Research highlights: manipulating cells inside and out

Abstract

We highlight recent work manipulating cells: from whole cells, to intracellular content, and even subcellular gradients in proteins. In the first manuscript, using interdigitated electrode arrays at a controlled tilt angle to a microchannel allows for an array of acoustic nodes that apply force and isolate larger circulating tumor cells from remaining cells in RBC-lysed blood. Moving to the subcellular scale, recent work shows the ability to use rapid bubble generation induced by a pulsed laser to transfect hundreds of thousands of cells in parallel, especially with larger cargo, such as live bacteria. Manipulating at an even finer level, our third highlighted paper applies magnetic nanoparticle-based techniques to the localization of proteins within the cytoplasm in gradient configurations. A recurring theme in the literature is how interfacing at the cellular scale is a key feature enabled by micro & nanotechnology. This feature can be exploited to achieve new capabilities for cell biologists which opens up new fundamental cell biology questions. This matching of scales and the unique advantages are well demonstrated in the articles highlighted.

Acoustic separation of circulating tumor cells

Circulating tumor cells (CTCs) can serve as a non-invasive means to evaluate tumor status for monitoring treatments or identifying targeted therapies. Various CTC isolation techniques have been developed, including immunomagnetic separation, dielectrophoresis, and size-based sorting. Some of the challenges in CTC isolation include high-throughput and high purity sorting. Huang et al.¹ address these challenges with an acoustic based technology to isolate CTCs based on physical properties of size, deformability and density. They use tilted-angle standing surface acoustic waves (taSSAWs) to generate standing waves in a microfluidic channel that guide particle trajectories based on these physical properties.

The authors develop a microfluidic device for continuous acoustic sorting. Fig. 1a shows the microfluidic device, it has two outlets, one for waste and one for CTC collection. Acoustic standing waves are generated in the microchannel using interdigitated piezoelectric substrates placed outside the device. The standing waves create pressure nodes at an angle to the flow direction (Fig. 1b). Particles experience an acoustic radiation force that is proportional to physical properties. The angled orientation of the pressure nodes cause the particles to pass through the nodes repeatedly, thus amplifying the trajectory deflection and resulting in higher particle separation distances. The acoustic force depends on the acoustic pressure, volume of the particle, wavelength, wave number of the acoustic waves, as well as acoustic contrast factor (which is dependent on the compressibility and density of the particle and medium).

Fig. 1 (A) Sheath flows are used to focus particles in the mainstream. Tilted interdigitated transducers (IDTs) generate acoustic waves in the microchannel. (B) The acoustic waves create pressure nodes at an angle to the main flow stream. A cell translating downstream will transit across a number of nodes that is dependent on the tilt angle. (C) A view of the whole PDMS based microfluidic device situated on the IDT substrate. A U.S. penny is used to show relative size of the device. Reprinted with permission.

In order to optimize the various parameters of the system, Huang et al. created a computational model that allows them to tune each parameter to identify optimal conditions for the separation of cancer cells. The separation distance between particle streams (ΔY) is the key parameter that was the focus of the simulations. An increase in separation distance of the target cell trajectories from the waste cell trajectories implies an increase in enrichment of target cells. Simulation results in Fig. 2a show that higher flow rates require smaller tilt angles to achieve larger separation distances. At high flow rates the acoustic force does not have sufficient time to act, thus interaction with more pressure nodes as it traverses downstream are needed to deflect the particles; this is achieved by changing the tilt of the IDTs. Similar simulations optimized the tilt angle as a function of input power (Fig. 2).

Fig. 2 (A) & (B) Simulation results show that for increasing flow rate and decreasing input power, the optimal tilt angle of the IDTs decreases. Reprinted with permission.

Using the flow rate of 1.2 ml h⁻¹ and 37.5 dBm power, they were able to achieve 87% recovery of MCF-7 cells and HeLa cells from a background of approximately 1 million WBCs. Further higher input power would potentially increase separation of particles, however the cells would experience detrimental levels of joule heating at higher powers. This study found that cells collected using 37.5 dBm were still viable and proliferated at normal rates.

As one of the first examples of using acoustic separation in isolation of CTCs from clinical samples, the taSSAW system achieved isolation from breast cancer patient blood. It was able to isolate 59 and 8 CTCs from only 2 ml of blood from two different patients. The CTCs were identified based on DAPI+, Cytokeratin, CK+, CD45− criteria.

This initial promising data indicates future effort is warranted to increase the throughput beyond 1.2 mL h⁻¹, further develop the technology to operate on whole blood, and reduce the background levels of WBCs. Solving these issues would further emphasize the strengths of a label-free physical property-dependent separation.

BLAST-ing large cargo into cells at high-throughput

Delivery of large elements into the cell cytoplasm or organelles is the necessary first step to several important biological techniques including physical intracellular manipulation using functionalized nanoparticles, transfection of DNA, the study of intracellular bacterial pathogenesis, imaging of internal cellular components, and others.^2–4 Although viral vectors and chemical delivery vehicles have been used with good efficiency, these methods are not applicable to arbitrary molecular cargo, may be limited in cargo size and risk endosomal trapping. Meanwhile, existing physical methods of creating pores through which cargo can be delivered may be damaging to cells and must make trade-offs between cargo size and throughput. Wu et al. developed a vastly improved physical delivery method called the biophotonic laser-assisted surgery tool (BLAST) that no longer needs to make this trade-off as it can deliver cargos up to several micrometers in size at rates up to 100 000 cells per minute while maintaining high cell viability.⁵

BLAST achieves high-efficiency delivery of large molecules into cells by first creating transient pores at precise locations in the membranes of adhered cells and then immediately injecting cargo-containing solutions towards and into these pores, all on a chip (detailed process is shown in Fig. 3). To do this, cells are adhered to and cultured on a porous SiO₂ layer containing uniformly spaced trans-membrane holes, the sidewalls of which are asymmetrically coated with titanium, which is thermally conductive. A rapid pulse scanning laser is used to store energy in this titanium coating triggering cavitation bubbles to rapidly form in these holes and locally disrupt the membranes of adjacent cells. From below, solution containing the wanted cargo is injected through these same trans-membrane holes and toward the transiently permeabilized cells. In the current implementation the area of the SiO₂ layer was 1 cm² allowing cargo to be delivered to 100 000 cells at once. Importantly, BLAST co-localizes the sites of pore-formation with the channels through which solution containing the cargo is injected ensuring direct delivery. Since the permeabilization and injection steps occur nearly simultaneously, large cargo can be delivered before the transient pores close – a process that begins immediately after they are created and severely limits the cargo sizes that other physical poration approaches (which often rely on diffusion rather than injection) can achieve.

Fig. 3 (A) Schematic drawing of the BLAST system and its components. Bubbles formed by the rapid heating of titanium dioxide coated pores generate holes in cell membranes of overlying cells. Delivery of molecules and small objects is achieved through a pressure applied from the bottom microfluidic channel. (B) Close up of the titanium dioxide coated silicon pores which are heated by the laser pulse. (C) Bubble formation occurs rapidly upon exposure of the pores to the pulsed laser. (D) Confocal images showing pore formation within a cell. Reprinted with permission from the Nature Publishing Group.

To demonstrate the efficiency and applicability of their technique the authors injected various cargos such as enzymes, antibodies, functional nanoparticles, polystyrene beads (up to 2 microns in diameter), high molecular weight dextrans and live bacteria into primary human fibroblasts, epithelial cells, monocyte-derived macrophages, and HeLa cells. They first showed that using optimized parameters, 3 day cell viability was >90% for all four cell types, while delivery efficiency (tested using 40 kDa dextran) was also >90% for three of the cell types and at least 50% for the human macrophages. Next, they demonstrated that the cargo remained functional post-delivery by assaying the activity of an injected enzyme (β-lactamase) and showing that it cleaved its substrate as normal. Finally, they demonstrated the ability to deliver live bacterial cells for the purpose of studying their pathogenesis within the cell, and unexpectedly discovered a new pathogenic role of a gene previously thought to be implicated only in phagosome escape. This technique is an innovative solution for delivering large cargo into cells that will enable the level of statistical significance needed for quantitative comparisons due to its high-throughput, and may allow multiplexed studies since multiple cargos can be injected into the same cells all at once.

Magnetic control of intracellular protein distribution

Moving beyond manipulating cells directly, or delivery of cargo into cells, in their paper, Etoc et al. describe the use of magnetic nanoparticles (MNPs) with hydrodynamic diameters below 50 nm and magnetic tips to realize control of the distribution of intracellular proteins within a cell.⁶ This work improved upon previous work using MNPs of 500 nm and 100 nm diameters, which because of their larger size suffered from poorer temporal and spatial resolution. With their new technique, the group claimed a temporal response in which particles migrated within a few tens of seconds (Fig. 4b).

Fig. 4 Control of intracellular protein distribution with magnetic forces. (A) 50 nm MNPs showed Brownian motion while 75 nm MNPs showed diffusion behavior in the viscous regime. (B) Magnetic forces influence intracellular MNPs distribution. (C) Top: magnetic forces localized bi-functional MNPs as shown in ratiometric images. Bottom: intensity profiles of ratiometric image. Reprinted with permission from Etoc et al. Copyright (2015) American Chemical Society.

The group first characterized the diffusion of MNPs in cytoplasm. It was found that 50 nm MNPs can be suitably manipulated magnetically as their diffusion behavior is suggestive of Brownian motion while the 75 nm MNPs displayed confined motion, indicative of a viscous diffusion regime and interaction with the cytosolic network of proteins (Fig. 4a).

Next, the utility of this technique was demonstrated in live HeLa cells. To overcome electrostatic repulsions from the lipid membrane and to induce binding to intracellular protein, the authors chemically modified the MNPs with carboxyl groups and HaloTag (HT-GFP). Upon application of the magnetic tip with femtonewton-range forces, a gradient of functionalized MNPs was established (Fig. 4c). The authors concluded that their technique could be valuable to manipulate activity of intracellular proteins due to its short response time, localized control and without any extensive genetic modification.

Nonetheless there are some questions that these studies do not sufficiently address yet: (1) besides the particle size, surface chemistry may affect the behavior and “stickiness” of particles within the intracellular milieu, modifying the diffusive constants independent of size; (2) the authors raised the possibility of using their platform for parallel cellular manipulation, perhaps using uniformly-created magnetic tips in arrays. Field-gradient enhancing magnetic tip arrays have been previously investigated⁴ and applied to applications such as to control intracellular distribution of TAU protein in neurons,⁷ and can be adopted for approaches described here.

References

1. P. Li, Z. Mao, Z. Peng, L. Zhou, Y. Chen, P. H. Huang, C. I. Truica, J. J. Drabick, W. S. El-Deiry, M. Dao, S. Suresh and T. J. Huang, Acoustic separation of circulating tumor cells, Proc. Natl. Acad. Sci. U. S. A., 2015, 112(16), 4970–4975.
2. J. Dobson, Gene therapy progress and prospects: magnetic nanoparticle-based gene delivery, Gene Ther., 2006, 13, 283–287.
3. A. K. Chen, W. J. Rhee, G. Bao and A. Tsourkas, Delivery of molecular beacons for live-cell imaging and analysis of RNA, Methods Mol. Biol., 2011, 714, 159–174.
4. P. Tseng, J. W. Judy and D. Di Carlo, Magnetic nanoparticle-mediated massively parallel mechanical modulation of single-cell behavior, Nat. Methods, 2012, 9, 1113–1119.
5. Y.-C. Wu, et al. Massively parallel delivery of large cargo into mammalian cells with light pulses, Nat. Methods, 2015, 12, 439–444.
6. F. Etoc, C. Vicario, D. Lisse, J.-M. Siaugue, J. Piehler, M. Coppey and M. Dahan, Nano Lett., 2015, 15(5), 3487–3494.
7. A. Kunze, P. Tseng, C. Godzich, C. Murray, A. Caputo, F. E. Schweizer and D. Di Carlo, ACS Nano, 2015, 9(4), 3664–3676.

---