Author: John Crabtree

Onur Gökçe at University of Zurich/ETH and Yuksel Temiz and Emmanuel Delamarche at IBM in Zurich, Switzerland, together with Samuel Castonguay and Thomas Gervais at École Polytechnique in Montréal, Canada recently authored a clever article in Nature that shows beautiful control of dried reagent dissolution and diffusion into the liquid buffer flowing over-top of the reagents in a microfluidic channel.

The authors demonstrated that, by including a partial-height longitudinal wall to act as a “capillary pinning line”, liquid flows through an empty channel and back along the adjoining channel with spotted reagents to create zones of dissolved reagents – a process they term ‘self-coalescing flow’ or SCM.  This is in contrast with the frontal accumulation of  reagents spotted conventionally along a single channel.  Further, by adjusting channel dimension, the concentration profiles could easily be manipulated to produce short zones of high-concentration or longer zones of lower concentration, with fairly homogeneous concentration profiles in all cases.  Several additional examples showed the ability to either segregate or coalesce such zones for a given reagent, mix and react with downstream reagents, etc.

From the perspective of microfluidic product development, this development is very interesting for several reasons.  First, it provides this control of reagents in solution with very simple channel structures that can be fabricated existing standard methods for any substrate material (plastic, glass, silicon).  Secondly, it may obviate the need for more complex channel networks and associated pumps and valves required to physically segregate parallel reactions of a single analyte with several target reagents for multiplexed analyses, common in many microfluidic products currently.  This could reduce the size, complexity and cost of the chip consumable, making the price and profitability of such products much more attractive.

Researchers from Wanida Wonsawat‘s group at Suan Sunandha Rajabhat University (Bangkok) and Takashi Kaneta‘s group at Okayama University recently published an article in Nature Research showing a relatively simple but elegant procedure to greatly increase the stability of hydrogen peroxide reagent absorbed into the fibres of the patterned Whatman filtre paper used the paper microfluidic devices in the study.  Hydrogen peroxide is an oxidising agent that is often used in enzymatic assays that may create or remove a separate species that can be measured via indirect optical detection using colourimetry, fluorescence, chemiluminescence, etc.

The paper microfluidic devices (or paper-based analytical devices, PADs) were created by with a commercial wax printer, and the exposed paper areas were treated  with the colourimetric reagents.  For these trials, simple circular areas were patterned for the proof-of-principle studies undertaken, in lieu of channel networks.  The reagents used were 3, 3′, 5, 5′-tetramethylbenzidine (TMB, the substrate and also colourimetric reagent), bovine serum albumin (BSA), phosphate buffer and hydrogen peroxide, with or without the poly(vinyl alcohol) (PVA).  Horseradish peroxidase (HRP, the enzyme ) was contained in the sample solution.  TMB changes to blue in the presence of HRP and hydrogen peroxide.

The authors discovered that blue colour seen upon initial reaction faded after a day, and isolated the cause to decomposition of the hydrogen peroxide on the PAD during this time.  The figure above shows PADs with the enzymatic assay occurring in the patterned dots, where a) both hydrogen peroxide and TMB were added to the PAD, and the results are captured shortly after reaction; or the PAD with only b) hydrogen peroxide or c) TMB is prepared and stored at room temperature for one day, and then the other reagents are added and the results captured.  Degradation of the hydrogen peroxide causes the near total loss of signal in b).

To stabilise the hydrogen peroxide, they added PVA to the hydrogen peroxide solutions prior to addition to the PAD, and found dramatic improvements in hydrogen peroxide stability.  The extent of the improvement varying with PVA concentration and chain length, as well as storage temperature.  2% solutions of PVA with a length of ~1650 monomers extended the life of the reaction at ~100 intensity to 10 days with storage at room temperature, while refrigeration at 4°C extended that to 30 days for PVA lengths of ~1650 or 2000 monomers.  The PVA is thought to protect the hydrogen peroxide by forming a liquid- and air-tight barrier, thus preventing attack by hydroxide anion that catalyses the degradation.

The importance of extending reagent lifetimes in microfluidic devices is a critical building block for the development of point-of-care/use devices.  In this case, isolation and stabilisation of the hydrogen peroxide reagent, in widespread use for enzymatic assays, is of obvious benefit.  30 days of refrigerated storage is much better than a few hours, but a long way from the requirements for a robust product with a minimum 6-month shelf-life.  However, one could imagine product development efforts with currently available packaging technology, e.g. desiccation followed by vacuum packaging or dry inert gas packaging, might easily eliminate the moisture-catalysed degradation, and thus extend product life significantly.  There are also other stabilising polymers to choose from that may be better optimised for other types of reagents requiring a liquid/gas barrier.

I guess it’s not a big stretch to go from paper to wood for microfluidic substrates, but I’m impressed nonetheless!  Govind Rao‘s group in Chemical Engineering at the University Maryland recently published an ASAP article in Analytical Chemistry (alternatively use this DOI) about microfluidic devices made in plywood.

The Rao group machined channels and features in 35 chips laid out on 12″ x 12″ plywood sheets using a CO₂ laser printer.  Devices were treated with a 0.1-1% PMMA or Teflon solution to seal up pores and inhibit capilarity which would otherwise promote seepage of sample and running buffers in the substrate.  Wooden channel layer substrate was bonded with cyanoacrylate glue to wooden cover plate with through-holes accommodating Luer lock fittings; PMMA covers were sometimes used for visualisation.  Channels were ~0.68 mm-deep X 1 mm-wide, and the system tolerated pressures of ~2 psi.  Different laser rastering speeds and line thicknesses were explored.  Identical devices were made out of PMMA for performance comparison purposes.

Runs with blue and red food dye mixed at Y- and T-intersection mixers were performed and imaged to compare performance in a simple flow experiment.  The figure at right shows comparable performance (PMMA in left column, wood in right column), though for both geometries, the wood chips do not provide an even amount of the two dyes (less red dye), and the signal is noisier, as shown by the black standard deviation lines above and below the red average signal lines.  No discussion was found concerning these two discrepancies.

The authors discuss the biodegradable advantage that wooden devices have vs. their polymer or glass counterparts, and how this is becoming increasingly important with new legislation banning single-use plastics coming into force in many jurisdictions.

An interesting article from Ashleigh Théberge’s group at the University of Washington reviews the relatively recent introduction of open capillary microfluidic systems (as distinct from electrowetted digital microfluidics).  The paper was recently pre-published in Analytical Chemistry as an ASAP article.

The authors review a number of different channel geometries and fibre bundle configurations that can be used to promote open capillary flow.  The helpfully provide the equations that use channel geometry and contact angle to determine whether capillary flow is energetically favourable.  This is done for both conventional monolithic (single material) channels, as well as composite material channels.

Pros and cons of the open capillary approach are surveyed.  The authors list several advantages:

1. simplified fabrication by obviating the need for bonding and potential associated use of solvents on the substrate, process development/trade secrets, manufacturing cost, etc.;
2. ease of performing surface modifications, such as for hydrophilicity/hydrophobicity, silanisation or other derivitisation, blanket or patterned exposure to UV, plasmas, chemical or physical vapour depositions (PVD & CVD), application of delicate bio-reagents that can’t withstand thermal bonding;
3. accessibility of channels for adding or removing reagents or components with pipettes, tweezers (as for tissue scaffolds)
4. elimination of air bubble issues, due to the open interface.

They also mention disadvantages primarily stemming from the open channel access such as higher evaporation, evolution and/or exchange of dissolved gases, liquid leaks to non-channel paths, and the inability to generate higher pressures in channels (beyond those of capillarity) and thus use valves, etc.

Lastly, a number of different applications are noted, though all appear to be academic in nature at this stage.  While I see the advantages of flexibility and reduced manufacturing cost offered by the open channel concept, I wonder how a product would be able to mitigate against evaporation and contamination issues in viable approach suitable for a robust consumable suitable for untrained users.  Perhaps the authors have the answers; they mention that they have financial interests in two companies, Salus Discovery and Stacks to the Future, involved in the commercialisation and IP related to some of the technologies presented.

South Korean researchers have recently shown that immature dendrite cells undergo chemotactic migration through microfluidic mazes preferentially towards healthy or cancerous cells versus cell-free medium.  The new research findings, published in a Lab on a Chip article, come from Cho’s group at the Institute for Basic Science, Grybowski’s group at the Ulsan National Institute of Science and Technology, and Jeon’s group at the Pohang University of Science and Technology.

Chemotaxis is the movement of cells towards or away from chemical stimulus (attractants or repellents, respectively).  Bacteria accomplish this through biased ‘random’ walk cycles, where the cells use their flagella to move in a given direction, then stop and sense whether they have moved up or down the stimulant’s concentration gradient to determine subsequent reorientation and straight-line translation.  Migration towards attractants by dendrite cells (surveillance agents and messengers for the immune system) is well documented for mature but not immature dendrite cells.

In this study, immature cells were allowed to migrate towards cell medium (control), EpH4-Ev healthy cells or beta-MEKDD 116 cancer cells.  In one experiment series, comparisons in migration were evaluated by allowing the immature dendrite cells to migrate from a single inlet towards either of two outlets that contained two of the three cell attractants.  Attraction bias was clearly shown to be (beta-MEKDD 116) > (EpH4-Ev) > (cell medium).  In another series of experiments, different cytokines drawn from the cancerous beta-MEKDD 116 cells were compared, and the protein Gas6 was found to have the largest attractive effect.  Large numbers of replicate analyses allowed the authors to nicely quantify the confidence limits that applied to their results.

A recent article in Nature: Microsystems & Nanoengineering by Ralf Hellmann’s group at the Aschaffenburg University of Applied Sciences and Thomas Walther’s group at the Technische Universität Darmstadt in Germany outlines impressive laser machined microfluidic microneedle devices.

Devices were fabricated in PMMA coated in OrmoComp® photoresist in a two-step process.  Direct laser written photoresist microneedle arrays (tip radius of 13-21 µm) and 3-D microfluidic PMMA channels were combined in a device using a single femtosecond pulsed laser for fabrication.  The fabrication process is claimed to be simpler than traditional fabrication methods, but unfortunately, no throughput figures were provided.  It would be nice to know what these are to understand how readily this fabrication technique could be applied as a manufacturing tool.

Devices were characterised with force vs. time curves for injections of a rhodamine B test solution through pork skin (which shares similar morphology to human skin) and to verify microneedle integrity during use.

The authors highlight that the devices could provide a significant leap forward for point-of-care applications including drug delivery and diagnostics.

New research out of Cullen Buie‘s lab at MIT’s LEMI lab shows effective sorting of different strains of bacteria using microfluidics-based dielectrophoretic (DEP) analysis.  The research was published in Science Advances article (DOI: 10.1126/sciadv.aat5664) and also profiled by MIT News; a second MIT news article provides some background to microbial digestion and power generation.

The cell manipulation works by virtue of a DEP trap. A cell suspension flows through a small orifice (~ 50 µm wide) between two microfluidic chambers under an electric field.  The electroosmotic and electrophoretic forces generated by the field determine each cell’s speed in the bulk solution.  At the constriction, there is additionally a dielectrophoretic force that acts to trap cells of a certain polarisability, depending on the applied field strength.  Since polarisability relates to extracellular electron transfer (EET), or a cell’s tendency to generate electricity during respiration, the trap can act as a filtre or screen for bacterial strains based on their tendency to generate electricity.  In other words, different strains can be selectively trapped or screened based on the applied voltage.

The value of being able to discern between the different strains of bacterial in terms of their tendency to generate power while digesting the components in wastewater can hardly be understated.  Apparently 3 of energy in the US is spent on wastewater treatment, while the wastewater itself contains 10 times the energy required for its own processing.

A Technology Networks article based on NIH-funded research describes how organ-on-a-chip microfluidic devices just reached the US’ NASA portion of the International Space Station (ISS), the ISS National Lab, last Wednesday as part of an investigation into the effects of aging, mimicked by weightlessness, on the immune system.  The research team is led by Professor Sonja Schrepfer at UC-SF.

The immune system chips incorporate immune, bone marrow and blood vessel cells.  The physiological effects of space flight on astronauts, and the tissues on these chips, are similar to those of aging, including bone and muscle loss and altered immune systems.  The several dozen chips will be incubated for 2 weeks, and then frozen until they return to Earth for analysis at the Schrepfer lab.  Future SpaceX re-supply missions to the ISS in March and April 2019 will bring kidney, bone/cartilage, and blood/brain barrier chip sets as part of the same research programme.

It is hoped that the research may provide better understanding of, and potentially improved therapies for the process of aging.

Impressive research relating to Alzheimer’s disease (AD) research on a microfluidic platform recently emerged from the Cho group at the University of North Carolina at Charlotte and the Tanzi group at the Massachusetts General Hospital & Harvard Medical School; it was published in Nature Neuroscience and highlighted in a News Medical Lifesciences article.

The work describes a flexible microfluidic system that allows the effectiveness of different AD treatment drug compounds to be evaluated based on the responses of different types of brain cells, particularly the microglia cells involved in the brain & CNS immune response.  The microfluidic device contains 2 chambers, one with neuron cells and astrocytes, and the other with microglia.  Reduced microglia ‘recruitment’ by the other cells, as visualised in the channels between the chambers, via excreted cytokines by the neuron and astrocyte cells indicates reduced neuronal damage and, presumably, reduced loss of memory or other mental capacity.

There are over 3000 approved drug compounds for treatment of Alzheimer’s disease, according to Joseph Park, first author on the research paper, so evaluation in a representative biological and disease environment such as created in their chip is crucial.  Human trials are lengthy, expensive and carry risk for the participants.  Selection of the most promising candidates for drug therapy through understanding their interaction with brain cells in this microfluidic system offers promise to accelerate the pace towards effective treatments of Alzheimer’s disease.

Research from Zhonghua Ni‘s group at Southeast University in China published recently in Analytical Chemistry (DOI: 10.1021/acs.analchem.8b02201) allows for effective cell concentration along a spiral microfluidic channel.

The device is connected directly to a syringe, and outward flow in the microfluidic spiral channel is generated simply by the operator pushing on the plunger.  Separation by inertial (centrifugal) forces partitions the cells to the inside channel ‘track’ such that a Y-shaped bifurcation at the outer spiral terminus yields a concentrated cell suspension on the inner outlet and almost cell-free solution on the outer outlet.

The device embodies simplicity as a virtue.  It has no active components, and requires no electrical power, only average human thumb power :-); it simply requires a sample-filled syringe and two collection vials.  It is well conceived for both resource-poor and lab environments alike.  Their bold claim that one could bring their concentrator “into commercial outcomes without additional redesigning” initially made me chuckle, but after absorbing their work, I have to admit they appear to be exaggerating only slightly!  An Abbott, BD or start-up would require some industrial product design, but technically, they appear to be at the finish line.  I can’t speak to the competition in this product segment, and sure hope they have a good patent, but it seems pretty clever!

An interesting review of lab-on-a-chip devices made on flexible substrates was recently published in Lab on a Chip by Anastasios Economou and Christos Kokkinos at the University of Athens, and Mamas Prodromidis at the University of Ioannina.  The authors cover devices made on polymer, paper and textile fabric substrates and which use electrochemical biosensing for detection.  Lab-on-a-chip & microfluidic devices based on flexible substrates are very attractive given the low costs of the substrate and deposited materials, inexpensive and well established high volume manufacturing processes, and physical flexibility as an attribute for body-wearable devices.

Sensor/device formats include patches, bandages and wristbands for dermal/transdermal sensing, dental mouthgards for salivary analyses, tooth patches for bacterial detection and contact lenses for tear fluid analysis.  Analytical applications include immunosensing, glucose monitoring, bovine serum albumin (BSA) sensing, metabolite sensing, cancer marker detection and HIV detection.  Deposition methods include lower-cost approaches such as screen printing, inkjet printing and electrodeposition for materials such as graphite, indium tin oxide, polyethylene terephthalate (PET).  Polymer substrates include thermoplastics like polyethylene, polystyrene, PET, polypropylene, polycarbonate, cyclic olefin (co)polymers (COC & COP), thermoset polymers like polyimide, and photoresists.  Filtre paper is the most common type of paper substrate, though newsprint is also mentioned.  3-D, folding or origami approaches are also discussed, as previously highlighted in this blog.

Fascinating new research by Felix Ellett and co-authors from Massachusetts General Hospital (MGH) and Harvard Medical School (HMS) shows elegant detection of sepsis in hospital patients via microfluidic blood analysis.  The study was published recently in Nature – Biomedical Engineering and profiled by MGH and Technology Networks.

Previous research from Daniel Irimia‘s group at MGH and HMS in 2014 showed that dysfunctional behaviour of neutrophils (a type of white blood cell) correlated well with sepsis.  This current research from his and Jarone Lee‘s (MGH) groups takes the next step and uses a microfluidic device to perform the analysis.  Their 5-mm device has 4.5-µm channel filtres (to screen out red blood cells) and meandering channel networks to allow the neutrophils to leave the sample drop of blood, and be observed in the microfluidic channel network.  Discrimination by neutrophil characteristics & behaviours (such as number of neutrophils, oscillations and motionless time in microfluidic channels, migration distance and tendency to migrate back to the central sample chamber) allowed the researchers to correctly predict the presence of sepsis in over 95% of patients in a small study of 19 ICU patients.

Detection of sepsis in patients is of critical importance.  According to the MGH and TN article profiles, about one million cases of sepsis occur annually in the US, with 25% of cases being fatal.  Comparable data from the UK showed that the mortality rate for hospital admissions with severe sepsis assessed withing the first 24 hours was 45%.  Current detection methods leave much to be desired, with misdiagnosis occurring in 30% of patients.  The arrival of fast, accurate diagnostic methods would lead to better patient outcomes, likely including saved lives.

Ground-breaking lab-on-a-chip research out of the van Nimwegen lab at the University of Basel, Switzerland and the Myers lab at the Max Planck Institute of Molecular Cell Biology and Genetics, Germany, has enabled both rapid environmental control of cell media, and the ability simultaneously monitor large numbers of bacteria as single cells.  The work, recently published in Nature Communications and highlighted in Technology Networks, describes a PDMS microfluidic device with ~1900 channels measuring ~ 1 x 1 x 25 µm that house the single cells, and which are connected perpendicularly to larger supply channels containing the cell medium solution.  The device also features continuously variable 2-input control of cell media via a split channel configuration, together with a mixing serpentine upstream of the smaller single cell channels.  New image analysis software was also created to handle the enormous amount of image data generated by this system when in operation.

The device is intended to provide a real-time monitoring window into gene regulation in bacteria, a laudable advance.  For example, bacterial response to different antibiotics can be measured, and insight into intra-cellular communication may be revealed.  The device’s performance was nicely illustrated in a video (image above from Nature Communications article) showing an experiment where gene regulation via the lactose operon was observed while alternating between glucose and lactose nutrients in the supply channel; green fluorescent protein can be seen to appear within 1-2 h of introducing the lactose-rich medium to the cells.

Fascinating research out of the Robinson, Kemere and Pasquali labs in Electrical & Computer Engineering and Chemistry Departments at Rice University, at Baylor College of Medicine and out of the Royer-Carfagni lab in Engineering at the University of Parma.  It was published recently in ACS NanoLetters and summarised in NewsMedical, and is focussed on the gentle introduction of microelectrodes for medical applications using microfluidics.  The team showed the insertion of carbon nanotube (CNT) electrodes into brain tissue surrogate (agarose gel) as well as rat brains, with subsequent observation of neural activity via these electrodes.  The ~30 µm CNT fibre is placed along the central channel in a glass/PDMS microfluidic device, and then carefully controlled sheath fluid flow is used to pull the wire down through a narrower ~100 µm-wide channel and into the tissue.  See impressive video showing microfluidic injection of CNT electrode into gel.

This technology is an excellent example of marrying microfluidics with nanotechnology, in this case for a medical application.  The microfluidic insertion of CNT electrodes in lieu of traditional approaches makes the process less damaging to surrounding brain tissue during insertion.  Neuro-sensing applications where gently introduced flexible CNT electrodes would be beneficial may include monitoring for epilepsy, repairing/patching connections for sight and hearing, as well as other areas of neuroscience research.

Researchers from the Mace lab at Tufts University recently published an article in Scientific Reports highlighting their new open-source CAD software that is tailored for the design of paper microfluidic devices.  The software works across all OS platforms, appears fairly straightforward and user-friendly, and likely capable of covering most of the normal aspects of geometrical design for channel networks, based on examples presented.

The software is provided by the authors gratis in a laudable effort to reduce the costs associated with chip design, especially for resource-constrained environments such as schools and labs in developing countries.  The price for subscription or purchase of industry-standard CAD software can be prohibitive, easily reaching tens of thousands of dollars, and scaling with the number of users, etc.  Removing the cost of design SW breaks down a significant cost barrier to innovation for paper microfluidics, a technology that otherwise boasts very economical costs for consumable materials (paper) and fabrication (wax printing with commercially available printers).