Research highlights

Driving liquids on-chip using SAWs

The field of microfluidics has experienced great advances in the past decade, in particular in the context of reagent mixing, cell culture, and sample storage. However, for transport of liquids on-chip, most microfluidic devices still rely heavily on external equipment, such as pumps, pressure sources, or robotic liquid handling systems.^1,2 This limits the portability and ultimately the usefulness of microfluidic chips in point-of-care settings. It also presents a barrier to industrial, large-scale production of microfluidic devices.

To address this issue, Cecchini and colleagues have recently applied surface acoustic waves (SAWs) to route fluids on-chip. Previously, SAWs have been used for particle and droplet sorting,³ but the approach taken by Travagliati et al.⁴ allows for more complex, feedback-free fluid manipulation and sensing. Fluid actuation was based on the resonant coupling between the stationary modes of an on-chip cavity and the travelling waves emanating from a SAW source. When the SAW frequency corresponds to a cavity eigenmode f_c, and the cavity is empty, the SAWs are fully transmitted; otherwise, the waves are reflected (a schematic is shown in Fig. 1). In addition, placing an object (e.g. a drop of a liquid) inside the cavity alters its f_c, causing the incoming SAW to be absorbed or slowed down. This characteristic was exploited to utilize the system as a sensor, such that a change in the wave velocity could be detected on-chip as a shift in the transmission spectrum, thus indicating the presence of liquid inside the cavity. In addition, a strongly absorbing liquid inside the cavity could affect the amount of transmitted and reflected energy. The SAWs could be used to detect the presence of the absorbing liquid without interacting with it strongly and suffering only minor energy losses.

Fig. 1 Schematic of liquid manipulation using surface acoustic waves, resonating at the characteristic frequency, f_c. of an on-chip cavity. Bottom figure: a typical transmission spectrum of the chip. Figure adapted and reprinted with permission from the Royal Society of Chemistry from Travagliati et al.⁴

The application of the system as an actuator was facilitated by the momentum transfer between an incident SAW and a liquid droplet placed on-chip, causing the droplet to move. When it reached the cavity, the incident SAWs were reflected. Since no acoustic field could be generated inside the cavity, the droplet remained trapped in it. This was demonstrated on the surface of a lithium niobate wafer. The electrodes were 3 μm wide with a 3 μm separation, and the cavity was formed by two grated mirrors (6.5 μm grooves with 5.5 μm separation). When the SAW source was activated, only waves with f = 163.33 MHz were transmitted through the cavity. For droplet actuation, the surface of the wafer was rendered hydrophobic (to avoid wetting) and a feeding signal of 22.5 dBm had to be applied. When a nl-sized water droplet moved (at a few mm s⁻¹) and got trapped inside the cavity, a red-shift in the transmitted frequency was observed, indicating a 10 °C rise in temperature of the cavity.

The application of resonant SAWs to droplet actuation, as described by Travagliati and colleagues, is advantageous for particle separation, fluid pumping and routing, and simultaneous sample detection. Moreover, the selective actuation and trapping of liquid samples could be extended to generate liquid logic gates. The system of electrodes can easily be scaled down to other length scales, requires little driving power (which can be provided by a battery) and can be produced in a cost-efficient manner. One disadvantage of the current setup is the large temperature change inside the droplets as well as the resulting atomization. While this side effect can be exploited in spectrometry studies, it is a disadvantage for the testing of biological samples. For this reason, a substrate with a different thermal expansion coefficient or a lower RF signal may be useful.

Cell cycle synchronization

The study of cellular events and related biological phenomena, such as molecular regulation or genetic expression, is of great importance for drug development and treatment of diseases.⁵ A large part of biological noise in such studies is due to the variation in cell cycle state. Since specific molecular processes tend to be tied to different cell states,⁶ it would be advantageous to synchronize the cellular cycle. To synchronize cells,⁷ they are attached to a membrane, such that during mitosis one half of the dividing cell remained connected to the membrane and the other half (in G1 phase) was eluted and could be collected for further studies. While this method did not introduce external stresses on the cells, it could only be used with adhesive cells.

Manalis and coworkers have recently adapted this approach to develop a microfluidics-based cell synchronization device. In this approach, cells are immobilized in miniaturized holes such that the daughter cells can be sheared off and collected. The device fabricated by Shaw et al.⁸ consisted of a glass substrate, into which the fluidic channels and the cell capture chamber were etched, and a silicon substrate containing over 2000 holes, each of which was 2 μm in diameter. After anodically bonding these two substrates, the entire system was aligned between two aluminium plates, which provided fluidic seals and had openings for fluidic connections and microscope imaging. The choice of these materials ensured that the device could be cleaned after each experiment and hence was reusable.

To demonstrate the utility of this device for synchronizing cells, an asynchronous suspension of mouse lymphocytic leukemia cells was introduced into the fluidic channel at 3.4 kPa, while the channel exit was kept at atmospheric pressure. The suspension filled the cell capture chamber and the solvent flowed through the membrane. The suspended cells were captured on top of the membrane holes due to the difference in pressure above and below the membrane (vertical pressure differential). Up to 80% of all holes were occupied with cells. When these cells reached mitosis, the daughter cells were easily removed by the applied perfusion flow. The shear stress acting on the cells during the continuous perfusion was much lower than the physiological shear stress.

In the validation experiments the cells were eluted every 12 h for 3 days in 40 μl batches, then fixed, stained and counted. To ensure that the eluted cell suspension was not contaminated with cells from the original sample, the elution was started 1 h after cell loading. This allowed sufficient time for the captured cells to settle inside the chambers. The purity of the eluent was shown to be affected by the collection time, such that –as expected– a short collection period resulted in a small number of eluted cells with a small age variation. Microscopy analysis confirmed that the mother cells were attached to the membrane and divided, and that the daughter cells were washed away. The measured interdivision times were 10–11 h, which was comparable to observations in conventional cell culture. At each division roughly 1000 daughter cells were collected. In the asynchronous cell population, close to 50% of cells were in the G1 state. This fraction rose to 83% in the synchronized cell suspension, a statistically significant increase, since a large number of daughter cells (∼1000) was collected at each elution.

The synchronization device described here utilizes a simple physical characteristic, namely the pressure drop across a membrane, to capture a large number of cells and to generate a highly synchronous G1 cell population. In contrast, previous membrane-based synchronization approaches required the complex tuning of the membrane chemistry to trap cells, and were not compatible with all cell types. The present platform has minor shortcomings, such as clogging of holes in the membrane or trapping of air bubbles that alter the pressure differentials. Another potential limitation of the microfluidic membrane-elution technique is that it only produces cells in the G1 state and not in G2, S and other phases of the cell cycle.

Aging of yeast

Aging is an essential component of the life cycle in all organisms including eukaryotes. The ability to continuously monitor the changes in single cells may provide an insight into the mechanisms involved in cellular aging processes.⁹Saccharomyces cerevisiae is considered one of the simplest models for studying eukaryotic aging.¹⁰ For example, the identification of senescence factors may be possible by monitoring the life cycle of mitotic cells.¹¹ In the past, budding yeast was used as a model to study aging phenomena due to their short life cycle and simple handling conditions.¹⁰

The continuous observation of the whole lifespan of yeast via high resolution microscopy is challenging.¹² The current approaches are not time-efficient and require labor intensive handling. Tracking the budding of an individual yeast cell over many cycles of division is technically difficult because the number of daughter cells increases exponentially. One method to overcome these limitations is to remove the daughter cells as the mother cells divide. However, the separation of the daughter cells from mother cells by manual dissection is cumbersome and labor intensive.¹³ There is no high resolution remedy for this challenge using traditional imaging techniques.

Conventional dissection approaches present a number of limitations, such as inefficient waste and nutrient transportation during the experiment or low optical resolution. To address this problem, Lee et al. used a microfluidic system to monitor the entire replicative aging process for budding yeast S. cerevisiae.¹² Their experimental system utilized a continuous high resolution imaging set-up to track budding yeast over their replicative life cycle. This platform consists of microarrays of polydimethylsiloxane (PDMS) pads and a focal plane to isolate daughter yeast cells from mother cells (Fig. 2).

Fig. 2 Schematic of the microfluidic dissection process. PDMS micropads are pressed onto a microscope cover slide and this set-up is used to entrap the mother cells. By flowing medium over the mother cells, daughter cells are continuously removed. Figure adapted and reprinted with permission from Lee et al.¹²

The device relies on the size difference between the mother cell and the buds to carry out the dissection process. In the beginning, the mother cell is entrapped within the PDMS micropad. Then, flow is established inside the microfluidic cell which continuously removes the buds. This method allows for the observation of the aging process for one single yeast cell throughout its life cycle.

The fully automated microfluidic dissection device enabled monitoring of phenotypical changes, such as the morphologies of cells and vacuoles during aging and cell death for budding yeast in a single experiment.

Although this approach was shown to be highly robust, one limitation of this system is that daughter cells may on occasion push the mother yeast cell out of the micropad. This platform is expected to find use in investigating the effects of cell-to-cell heterogeneity on cellular aging, for exploring phenotypic differences at different stages of life, and for studying the molecular mechanisms relevant to the aging process.

References

1. Lab Chip, 2006, 6, 1118–1121.
2. T. Ward, M. Faivre, M. Abkarian and H. A. Stone, Electrophoresis, 2005, 26, 3716–3724.
3. T. Franke, A. R. Abate, D. A. Weitz and A. Wixforth, Lab Chip, 2009, 9, 2625–2627.
4. M. Travagliati, G. De Simoni, C. M. Lazzarini, V. Piazza, F. Beltram and M. Cecchini, Lab Chip, 2012.
5. K. Vermeulen, D. R. Van Bockstaele and Z. N. Berneman, Cell Proliferation, 2003, 36, 131–149.
6. K. W. Kohn, Mol Biol Cell, 1999, 10, 2703–2734.
7. C. E. Helmstetter and D. J. Cummings, Proc. Natl. Acad. Sci. U. S. A., 1963, 50, 767–774.
8. J. Shaw, K. Payer, S. Son, W. H. Grover and S. R. Manalis, Lab Chip, 12, 2012.
9. A. E. Roux, P. Chartrand, G. Ferbeyre and L. A. Rokeach, J. Gerontol., Ser. A, 2010, 65, 1–8.
10. M. Kaeberlein, Nature, 2010, 464, 513–519.
11. C. D. Powell, S. M. Van Zandycke, D. E. Quain and K. A. Smart, Microbiology-Sgm, 2000, 146, 1023–1034.
12. S. S. Lee, I. A. Vizcarra, D. Huberts, L. P. Lee and M. Heinemann, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 4916–4920.
13. K. Steffen, B. Kennedy and M. Kaeberlein, J Vis Exp, 2009, 25, 1209.

---