Research highlights: microtechnologies for engineering the cellular environment

Abstract

In this issue we highlight recent microtechnology-enabled approaches to control the physical and biomolecular environment around cells: (1) developing micropatterned surfaces to quantify cell affinity choices between two adhesive patterns, (2) controlling topographical cues to align cells and improve reprogramming to a pluripotent state, and (3) controlling gradients of biomolecules to maintain pluripotency in embryonic stem cells. Quantitative readouts of cell-surface affinity in environments with several cues should open up avenues in tissue engineering where self-assembly of complex multi-cellular structures is possible by precisely engineering relative adhesive cues in three dimensional constructs. Methods of simple and local epigenetic modification of chromatin structure with microtopography and biomolecular gradients should also be of use in regenerative medicine, as well as in high-throughput quantitative analysis of external signals that impact and can be used to control cells. Overall, approaches to engineer the cellular environment will continue to be an area of further growth in the microfluidic and lab on a chip community, as the scale of the technologies seamlessly matches that of biological systems. However, because of regulations and other complexities with tissue engineered therapies, these micro-engineering approaches will likely first impact organ-on-a-chip technologies that are poised to improve drug discovery pipelines.

“Sticky” choices between two micropatterned surfaces

Cell behavior, such as adhesion and motility, is largely governed by interactions between cell surface proteins and proteins of the extracellular matrix (ECM). Integrins, for example, are transmembrane proteins that bind to the ECM and serve as mechanotransducers, converting binding and subsequent mechanical forces into chemical signals for the cell to initiate a cascade of responses. Precise patterning of ECM components on a substrate has become a popular approach for spatial control to gain better intuition concerning cellular contact-mediated responses. However, across the literature there appear to be inconsistencies in jargon and a lack of control and definition of the surfaces adjacent to the protein patterning, making it difficult to compare results and to develop a more complete picture of cell–ECM interactions.

To clarify these discrepancies in cell–surface interaction assays, Ricoult et al.¹ promulgated the use of the term “cell–surface affinity” as opposed to “cell adhesion”, which does not always imply cell preference.² Furthermore, they defined the “reference surface” (RS) to be the surface contiguous to the patterned regions of interest, and described a simple approach of controlling the cell–surface affinity of the RS by adjusting the ratio of its constituents, poly-D-lysine (PDL) and polyethylene glycol (PEG), which serve as high- and low-affinity polymers, respectively. The authors used a micro-contact printing approach to create stripes of adhesive surfaces next to reference surfaces.

With these concepts in mind, the cell–surface affinity to a patterned protein was adjusted by only varying the RS composition, as demonstrated with Rat2 fibroblasts in Fig. 1. A low affinity RS facilitated preferential binding to fibronectin, while a high-affinity RS made both surfaces indistinguishable, i.e. gave a random distribution of cells (~18% of cells on the stripes). Furthermore, C2C12 myoblasts subjected to the same conditions displayed similar trends to the Rat2 fibroblasts, but with an inflection point in adhesion to fibronectin stripes adjacent to a stickier reference surface. Importantly, one can now obtain a quantitative value similar to an “IC50” used to evaluate drug effectiveness, but for relative cell–surface affinity.

Fig. 1 Cell–surface affinity response of Rat2 fibroblasts (stained red and blue for F-actin and nuclei, respectively), to stripes of fibronectin (green) when presented with (a) a low affinity and (b) a high affinity RS. Scale bar is 100 μm. Adapted from Ricoult et al.¹

Beyond cell lines, to study the response of primary polarized cells under these conditions, neurons from embryonic rat spinal cords were tested in the presence of patterned netrin-1, and were quantified by the locations of both their somas and their axons. When subjected to the lowest affinity RS, ~98% of both neural somas and axons were found on the netrin-1 stripes, as opposed to ~18% of the soma (no preference) and ~40% of the axons in the highest affinity case. Similar trends, yet mitigated preferences toward patterned IgG were observed. A 75 : 25 %PEG : %PDL RS yielded the highest number of cells with only their axons on the pattern, which is the condition that best indicates a guidance response to substrate-bound cues. With this RS, the guidance response was confirmed by a significant decrease in the number of neural cells that adhered to the patterned stripes after integrin deactivation via treatment with src family tyrosine kinase inhibitor PP2. Additionally, there was no difference in cell adhesion to IgG stripes under the same conditions, further assuring cell preference to netrin-1 via an integrin response.

The authors proposed that the concept of a reference surface could be fine-tuned further with more complex combinations of other non-specific polymers, biologically active proteins, and ECM components, and can be rigorously characterized by generating cell–surface affinity response curves as quantitative measures of the response magnitude and slope. This quantitative information on relative affinity to extracellular matrix proteins should be especially useful towards applications in tissue engineering and developing physiologically-relevant organ system models (organ-on-a-chip).³

Physically reprogramming cell epigenetics

Induced pluripotent stem cells, from which any cell type of the body can be derived, have the potential to revolutionize basic research, disease modeling, and regenerative medicine, amongst other applications. It is widely known that biochemical factors, such as transcription factors, microRNAs, and various chemical inhibitors and compounds, can reprogram cells into pluripotent cell lines. In general, these approaches attempt to modify the expression of various genes within the cells, for example pluripotency markers (OCT4, SOX2, etc.) and proliferation markers (MYC, KLF4, etc.), which are induced via cocktails or viral transfection.⁴ Additionally, the epigenetics of cells (or alterations to cell transcription via DNA methylation or histone modification) are a targeted mode of reprogramming. For example, histone deacetylase and methyltransferase inhibitors have been shown to affect the efficiency of cell reprogramming.

As of yet, the role of biophysical cues such as substrate topography, substrate adhesiveness, and cell shape in cell reprogramming is unknown, despite having a known role in a myriad of cell functions. To shed light on the importance of biophysical factors, Li and colleagues used microfabricated grooves and nanofibers to control topography, or micropatterned adhesive surfaces to control cell shape, as means for cell reprogramming and epigenetic modification of cells⁵ (Fig. 2). Substrates were microfabricated using soft lithography procedures. Various topographies and geometries were molded on PDMS membranes, or patterned flat surfaces were produced using a stencil plasma process to define reactive regions. These consisted of grooves of 10 μm width, 10, 20, and 40 μm spacing, and 3 μm height. Finally, to facilitate cell adhesion, these substrates were coated with gelatin. The nanotopographies were studied via the generation of aligned nanofibers on PDMS.

Fig. 2 a) SEM of microfabricated grooves. b) Various geometries of studied microtopographies. c) Protocol for the reprogramming and replating of adult cells. d) Cell morphology when cultured on flat versus microtextured substrates. e) and f) The effect of topography on reprogramming efficiency, assessed via Nanog+ colonies. Adapted from Li et al.⁵

Li and colleagues first studied how these substrates influenced the process of cell reprogramming. Mouse fibroblasts, transduced to be pluripotent via a lentiviral approach, were seeded onto their substrates and tracked over time. They found that the substrates had a significant effect on cell alignment, similar to previous work with other cell types,⁶ and experienced a corresponding elongation of nucleus size, presumably induced directly by the substrates. Additionally, they observed that 10 μm width grooves and spacing increased the efficiency of cell reprogramming by almost four fold. This was assessed via Nanog+ fluorescent staining, and the presence of pluripotency markers OCT4, Sox2, Nanog, and SSEA-1.

The source of this improved efficiency was found to be mediated through chromatin modification (changes in epigenetic marks) of the cells with elongated whole cell and nuclear morphologies. Western blotting and immunostaining confirmed this directly, through modification markers AcH3, H3K4me2, and H3K4me3. This biophysical approach was further contrasted by comparing its effect to known small-molecule epigenetic modifiers; both of these epigenetic approaches had similar effects on reprogramming efficiency.

The mechanical origin of these changes appeared to be linked to cytoskeletal tension induced by the microgrooves or shape changes. Critically, it was found that blebbistatin, an inhibitor of cell contractility, removed any epigenetic modifications normally induced through the microtopography. Both nanotopographies and the direct adhesive micropatterning of cells induced changes in cellular epigenetics, indicating that cellular (or nuclear shape) may be the critical factor.

In summary, the authors utilized microfabricated substrates to elucidate a biophysical link in cell programming into pluripotent stem cells, which could have numerous implications in the optimization of approaches to induce pluripotency in adult cells. In particular, the dependency of these approaches on chemistries and material surface characteristics may become an interesting application that lab-on-a-chip approaches could optimize and improve.

Gradient patterning of soft surfaces using hydrodynamic flow focusing

Beyond patterning of cellular environments with defined areas or strips, there is significant interest in controlling gradients of biomolecule surface patterns for applications in tissue engineering or mimicking physiological systems. Hydrodynamic flow focusing (HFF) describes a form of fluidic flow shaping in microfluidic channels, where two outer streams squeeze a middle stream down to a thin region. Through programmable flow rate control, the position and width of the middle stream can be tuned and used to generate adsorbed patterns of biomolecules on surfaces.⁷ The advantages of hydrodynamic focusing approaches include: (1) a very simple microchannel design, offering the potential to parallelize gradient cell studies on small surface areas and (2) an integration capacity to use further flow shaping techniques.⁸

While microfluidics provides a huge variety of gradient formation tools, e.g. complex gradient channel mixers or microjet-derived gradients for diverse cell behavior studies,⁹ these tools have been mostly applied to create gradients on stiff polymers (>10 kPa) or glass. For stem cells, such hard surfaces are known to trigger a specific, but unintended progression towards a differentiated cell fate, making current gradient tools less suitable for stem cell research.

To overcome uncontrolled stiffness-induced cell fate determination, Lutolf and colleagues used HFF to generate biomolecular gradient patterns on surface-bound PEG hydrogels.¹⁰ They studied tethered arrays of leukemia inhibitory factor (LIF) gradients on the fate of embryonic stem cells (ESCs) using microfluidic based HFF on chip. LIF has been shown to act as a differentiation inhibitor for ESCs through LIF receptor signaling. The authors sought to identify a threshold gradient value for immobilized LIF to maintain ESC pluripotency locally on soft hydrogel substrates.

The culture system consists of a PEG-based NeutrAvidin or Protein A hydrogel mixture covalently linked to round silanized glass slides, providing a soft surface for ESC culture (Fig. 3A). To facilitate ESC adhesion, the PEG hydrogels were initially immersed in 0.2% (w/v) thiolated-gelatin followed by HFF LIF gradient imprinting. Prior to cell culture, the hydrogel-covered chips with tethered LIF gradients were thoroughly washed in PBS. A PMMA holder firmly sealed the microfluidic chip to the top of the PEG-hydrogel (Fig. 3B).

Fig. 3 Microfluidic-based hydrodynamic flow focusing (HFF) tethers cell signaling gradients onto PEG-hydrogels. (A1) A parallelized three phase flow microfluidic device design allows for parallel cell response studies on gradients patterned on hydrogel surfaces. (A2) PEG-NeutrAvidin-Protein A crosslinked hydrogels were covalently linked to silanized glass cover slides prior to HFF. (B1–B3) Protein gradient patterning on PEG-hydrogels using time-variant hydrodynamic flow focusing. Different numbers of flow focusing steps and biomolecular diffusion coefficients define gradient slopes, here shown for Alexa 488-BSA-Biotin/NeutrAvidin (B1), DsRED-HlgG/ProteinA (B2) and DsRED-FcLIF (B3). (C1, C2) The LIF gradient locally signals embryonic stem cells, which results in ESC colony formation. (D1, D2) Multiple gradient profiles allowed for parallelized systematic cell differentiation studies, which yielded a minimum LIF gradient requirement of 85 ng cm⁻² to sustain ESC self-renewal based on (D1) the normalized GFP-Rex1 signal and (D2) ESC colony area. Adapted from Lutolf et al.¹⁰

The microfluidic design comprises four parallel flow focusing units (1.2 mm × 0.9 × 0.1 mm) with combined inlets for each buffer unit (outer fluid streams). After priming the microfluidic channels in the chip, gradients of biomolecules were generated through a time sequence of discrete patterning steps, considering immobilization kinetics for each protein on the PEG hydrogel surface. Starting with a wide middle flow stream and subsequently narrowing down the stream width allows generation of complex shaped gradient profiles. The number of steps (changes in flow rate) permits tunability to achieve either step-wise (5 steps) or smooth gradient profiles (>20 steps).

The ESC response to tethered LIF gradients was quantified based on Rex1-GFP signals and cell colony formation (Fig. 3C). A higher Rex1-GFP signal correlated with a higher likelihood of ESC pluripotency. Cell colony formation was characterized based on colony size, compactness and morphology. Both, Rex1-GFP signals and cell colony formation showed LIF gradient dependency. The authors reported a threshold value for the PEG hydrogel-bound LIF gradient of 85 ng cm⁻² to obtain small compact cell colonies (Fig. 3D2) with high Rex1-GFP signal expression (Fig. 3D1). Finding a threshold concentration to maintain the self-renewal properties of ESCs is an important step in understanding self renewal and differentiation processes to control complex tissue differentiation.

In summary, the HFF microfluidic approach to impose patterns of biomolecular gradients on hydrogels provides two benefits for cell studies: (1) a tunable extracellular matrix depending on the required mechanical stimulation of cells by their environment, and (2) specific gradient based chemical stimulation. This gradient tool should be of use to the stem cell community to examine cell fate responses to multiple and overlapping stimuli derived from complex surface-bound gradients on tunable stiffness surfaces.

References

1. S. G. Ricoult, G. Thompson-Steckel, J. P. Correia, T. E. Kennedy and D. Juncker, Tuning cell-surface affinity to direct cell specific responses to patterned proteins, Biomaterials, 2014, 35, 727–736.
2. A. L. Calof and A. D. Lander, Relationship between neuronal migration and cell-substratum adhesion: laminin and merosin promote olfactory neuronal migration but are anti-adhesive, J. Cell Biol., 1991, 115, 779–794.
3. D. Huh, Y. Torisawa, G. A. Hamilton, H. J. Kim and D. E. Ingber, Microengineered physiological biomimicry: organs-on-chips, Lab Chip, 2012, 12, 2156–2164.
4. F. González, S. Boué and J. C. Izpisúa Belmonte, Methods for making induced pluripotent stem cells: reprogramming à la carte, Nat. Rev. Genet., 2011, 12, 231–242.
5. T. L. Downing, J. Soto, C. Morez, T. Houssin, A. Fritz and F. Yuan, et al., Biophysical regulation of epigenetic state and cell reprogramming, Nat. Mater., 2013, 12, 1154–1162.
6. A. Chen, D. K. Lieu, L. Freschauf, V. Lew, H. Sharma and J. Wang, et al., Shrink-Film Configurable Multiscale Wrinkles for Functional Alignment of Human Embryonic Stem Cells and their Cardiac Derivatives, Adv. Mater., 2011, 23, 5785–5791.
7. S. Allazetta, S. Cosson and M. P. Lutolf, Programmable microfluidic patterning of protein gradients on hydrogels, Chem. Commun., 2011, 47, 191–193.
8. H. Amini, E. Sollier, M. Masaeli, Y. Xie, B. Ganapathysubramanian and H. A. Stone, et al., Engineering fluid flow using sequenced microstructures, Nat. Commun., 2013, 4, 1826.
9. S. Kim, H. J. Kim and N. L. Jeon, Biological applications of microfluidic gradient devices, Integr. Biol. Quant. Biosci. Nano Macro., 2010, 2, 584–603.
10. S. Cosson, S. Allazetta and M. P. Lutolf, Patterning of cell-instructive hydrogels by hydrodynamic flow focusing, Lab Chip, 2013, 13, 2099–2105.

---