Repositorio Institucional de Documentos

Trabajo Fin de Máster Development of microfluidic devices with 3D collagen gels for Traction Force Microscopy Autora Júlia López Guimet Directores José Manuel García Aznar Gracia M. Mendoza Cantos Escuela de Ingeniería y Arquitectura 2013 1 Development of microfluidic devices with 3D collagen gels for Traction Force Microscopy RESUMEN En las últimas décadas, se ha demostrado sobradamente que las células ejercen fuerzas de manera constante sobre las células y matriz que las rodea. Este comportamiento es fundamental en procesos fisiológicos y patológicos, como son la embriogénesis y la metástasis; por consiguiente, el estudio de cómo las células interactúan con su entorno es de vital importancia para entender estos procesos. Con el objetivo de contribuir en este campo de investigación, el presente proyecto de máster pretende adaptar un dispositivo de microfluidica dado a la técnica de traction force microscopy , que actualmente es el método más fiable para medir fuerzas celulares. Resulta de gran interés la cuantificación de las fuerzas ejercidas por las células sembradas dentro del dispositivo de microfluidica ya que este dispositivo en concreto permite el depósito de colágeno de manera tridimensional, lo cual se asemeja más al ambiente celular fisiológico que los cultivos celulares tradicionales en 2D. La adaptación del dispositivo consistió en la introducción de microesferas fluorescentes en el colágeno, y la posterior toma de imágenes de las células mediante microscopía confocal. En ambas etapas fue necesaria la optimización de diversas características, como por ejemplo: concentración y tamaño de las esferas, viabilidad y marcaje celulares, complicaciones al realizar las tinciones fluorescentes, etc. Mediante varios experimentos y búsqueda de información, los puntos descritos fueron mejorando hasta que finalmente el dispositivo de microfluidica fue adaptado a traction force microscopy con éxito. Por tanto, los últimos ensayos generaron datos útiles para la realización del cálculo de las fuerzas celulares. ABSTRACT In the past few decades, it has been widely demonstrated that cells constantly exert traction forces on the cells and matrix surrounding them. This behavior is fundamental in physiological and pathological processes such as embryogenesis and metastasis, therefore studying how cells interact with their environment is vitally important to understand these processes. To contribute to this research field, the present master project aims to adapt a given microfluidic device to the traction force microscopy technique, which is currently the most reliable approach for measuring cell forces. It is of great interest to quantify the forces exerted by cells seeded inside this microfluidic device because this specific device allows the deposition of collagen in a 3D arrangement, which resembles better the cell’s physiological environment than traditional 2D cell cultures. Device adaptation consisted in introduction of fluorescent microbeads into the collagen matrix, and subsequent confocal microscopy imaging of the cultured cells. Both stages required optimization of diverse features, for example: bead size and concentration, cell viability and labeling, fluorescence staining complications, etc. After numerous experiments and information search, the aforementioned features were improved, and ultimately, the microfluidic device was successfully adapted to traction force microscopy. Thus, the final assays produced useful data for performing the cell force calculation. 2 TABLE OF CONTENTS 1. INTRODUCTION ……………………………………………………………………………………… Page 3 1. 1. Biological background ……………………………………………………………… Page 3 1. 2. Cell force measurement approaches ………………………………………… Page 4 1. 3. In vitro cell cultures: 2D versus 3D environments ……………………… Page 6 2. PROJECT AIM ………………………………………………………………………………………… Page 10 3. DEVICE DESCRIPTION ……………………………………………………………………………… Page 11 4. MATERIALS AND METHODS ……………………………………………………………………… Page 13 4. 1. Microfluidic device fabrication …………………………………………………… Page 13 4. 2. Fibroblast cell culture ……………………………………………………………… Page 13 4. 3. Hydrogel solution preparation ………………………………………………… Page 13 4. 4. Cell staining and microscopy imaging ……………………………………… Page 14 5. RESULTS ………………………………………………………………………………………………… Page 15 6. CONCLUSIONS ………………………………………………………………………………………… Page 26 7. REFERENCES ………………………………………………………………………………………… Page 27 3 1. INTRODUCTION Cells within a tissue interact with neighboring cells and with the surrounding extracellular matrix (ECM) through biochemical and mechanical signals. Cell-cell and cell-ECM interactions establish a 3D communication network that maintains the specificity and homeostasis of the tissue1. In this context, Mechanobiology studies how cells detect and respond to environmental mechanical stimuli2. Particularly, the cell cytoskeleton is responsible of these interactions. It generates intracellular mechanical forces that actively respond to the perturbations occurring in their underlying substrate3. In addition, intensity of such cell forces is adapted to and modulated by both environmental biochemical stimuli and physical properties4. For instance, even for cells with the same genetic material, they behave very differently depending on the attributes of their microenvironments5. Several studies suggest a ubiquitous role for cell forces in regulating cell signaling and function in vivo 6-8. Therefore, a deeper knowledge of cell force generation and modulation is crucial to understand physiological and pathological events at tissue and organ levels9. Thus, obtaining new information on this concern appears as an important step in numerous fields of biology, including cancer research, regenerative medicine or tissue bioengineering5. Surprisingly little has been reported that sheds light on this issue, probably because there is limited access to approaches that precisely measure physiologic cellular forces in space and time10. In order to overcome this drawback, the present project develops a microfluidic device for cell culture under diverse controlled conditions and prepared for investigating the behavior of such cells. Furthermore, this platform is also adapted to fit in the Traction Force Microscopy technique (TFM), which is currently the most reliable method to measure the forces exerted by cells on their surrounding substrate9. In detail, the proposed platform consists in a microfluidic device containing cell-loaded 3D collagen hydrogel, which indeed assures a tissue-like surrounding with constant nutrient supply. The design of this approach is addressed toward reaching the most similar in vitro environment to the natural physiologic milieu, in order to obtain more realistic cell behavior data. 1. 1. Biological background The minimal structural unit that defines living organisms is a single cell, for example fibroblasts and smooth muscle cells. By proliferating and interacting with each other, cells can build complex arrangements such as tissues that ultimately organize into even more complex multicellular living organisms5. The supporting substrate within tissues, termed extracellular matrix (ECM), materializes as a mesh of crosslinked proteins (collagen, proteoglycans, elastin, and other tissue-specific molecules) and carbohydrates surrounding the cells3. Indeed, it establishes their microenvironment, where cell-cell and cell-ECM interactions occur through biochemical and physical stimuli; for instance, diffusion of soluble molecules, electrical signal transmission and transduction of mechanical cues11. Moreover, key events in the life cycle of a cell, like proliferation, migration and apoptosis, are regulated by organizing principles that are determined by the cellular context1. As opposed to passive objects such as water droplets, living cells constantly probe, push and pull their environment (ECM and adjacent cells) by exerting forces on it2 (figure 1). 4 Forces of this kind not only drive mechanical events like cell shape deformation but also are essential for mitosis and cell migration12, and have a large influence on other cellular functions such as cell adhesion13,14, gene expression15 and differentiation16. Adhesion and migration then play a key joint role in embryogenesis15, wound healing17, inflammatory responses18 and many other biological processes5. Moreover, a dysregulation of cell attachment has dramatic effects and can cause pathological states like developmental defects, cancer invasion and metastasis19. As an example, both fibroblasts’ and myofibroblasts’ mechanical activities have a critical purpose in wound healing by generating traction and contractile forces, respectively, to enhance wound contraction; but an excessive force usually results in tissue scarring20. In order to survive and grow, cells like fibroblasts and smooth muscle cells must attach to and spread on the surrounding substrate9. Once adhered to the ECM, these cells generate internal forces by their cytoskeleton, as roughly mentioned before, and these forces are then transmitted to the surrounding environment. The cytoskeleton is a highly dynamic cellular scaffolding structure fundamental to maintain and modify the cell’s shape. Besides, it is composed of diverse polymers: filamentous actin (6 nm in diameter), intermediate filaments (10 nm), and microtubules (23 nm)3. These three cytoskeletal elements are not single proteins, but consist of many monomers able to span large distances within the cell. In particular, actin filaments together with myosin II proteins form the cytoskeletal contractile apparatus at nonmuscle cells, which connects multiple parts of the cell membrane as well as the cell membrane to the nucleus21. This structure is responsible for generating the contractile forces at adherent cells so that the tension applied to the substrate is directed inward, towards the centroid of the cell10,22-24. Nevertheless there are instances in which this directionality is not true, such as when migrating cells extend protrusions at their leading edge25. In this case, traction forces are produced by actin polymerization alone. All these in and outward traction forces are transmitted to the ECM through clusters of proteins located at the cell membrane, so-called focal adhesions (FA). These structures physically link the actin cytoskeleton with the ECM. Hence, FAs are sites of tight adhesion between the cell and the underlying ECM, where both forces and signals pass across9. Furthermore, once transmitted, forces manage many cellular functions involving ECM interaction, including cell migration, mechanical signal generation, and structural deformations and rearrangements of the ECM, as aforementioned. That is the reason why a detailed knowledge of cell traction forces is crucial to understand many fundamental biological processes3. 1. 2. Cell force measurement approaches The existence of cell traction forces was first demonstrated by the ability of adherent cells to wrinkle thin films of silicone rubber26, as well as to drive matrix reorganization27. However, due to the inherent nonlinearity of wrinkling, there is currently no mathematical solution available for calculating cell forces determined through this approach20. Therefore, nowadays this technique remains as a qualitative tool and is not applicable to accurately quantify cell forces. Since then, various methods have been developed to quantitatively measure the direction and magnitude of traction forces exerted by cells on artificial substrates, with the intention of better understanding the cell-ECM interactions and their regulation. Albeit, such measurements remain a challenging problem owing to the small dimensions of the 5 adhesion sites (~µm2) and the range of magnitudes of these traction forces (nano-Newton scale)5. Due to these concerns the precise characterization of physiological cell forces in space and time remains elusive nowadays2. Nevertheless, there are currently some established methods to perform traction force measurement on cell populations or on single cells. On the one hand, to measure forces of a cell population, many studies use the cell-populated collagen gel method, in which cells are embedded in a collagen gel and shrink the gel as a result of cellular traction forces; the forces are then indirectly estimated by changes in gel volume or area. However, cells are heterogeneous and the forces they generate are often widely variable, therefore cell population-based techniques only provide an estimation of the averaged forces of a group of cells20. On the other hand, traction forces of individual cells can be determined by a variety of approaches, including the aforementioned wrinkleable substrate assay, and also the micromachined cantilever beam array28, the micropost force sensor array23,29, and traction force microscopy22 (figure 2). Cantilever beams fixed on micro-machined devices bend when an isolated cell exerts a traction force, and the bending degree is recorded and used to determine the force magnitude28. Such a technique can reliably determine the traction forces of individual cells, however it is limited in that it cannot resolve the traction force field within the whole cell area, but it can only obtain the forces in one direction. To overcome the former limitation, techniques based on micropost force sensor array (MFSA) have been developed to detect traction forces in all directions23,29. Each micropost or pillar placed in an MFSA functions as an individual force-sensing unit since it bends due to the locally applied cell’s traction forces. The development of traction force microscopy (TFM) by Dembo and Wang in 1999 has been a significant improvement to measure cellular forces3, and at present, this technique provides the most reliable, comprehensive information on cell traction forces underlying an entire cell9. In this method, elastic hydrogels with flat surface, commonly polyacrylamide gel (PA), are used as a cell culture substrate (figure 3). Fluorescent microbeads are embedded into the gel in order to act as markers of its deformation, caused by movement of the also embedded cells in there5,9,20. The substrate’s surface is further coated with ECM proteins to stimulate cell adhesion and spreading. Those cells deform the surrounding matrix, which is visualized by tracking the displacement of the beads situated in the vicinity of each cell. With that, a displacement field is determined around each cell. Finally, the displacement field values are used to calculate the cell traction forces by applying complex computational algorithms to solve the inverse problem. In TFM, bead tracking is performed through fluorescent microscopy by imaging an isolated cell and its surrounding. A pair of images of the same cell, referred to as “force-loaded” (or deformed) and “null-force” (or non-stressed), are taken during TFM measurement5,9,20. The force-loaded image is taken while the adherent cell remains on the substrate, whereas the nullforce image is taken after the cell has been removed by tripsinization or detergent treatment. The comparison of both images determines the displacement vector for every bead, which all together builds up the displacement field around a single cell; then, this data is employed in cell force calculation. There are different alternatives to this proceeding, such as taking images during cell movement instead of before and after exerting forces on the substrate30, or like treating cells with a cytoskeleton inhibitor that will prevent them from force generation31. There are three strategies to execute TFM, each of which is unique in both how displacement field is extracted from images and how cell forces are subsequently estimated9: 6 Dembo and Wang 1999, Butler et al. 2002, Yang et al. 2006. Characteristically, the latter33 introduces the finite element method in the TFM process to model the substrate as a 3D object, in order to improve the efficiency of computational calculation20. Almost all TFM studies have been applied to two-dimensional (2D) cell culture systems (figure 3). Regarding it has been shown that cellular traction forces in all three dimensions matter to fully understand cell–substrate interactions34, recently some research groups are applying more natural 3D cell cultures to the TFM measurement technique24,35. The substrate of these cultures is made of PEG (polyethylene glycol) or collagen gel shaped into 500µm-tall disks, and cells are grown in its inside (figure 4). This new feature requires the use of confocal microscopy, so as to properly visualize the bead and cell localization in the axial plane (z axis). Confocal microscopy shows an individual section or slice of the z axis, with controllable thickness. This approach can only be performed by applying laser light to the sample at the excitation wavelength of the fluorochrome inserted in the sample, in this case, the fluorescent beads. 1. 3. In vitro cell cultures: 2D versus 3D environments Although in vivo studies have been carried out to monitor cell-cell interactions and cell signaling within their native microenvironments, these studies are limited by expensive experimental manipulations (e.g., animal models), lack of control over local experimental conditions, and complex imaging setups11. Study of cell-cell interactions in vitro is advantageous due to more tightly controlled experimental conditions, higher experimental throughput, and lower costs. The traditional in vitro approach consists in culturing cells on flat plastin or glass plates, where only cell monolayers can be grown. In fact, tissue-specific architecture, mechanical and biochemical cues and physiological cell–cell communication are missed by planar, two-dimensional (2D) cell cultures, as a result of their simplified and highly biased conditions1. Indeed, this limits the culture’s potential to predict the cellular responses of real organisms. Cell culturing within 3D scaffolds mimics the specificity of real tissues better than conventional 2D cultures1. Cells cultured using traditional 2D tissue culture methods migrate different from cells in 3D environments, and moreover, they show different gene expression levels of a variety of proteins compared with their counterparts in native in vivo environments. In addition, natural cell-ECM adhesions differ in structure, localization, and function from classically described in vitro adhesions36. Hence, 3D cell cultures re-establish the physiological cell-cell and cell-ECM interactions, reducing the gap between cell cultures and physiological tissues1. The need for quantitative and physiologically relevant cellular systems has driven to develop several 3D assays, like suspending cells in a gel or plating cells on beads within a gel. However these fail in the ability to produce a well-controlled microenvironment with similar dimensions to tissue structures in vivo 37. To overcome this obstacle, a hydrogel-incorporating microfluidic cell culture assay has recently been elaborated11,37,38, which is a multipurpose platform (it has been used to study angiogenesis and tumour cell interactions, for example). Particularly, a microfluidic device or chip is a set of micro-channels molded into a material (glass, silicon or polymer), filled with a desired liquid, and connected together so as to achieve a desired function like mix, pump, redirect and/or allow chemical reactions39. This technology 7 has opened the door for creating more realistic in vitro cell culture methods that replicate many aspects of the true in vivo microenvironment, by enabling the containment of ECM-mimicking hydrogels inside the micro-channels and the introduction of biochemical gradients, shear stress or other factors, with precise spatial and temporal control. Besides, the microfluidic device can be custom-designed, i.e. its channel architecture is adaptable to the specific biological process in study and the applied methodology. Figure 1. Cell adhered to the ECM. a. Cells’ actin cytoskeleton (green) generates traction forces, which are transmitted to the ECM (blue) through focal adhesions (red spot and inset). b. Confocal fluorescence microscopy image of adherent cells. Cells’ nuclei are dyed in blue, actin fibers in red, and focal adhesions in green. Figure adapted from refs. 10, 5. Figure 2. Techniques for measuring traction forces of single cells. a. A cell produces wrinkles on a silicone membrane. b. Different magnifications of a micromachined cantilever beam array. c. Scheme and phase contrast microscopy image of a micropost force sensor array. Figure adapted from refs. 23, 28, 40. 8 Figure 3. Traction force microscopy in two-dimensions. a. Schematic representation of the TFM: cell cultured on a gel loaded with fluorescent beads. b. Deformation of a collagen gel denoted by bead movement. Scale bar, 20µm. c. Human patellar tendon fibroblast cell on a polyacrylamide gel with embedded fluorescent beads. d. Substrate displacement field. e. Recovered cell traction force field. Figure adapted from refs. 5, 33, 41. 15 5. RESULTS To begin with the project, progressive learning of the microfluidic device fabrication proceeding was carried out, and then collagen hydrogel injection required some training. Once the process was known (aprox. 74 hours), adaptation of the device to TFM was developed through numerous experiments. First assays consisted in filling microfluidic devices with collagen hydrogels containing fluorescent beads and no cells, at different conditions. This was performed in order to find the optimal bead concentration, while comparing the behavior for different bead sizes. With this, it was shown that in fact 0.5µm-diameter beads were too tiny, therefore didn’t get physically encapsulated in the hydrogel and they diffused within it in a Brownian motion (figure 6). As opposed to them, 1µm particles got well encapsulated in the collagen hydrogel and showed a uniform distribution along the three dimensions of the substrate. Bead concentration was notably differentiated at both tested sizes. And so, 1:100 and 1µm were determined as the optimal concentration and bead size conditions. The selected dilution factor is similar to the ones used in many other publications: 1:12522,32, 9:10030. Before sampling, vortex mixing was applied to the bead solution during 10 seconds, as recommended by the manufacturer, in order to avoid particle agglomeration. However, observing fluorescent microscope images of 1µm particles at 1:100 dilution assays, single fluorescent dots seemed small accumulations of some beads instead of single beads; and this was not attributable to the collagen solution’s pH, given that it was always mixed to reach a pH 7.42, while microspheres get neutralized and agglomerate at a pH inferior to 5.044. To solve this problem, it was carried out a set of experiments to analyze the effect on bead aggregation of vigorous shaking via sonication, since this treatment generates millions of microscopic vacuum bubbles within the sample that collapse violently (cavitation) driving liquid into all openings and corners of the sample particles. This strategy was both also recommended by the manufacturer and applied in some previous studies22,43. Further search was performed in order to verify that temperature increase due to sonication (until 40°C, maximum) wouldn’t affect bead properties. Polystyrene non-coated microspheres’ glass transition temperature is 100-110°C and carboxylate-coated microspheres’ is 120°C45, beyond those values particles suffer deformation. So indeed, the carboxylic surface groups provide thermal stability to polystyrene beads, hence sonication temperature rising wasn’t an obstacle for the project’s experiments. Here, 1µm beads were briefly sonicated after their suspension in culture medium supplemented with 1% penicillin and 1% streptomycin (medium-PS) in volume ratio 1:1, as described by Petroll et al. 43. Two kinds of sonicator were compared, i.e. probe and bath sonicator, and the resulting data was that beads treated with a bath sonicator were the less aggregated, even though probe sonicator was applied at multiple times and intensities. This difference occurred owing principally to the distinct application system between the two instruments. The concluded optimal bath sonicator time was 120”, since there was almost no bead accumulation. Additional assays demonstrated that both bead sonication treatment and fluorescence were maintained after 12 days in medium-PS-bead samples stored at 4°C, protected from light. Those samples appropriately mixed with collagen solution were injected into microfluidic devices the sonication day, and also 5, 7 and 12 days later, with same results. 16 A further step in this project was to add human dermal fibroblast cells to the 1:100 bead-collagen solution (figure 7). The first approach consisted in checking the distribution and viability of the added cells at 120,000cells/mL density, and qualitatively validating the size dissimilarity between fibroblasts and microspheres. Indeed, cells were randomly distributed and in a extremely low density, just as needed in TFM, since each cell has to be isolated from others to measure its single mechanical influence on the surrounding substrate and avoid confounding effects exerted by neighboring cells. Moreover, size difference between fibroblast and beads was unconditionally confirmed; and there was evidence of good cell viability as plenty of cells adhered within the collagen matrix and spread acquiring an elongated shape, and after 4 days cells had proliferated. To support this indirect viability evidence, a viability assay for mammalian cells (Live/Dead Viability/Cytotoxicity Kit from Invitrogen) was performed, which applies ethidium homodimer-1 (EthD-1) and calcein AM to the sample. Nonfluorescent cell-permeant calcein AM is converted to the intensely fluorescent calcein by intracellular enzymatic activity, which is only present in live cells. Further, EthD-1 enters cells with damaged membranes and undergoes a 40-fold enhancement of fluorescence upon binding to nucleic acids, thereby producing a bright red fluorescence in dead cells. Moreover, EthD-1 is excluded by the intact plasma membrane of live cells. Hence, this kit dyes dead cells in red by EthD-1 and live cells in green by calcein. As expected, this assay performed on sample devices revealed a 100% of viability, as all the cultured cells within a device were marked as living cells (figure 8). Next step consisted in confocal microscopy visualization of the device’s central channel, with a higher objective (63x instead of 40x), so as to closer observe cells. As aforementioned, confocal microscopes show only fluorescently marked structures. Therefore, cells were dyed in green in order to see them together with the red beads. Also, collagen fibers’ reflection at 488nm was useful to visualize them as well. To begin, some control samples were processed. With them, collagen fiber distribution was obtained: at an only collagen hydrogel sample device, fibers were randomly deposited all over the central channel; there was no difference between axial plane bottom and top fibers, nor between central and lateral situated fibers within the x axis (figure 9). The same distribution was achieved at the collagen-beads hydrogel sample device (figure 10). In addition, reflection spots and red bead spots colocalized, hence beads were wrapped by collagen fibers. However, beads didn’t act like nucleation sites for collagen fibers, altering the collagen assembling network, as reported by Newman et al. 46. Another control consisted in collagen-beads-cells hydrogel sample device, with cells not yet fluorescently marked (figure 11). This assay revealed a convergence of fibers towards the cells, which perhaps was evidence for local degradation-remodelation of the collagen matrix by the cell31,41,43. Furthermore, the number of beads surrounding each cell was observed insufficient along the axial plane. Therefore, another experiment was performed introducing cells into collagen solutions with distinct fluorescent particle concentrations, from volume ratio 1:50 to 1:1000. This assay revealed that at more concentrated solutions (1:50, 1:100 and 1:200), several beads gathered really close to the cells, particularly around the not yet spread cells, namely the round-shaped ones. Interestingly, this tendency diminished in accordance with bead concentration decrease, until the 1:500 dilution, in which beads no longer gathered (figure 12). The 1:50 dilution was selected as optimal regarding the number of beads surrounding each cell. 17 In order to better visualize the cell-bead interaction, cells were marked with CellTrackerTM Green, which is a green fluorescent dye for the whole cell. With that, cells were stained, but surprisingly they had a rounded shape instead of elongated; and also beads looked thicker than in previous samples. The washing process of cell staining with CellTracker was modified by employing NaCl instead of PBS, so as to prevent cell shape alteration. PBS was also deleted from the protocol since it contains monobasic phosphates that occasionally attach to the collagen matrix, as a lab colleague reported, therefore these phosphates could be adhering to the beads and thickening them. Confocal visualization improved a bit with the employment of NaCl, but cells were still round-shaped and, as well, fluorescence faded away rapidly with laser exposition. To overcome this obstacle, cells were dyed green with calcein, which specifically marks cell cytoplasm. This marker was the same as in the viability assay, thus only live cells would be stained. Indeed, all cells got stained, for a longer time, and were visually better defined, but even so they became round-shaped. This circumstance was not favorable to the project, because those rounded cells didn’t change their shape due to treatment with a cytoskeleton inhibitor. Despite this setback, some interesting confocal images were taken that revealed a colocalization of numerous beads and cell cytoplasm (figure 13). This demonstrates that the cultured fibroblasts were phagocyting the fluorescent beads44,47, which were probably the beads previously localized at the collagen matrix remodeled by the cell. However, there were still diverse beads surrounding the cell, fact that is still useful for TFM. Regarding that for an unknown reason cells lost their fusiform shape when stained with fluorescent dyes, transfection of the fibroblasts with GFP (green fluorescent protein) was decided to be done, in order to obtain GFP-expressing cells. This protein exhibits bright green fluorescence when exposed to light in the blue to ultraviolet range, and it has been employed to label cells in diverse TFM publications24,30,43. However, this process has to be carried out by specialized scientists and requires a long time to get fully accomplished. Therefore, in order to continue with this project, subsequent experiments were performed without cell staining, so as to work on elongated cells. Diverse assays involving collagen-beads-cells hydrogel sample device were prepared to take confocal images 5 minutes before and 15 minutes after treatment with triton X-100, which is a detergent that kills cells by disrupting their membranes (figure 14). In fact, triton influence was observed on the fibroblasts: their body lost elongation. Images stacks of this phenomena were taken every 30 seconds at the same z-axis section, situated near the centroid of a selected isolated cell. Bead movement could be observed, therefore, cells were changing their traction force exertion. However, collagen reflection images revealed that the detergent altered also the collagen matrix’s reflection pattern. Another experiment consisted in treating the sample with cytochalasin D (figure 15). This cell permeable micotoxin acts as an inhibitor of actin polymerization48, therefore it halts cell movement and force generation. As expected, the effect on cell body was weaker than with triton, but collagen reflection remained unmodified. With the cytochalasin D and triton image stacks, bead spatial localizations and displacement vectors can be recovered through computational algorithms by other members of the research group; and then, this data will be used to calculate cell traction forces. Therefore traction force microscopy has been successfully applied to a microfluidic device. 18 Figure 6. Bead distribution within the collagen hydrogel. a. Collagen hydrogel with 0.5µmdiameter beads. b. Collagen hydrogel with 1µm-diameter beads. Figure 7. Cell distribution and shape within a collagen-beads hydrogel. a. Cells adopt an elongated shape. b. Bead distribution at 1:100 dilution. 19 Figure 8. Cell viability assay. a. Cells marked in green demonstrating they are alive. c. There is nothing labeled in red. 20 Figure 9. Collagen fibers distribution in a collagen hydrogel by confocal microscope with 63x objective. a. Fiber distribution near the central channel’s posts. b. Zoom image of a. at 3.7x. c. Fiber distribution in the middle of the central channel. d. Zoom image of c. at 2.3x. 21 Figure 10. Confocal image of collagen-beads hydrogel. a. Phase contrast image (not confocal). b. Bead distribution. c. Collagen fiber distribution. d. Overlap of b. and c. to emphasize colocalization (colored in purple). 22 Figure 11. Collagen-cells hydrogel. a. Phase contrast image of two cells. b. Confocal image of collagen fiber distribution underneath the cells. Figure 12. Gathering of beads near the cells. a. and b. Cells within a collagen hydrogel with beads at 1:200 dilution. c. and d. Cells within a collagen hydrogel with beads at 1:500 dilution. a. and c. Phase contrast images of cells. b. and d. Confocal images of beads. 23 Figure 13. Calcein fluorescent staining of cells, in green. Two different samples. 24 Figure 14. Triton detergent treatment. a. Image stack before. b. Image stack after 15 minutes administration of Triton 1x. a. and b. Upper images are phase contrast for cell visualization, middle images are confocal imaging of the embedded beads, and lower images are confocal imaging of collagen fibers’ reflection.