. The device contains two microchannel networks perpendicular to each other: a) the "main channel" contains the cells and starts at the point of convergence of three inlet channels, which are used to switch liquids required for short biochemical assays and also serve the purpose of generating heterogeneous laminar flow streams; b) the 16 "side channels" orthogonal to the main channel are used to perfuse the cells during long term cultures. A microfluidic perfusion network design that utilizes segments of high flow resistance to limit the flow rate allows for obtaining reliable low flow rates using moderate head pressures.

While traditional non-microfluidic (static) cultures can be straightforwardly maintained by occasional medium replacements, microfluidic cultures cannot be treated as static because the average volume of medium per cell is typically one order of magnitude smaller than in macroscopic cultures; constant flow poses a challenge because the medium must be pre-warmed (for mammalian cells) and pre-equilibrated with specific gas compositions. Our microfluidic system allows for maintaining cell cultures for more than two weeks using gravitational flow. Furthermore, the system can be combined with cell-adhesive and cell-repellent micropatterns to produce cellular microstructures within the microfluidic environment.

With this setup, the target cells can be focally labeled by a fluorescent tag and/or selectively exposed to biologically-relevant molecules to study spatiotemporally-confined extracellular stimuli. Various examples of regioselective fluorescent labeling using live "CellTracker" dyes are shown below:

This work was performed by Anna Tourovskaia, a Ph.D. student in the lab, and Xavier Figueroa-Masot, a postdoctoral fellow.

For more information, please read:

A. Tourovskaia, X. Figueroa-Masot, and A. Folch, "Differentiation-on-a-chip: A microfluidic platform for long-term cell culture studies", Lab on a Chip 5, 14 (2005).