CFM: Confinement Force Microscopy-a dynamic, precise and stable microconfiner for traction force microscopy in spatial confinement

In vivo, cells experience complex tissue environments with various chemical and physical features. They sense and respond to tissue morphology and mechanical properties and adjust their behavior and function based on the surrounding. In contrast to the free environment experienced on 2D substrates commonly used in research, the 3D natural environment represents a major physical obstacle for cells. Here, cells are usually confined either by the extracellular matrix (ECM) or neighboring cells. The importance of such confinements has been demon-strated in the past decades by showing its influence on cell decision-making in many vital biological processes such as migration, division and cytoskeletal reorganization. Despite these insights, the sheer level of complexity faced when studying cell biological questions in biomimetic confined situations, led to an indispensable need for a 3D system which can simulate the in vivo confined condition, while being capable of providing microenvironments with different chemical and physical properties for the cells and capturing the mechanical forces and properties of the studied biological sample. Here we introduce a microconfiner that finally provides a new imaging capacity, namely the confine-ment force microscopy (CFM). We are able to adjust the confinement level in real time during microscopy while measuring not only the the cellular traction but also the cellular compression forces. Furthermore, the chemical and physical properties of the microenvironment can be optimized for the respective questions. We demonstrate the power of this confinement system by the mechanical response of cells, migration analysis of immune cells, the timed force generation during durotaxis driven adhesion switching and the viscoelastic properties of cancer tissue.


Abstract
In vivo, cells experience complex tissue environments with various chemical and physical features. They sense and respond to tissue morphology and mechanical properties and adjust their behavior and function based on the surrounding. In contrast to the free environment experienced on 2D substrates commonly used in research, the 3D natural environment represents a major physical obstacle for cells. Here, cells are usually confined either by the extracellular matrix (ECM) or neighboring cells. The importance of such confinements has been demonstrated in the past decades by showing its influence on cell decision-making in many vital biological processes such as migration, division and cytoskeletal reorganization. Despite these insights, the sheer level of complexity faced when studying cell biological questions in biomimetic confined situations, led to an indispensable need for a 3D system which can simulate the in vivo confined condition, while being capable of providing microenvironments with different chemical and physical properties for the cells and capturing the mechanical forces and properties of the studied biological sample. Here we introduce a microconfiner that finally provides a new imaging capacity, namely the confinement force microscopy (CFM). We are able to adjust the confinement level in real time during microscopy while measuring not only the the cellular traction but also the cellular compression forces. Furthermore, the chemical and physical properties of the microenvironment can be optimized for the respective questions. We demonstrate the power of this confinement system by the mechanical response of cells, migration analysis of immune cells, the timed force generation during durotaxis driven adhesion switching and the viscoelastic properties of cancer tissue.

Introduction
The physical environment experienced by a living cell in an organism has only little to do with the situation faced by cells commonly cultivated in research labs. While the optimized situation in a petri-dish often ensures that cells do hardly interact with each other, cells in living tissues are constantly pushed and pulled by their complex environment. Throughout their life, cells experience distinct microenvironments of varying mechanical and geometrical properties. It turns out that these physical differences can direct cell signaling pathways, decisions, responses and fate. [1][2][3][4][5] Among these physical environmental features, confinement is of great importance. Cells could be confined either by the extra cellular matrix (ECM), 6 or the fellow cells during the tissue formation 7 and collective cell migration. 8 Therefore, many cells in the body have to cope with the limited spaces, small junctions and narrow tracks during their emergence, growth, proliferation and migration. 9 To perform their task flawlessly in many different environments, cells need to arrange with such constrictions. Immune cells and epithelial cells are well-known examples. For instance, circulating from bone marrow to the site of infection, immune cells have to squeeze through many small constrictions, starting with the tiny junctions between endothelial cells when transmigrating, and numerous small pores when passing through the connective tissue. 10 Similarly, while proliferating and migrating in the epithelial sheet, epithelial cells experience high confinement imposed by their neighboring cells. 11 Therefore, confinement by physical barriers is a ubiquitous phenomenon in the body, and plays a critical role in directing cell's decision-making in many biological processes and exploring its effects on cell behavior is of great importance. To explore these impacts requires an instrument that closely mimics the conditions experienced in the body while still being capable of providing controllable chemical and physical microenvironments for them. It should be fully biocompatible, precisely adjustable and accessible to media flows, temperature and pH control. There are many studies that used confined cells to investigate different parameters, thereby exploiting several methods to achieve this confinement. For instance, microchannels with different shapes 12 were used, but also collagen matrices 13 and 2 parallel planes with spacers 2 to keep a constant distance. While providing excellent stability, these approaches are limited to fixed confinement levels and cannot change the confinement during imaging. Besides the instruments created by researchers, 14 also a commercial confiner was introduced (Microsquisher-CellScale). 15 However, this design only provides a side view on the confinement, thus not allowing to record cellular properties and responses during key biological processes like migration. Additionally, this optical access prevents the usage of high numerical aperture objectives, which are necessary for sub-micrometer resolution. Hence, it is not possible to observe and record the effects of confinement and forces on single cells, or on larger systems with subcellular resolution. Moreover, it does not allow to confine the sample to the µm range, and long-term stability is not assured. Atomic force microscopy (AFM) is another powerful approach use by researchers to confine cells and measure their mechanical properties. 16,17 While AFM provides excellent force and confinement resolution it can only measure single cells without access to lateral forces, and has hence severe drawback regarding throughput. As a consequence, a comprehensive and adjustable instrument is required to create a dynamic confinement while providing access to live imaging and extraction of traction and compression forces during controlled confinement on the cell and tissue scale. To fulfill this need and overcome the mentioned limitations, we introduce Confinement Force Microscopy (CFM) and a corresponding CFM-instrument (also called CFM in the following) using a piezo driven motorized linear stage and a precise capacity sensor, which is able to adjust confinement during live imaging and to simulate the in vivo mechanical and chemical environments experienced by cells and tissue. It is compatible with a standard glass bottom petri-dish and provides optical access via a high resolution objective (NA>1) in a typical commercial inverted microscope. CFM is small (<20cm) enough to be placed on common microscope stages, or to locate it in a incubator, which enables long term studies. It is also lightweight allowing mounting on a high speed translational microscope stage for imaging. Using well characterized polyacrylamide (PAA) hydrogels as substrates, allows a large range of environment stiffness as well as measuring the forces exerted on the substrates. This leads to several important applications of CFM. First of all, we are able to measure the mechanical properties of the biological entities from the cell and nuclei scale up to spheroids and whole embryos. Here, parameters such as viscosity and elastic modulus are obtained which would be specially helpful for modelling cell and tissue behavior. Moreover, it is able to provide information about changes in mechanical properties during biological events, such as cell migration, growth, mitosis and even development. Additionally, we are able to study 3D migration of cells, for example 3D durotaxis. Here, the strengths of CFM become a key advantage as it allows to adjust the confinement during the live imaging based on the cell height and spreading. Furthermore, another important feature of CFM is to exert a periodic or oscillatory force and confinement in many temporal patterns on the cells and bigger biological entities such as organoids, cell spheroids or even embryos, and track their mechanical deformation and force transmission and response. Finally, CFM enables its user to measure forces exerted by cells on their confined environment during different vital biological processes such as cell migration, division, apoptosis and embryogenesis, which provides us an outstanding opportunity to shed light on many critical open questions in the field of cell mechanics and mechanotransduction.

Design and characterization of CFM
CFM is composed of a piezo driven motorized linear stage and a precise capacity sensor (Figure 1a). The piezo stage enables us to dynamically adjust the confinement, from mm down to smaller than µm, hence covering the working distance of any common objectives. As both cell size and spreading levels vary during experiments, the relative confinement is a cell dependent number even when fixing the distance between the substrates. Here the motorized stage enables us to easily change the confinement level during the live imaging based on the experimental condition in real time and image and record cellular behaviors like migration and cell area change under the desired confinement level. Additionally, the sensor provides a long-term stability of the confinement level (experimentally tested for up to 20h) by a closed feedback loop with the piezo motor ( Figure 1a). After adjusting the distance between the substrates, even 500 nm of change in the sensor position will lead to a motor movement to compensate for the change. The CFM instrument consists of a metal body (Figure 1a, number 1), that is designed to be inserted in standard microscopy stages for imaging ( Figure 1d). The upper part of the confinement is created by a lever-arm (Figure 1a, number 5) and a cylinder (Figure 1a (Figure 1a-b, number 6), which itself is screwed to the metal lever arm (Figure 1a, number 5) of the CFM. The mechanical stability of the glue is crucial for the stability of the CFM and can be achieved by a UV activated glue that can be washed away with warm water after the experiment to reuse the metal cylinder. The cylinder is made of 2 different materials: plexiglass to combine the CFM with brightfield microscopy and stainless steel for more stable fluorescent microscopy. Cells are later seeded either on the bottom gel or the top gel based on the question (Figure 1c). The bottom plane of the CFM has the same dimensions as a standard 6-well dish, which allows insertion in microscope stages for live cell imaging (Figure 1d).

Confinement stability over time
One of the most important requirements to study biological systems under confinement is to ensure a stable confinement level during a long-term experiment. Effectively, the smaller the biological entity to be studied is, the more precision and stability of the instrument is required. Here, the aim is to confine the cells down to 1 µm or less, corresponding to a required accuracy of less than 1 µm. To study and quantify the CFM stability capacities, we measured the distance over a 20h time window. Figure 1e and supplementary video 1 give an example of such measurements, in which the maximum change of distance between the  The main goal to develop CFM was to overcome the lack of a suitable method that combines confinement, live microscopy and traction force measurement. After demonstrating the stability of confinement, it was next tested if CFM allows to change the confinement while recording 3D image stacks and to measure the traction forces during the confined cell behavior. By incorporating fluores-cent beads in the PAA gels during gel fabrication (Figure 1c), it is possible to track the 3D deformation of the gels during confinement and hence to measure the 3D stress following an adapted version of traction force microscopy (See materials and methods). Figure 2a and supplementary video 2 demonstrate a step by step confinement of a HeLa cell during live imaging while extracting the lateral (Figure 2b) and normal stress (Figure 2c) that the cell exerts on the bottom PAA gel. Please note that the stress map images are given as the 2D (xy) lateral stresses obtained from the 3D stacks. As it is observed in this figure we have changed the distance between the gels from around 12µm to 5.4µm. In principle, the distance can change from that maximal range of the linear motor down to 0µm. The red planes in Figure 2a are the surface of the bottom and top gels and the thickness of the PAA gels is usually between 50-80 µm. Figure 2d represents the average lateral and normal stress of more than 100 HeLa cells in response to different strain levels extracted from the step by step confinement experiments. Cells have undergone 3 steps of consecutive confinement with the time interval of 5 minutes. Surprisingly, there was no significant trend (increase or decrease) in the normal stress when cells were more confined and there was a considerable reduction in the lateral stress in most of the cases with increasing the confinement level. This suggests that the observed cells regulate traction forces to a constant value despite a confined situation. To assure that this was the active response of the cells, the experiment was repeated with fixed cells using 4 % PFA solution, to stop cellular activity. As expected from a passive elastic object, fixed cells led to a very different result, where we observe a normal and lateral stress increase upon a cellular strain (Figure 2e). This implies that living cells are capable of adapting their lateral and normal stress generation to new mechanical environments.

Results
Confinement impacts various biological phenomenon such as embryogenesis and cancer metastasis. 18 For instance, it is known that the mechanical properties of a microenvironment 19 and the confinement level 20 can affect the initiation and type of cancer metastasis formation, which emphasizes the importance of studying their impact. CFM gives us the ability to study the simultaneous effect of these 2 key variables in an efficient way. In this regard we demonstrate CFMs capacity for various biological applications such as: 1. measurement of confined cells' traction stress, 2. investigation of cell durotaxis under confined environment, 3. application of periodic forces on cells and organoids with arbitrary local and temporal patterns, 4. measurement of forces during biological processes such as embryogenesis, 5. measurement of stress on the microenvironment during cell death, and 6. measurement of the mechanical properties of live cells and tissues, such as elastic modulus, viscosity and stress relaxation.
In the following, we represent some of our experimental results for each of the mentioned applications.

Effect of confinement level and microenvironment stiffness on neutrophil migration and its nucleus
Neutrophils are born in the bone marrow and to reach the site of infection they have to move through very tight junctions and undergo extreme cellular and nuclear deformations, such as in transendothelial migration. 21 To study how these fast migrating cells can undergo such an extreme compression, a confinement instrument is required which can capture neutrophil response immediately after confining them; otherwise, due to their very quick response to the external mechanical signals, it would not be possible to find out the mechanism behind their confined behavior. Using CFM it is possible to confine these cells during live imaging and thereby capture their very early responses to external forces. To investigate how the microenvironment stiffness affect the responses of neutrophils to confinement level, we varied both, confinement and stiffness, and probed the cells 3D traction stress in response to these changes (Figure 3a-b and supplementary video 3). In order to simulate physiological in vivo conditions experienced by neutrophils, we exposed them to three different substrates stiffnesses, namely Young's moduli of 1.2, 9 and 21 kPa. In an initial step (Figure 3a), we categorized our experimental design to 2 main groups: First, unconfined cells, seeded on a single PAA gel layer, and second, confined cells, seeded between two layers of PAA gels of the same stiffness. The confinement level was adjusted in a range from 5.4 to 7.8 µm (Figure 3a). Our results demonstrated that independent of the microenvironment stiffness applied, neutrophils exert in general higher lateral and normal stress on their environment when confined. This suggests that neutrophils respond to the external confinement by an increase in pressure and traction forces. We also observed an increase in lateral and normal traction stress as function of substrate stiffness both in unconfined and confined situations (Figure 3a), which is consistent with the well known effect of mechanotransduction, so far only studied in unconfined systems. [22][23][24] Lateral and normal stress increased by enhancing the substrate stiffness but saturated in confinement already at 9kPa substrate stiffness. Besides comparing confined and unconfined situations, we also tested the stress response upon increasing confinement levels at different substrate stiffness (Figure 3b). To this end we applied 3 different confinement ranges of 2.4-4.8 µm, 5.4-7.8 µm and 8.4-10.8 µm. We also used a higher distance (about 30 µm) between the gels as our control group, which assured a situation without contact between the neutrophils and the top gel, hence resembling a 2D substrate for the cells. We observe that increasing the confinement level leads to an increase of both lateral and normal stress in a soft environment (1.2 kPa). In the medium and stiff environment (9 kPa and 21 kPa, respectively) the picture changes slightly as here both lateral and normal stress still increase for higher confinements, but after a critical confinement below 5.4 µm compressed neutrophils reverse their stress response which is measured as a decrease in forces at high confinements. Please note that the normal stress in the 9kPa situation was not significantly decreasing at high confinements, but still breaking the trend of increased normal forces upon increased confinement as it would be intuitively expected ( Figure 3b). In short, these results suggest that confined neutrophils adjust their pressure and lateral forces on the microenvironment as a strategy to cope with the external compression and difficulties on their way. The fact that such adjustment is not observed in HeLa cells points out the importance to study such properties for a better understanding of cell type dependent mechanical response to confinement. An interesting fact of neutrophil confinement is the reorganization of their nuclei during their confined migration, which was suggest to be a very effective factor for their efficient migration even in strong confinements. 25 To further demonstrate the potential of CFM, we labeled the nucleus with a fluorescent life marker (SYTO™ Red Fluorescent Nucleic Acid Stain Sampler Kit, Nr.60, ThermoFisher) and tracked the nuclear shape changes in response to confinement. We observed nuclear shape changes in response to the high confinement ( Figure 3c and supplementary video 4) and specially when facing another cell on their migration way (Figure 3d and Supplementary video 4). To test this further, we measured the surface area of the nucleus under different levels of confinement and compared them with the cell surface area. Surprisingly, we did not observe any significant change in the surface area of neutrophil nucleus under different levels of confinement. In contrast, the cell surface area increases drastically for confinements below 5.7µm (Figure 3e). It seems that this multilobular nucleus helps neutrophil to migrate through the small junctions by its capability of remarkable shape change.
In short, CFM provides the possibility to correlate nuclear shape changes with force generation and migration characteristics, which will pave the way to new insights about active migration properties of immune and other cell types.

Neutrophil 3D durotaxis
Although 3D confinement is one of the major physical obstacles for cell migration in their natural environment, it remains difficult to study under controlled conditions. In addition, most migrating cells experience various microenvironmental stiffnesses while migrating through different tissues. Neutrophils are among these cells. Being born in bone marrow they have to migrate through very small junctions and pores of various tissues with different stiffnesses to reach the cite of infection and accomplish their task. Therefore, they experience a large range of tissue stiffnesses during their journey, ranging from 300 Pa (brain) to 10 GPa (bone). 26,27 These cells are also responsive to their microenvironmental stiffness by adjusting their adhesion and spreading. 28 Therefore, potential stiffness gradients in the environment can be a directional factor for confined migration, a process called durotaxis, which is well document in 2D cell culture. CFM gives us the ability to study 3D durotaxis of these rapid moving cells in a confined environment, which is difficult to achieve in classical gels that expose the cells to a 2D stiffness gradient along the gels. By preparing two different gels for the top and bottom and exposing the cells to two different stiffness, we are able to precisely trigger the moment where neutrophils are experiencing a very well defined stiffness gradient. To this end CFM allows to capture the cells' immediate response to confinement and steep stiffness gradients that trigger durotaxis. Here, we confine neutrophils in between 2 layers of PAA gels with 2 different stiffnesses (Bottom: 1.2 kPa and top: 114 kPa) while maintaining the confinement stable for the desired imaging time period until we observed the complete durotactic switch from the soft to stiff substrate (Figure 4a and supplementary video 6). By registering the recorded traction forces exerted on the soft substrate with the shifting process we observe a surprising and transient stress peak on the soft substrate during neutrophil durotaxis (Figure 4b). This peak occurred when neutrophils adopt a cylindrical configuration between the bottom and top substrates (Figure 4a, at 5:33 min the stress peak occurs). This situation resembles the moment when the attachment area was equal on the soft and stiff side. By calculating the forces based on the stress and the cell contact area with the soft gel, we realized that the total forces are almost unchanged on the soft substrate from the first time point (cells completely spread on the soft gel) to the time point that cells adopt the cylindrical shape between the gels. Shortly after the equal area situation, the forces suddenly decrease on the soft gel and cells start to completely spread on the stiff substrate while detaching from the soft substrate (Figure 4c). That is a very intriguing behavior which might originate from the focal adhesion or cytoskeletal reorganization during durotaxis.

Periodic force application on the cells with many arbitrary temporal and local patterns
Cells and tissues in the body are often under periodic forces with variable patterns. Bladder, 29 cartilage, 30 muscle ? vessels and endothelial cells 31 are only a few examples of them. Given this high relevance of periodic forcing in cell and tissue, the systematic reaction of cells to such mechanical interactions becomes critical for a full understanding of the biomechanical circumstances that influence general cell biology. 32 To this end, CFM gives us the opportunity of exerting periodic compression forces with controlled local and temporal patterns on the cells and tissues in different scales while simultaneously tracking their behavior. Thereby we can monitor not only possible changes in their mechanical properties but also the stress that cells exert on their environment in response to the periodic forcing and investigate possible fatigue like long term changes in force generation upon mechanical stimulation. To assess the cell mechanical response to periodic confinements, we expose HeLa cells to periodic confinement with different temporal patterns.  Due to the high range of imaging in Z direction for capturing both gels, a very high speed camera is required to be able to capture both gels every 6 seconds.)

Measuring force generation during development
Embryogenesis presents the fundamental process during the development of any organism, and any significant disturbance of this vital process has catastrophic consequences. Many species are developing outside the safe harbour environment represented by the mother, and are hence often subject to drastic mechanical challenges. Although many mechanical forces are shielded by a shell-like structure, the study of force generation during development, and the reaction of embryogenesis to external loads gives deep insights into developmental processes. To demonstrate the power of the CFM to study classical developmental model systems, we confined a developing Drosophila melanogaster embryos and observed/measured the forces/ the force generation during cellularization.
For Drosophila development, cellularization as the first tissue formation step is of great importance and includes membrane-coupled actin remodeling and force generation. 33 However, the role of mechanical forces exerted by the outer embryo wall while triggering cellularization and tissue formation has never been studied on a global scale. To demonstrate the versatility of the CFM to measure the stress and force generation during cellularization of the Drosophila embryo, we started to image Squash-GFP (Drosophila myosin II homologue) expressing embryos just prior to cellularization (Figure 6a and supplementary video 8).
Our results demonstrate that despite the controlled confinement, the embryo can go through the process of cellularization without a detectable phenotype. Furthermore, the obtained PAA gel deformation images indicate that there is a small increase in the normalized deformation equivalent of bottom gel while the actomyosin ring is separating from the embryo membrane (Figure 6b). This can be interpreted as an increase in pushing forces on the environment during cellularization.

Measuring the stress on the microenvironment during cell death
Apoptosis, or programmed cell death, is vital for forming structures and the elimination of unnecessary cells during embryonic development, tissue homeostasis, and in some pathological conditions. 34 The mechanical forces produced during apoptosis are not only important for forcing dying cells out of tissues in order to keep tissue integrity, but also for altering the morphology of neighboring cells to fill the space originally occupied by the dying cell. 35 These forces could also play a role in some other biological processes such as tissue fusion process (dorsal closure) during Drosophila embryogenesis 36 or hair follicle regression. 37 To test the capacities of CFM to study the general process of cell death we recorded C2C12 cells (Precursor muscle cells) that were confined in between the PAA gels with the stiffness of 1.2 and 21 kPa for the bottom and top gels, respectively. The C2C12 cells were stained with CellMask Deep Red plasma membrane stain which is known to be toxic for the cells when exposed to the high power laser. 38 Cells were kept in confinement for about 12 hours and exposed to the high power imaging laser every 30 min. The process of cell death initially showed a reduction in spreading area, while the cell simultaneously enhance their sphericity, thereby effectively indenting the gels before the plasma membrane ruptures and the cells disintegrate (Figure 7a and supplementary video 9). Since the bottom gel is about 20 times softer than the top gel, the cell is able to deform it readily to make space for its height increase. The lateral and normal stress were measured on the bottom gel during the cell death, which consistently exhibit a stress peak exactly at the moment of the highest deformation and subsequently the stress drops down to almost 0 (Figure 7b), confirming the cell death. Figure 7c and d represent the average lateral and normal stress of several cells, which is normalized by the minimum stress on the bottom gel for each cell. The results indicate an average increase of about 2 times for the lateral stress and about 5 times for the normal stress during the cell death. Viscoelasticity is one of the unique properties of each cell and tissue and is different in various cell lines and tissues. [39][40][41][42][43] The residing tissue viscoelasticity could be a determinant factor of the cell fate. 44,45 Likewise, tumor viscoelasticity could affect the phenotype and proliferation of malignant cells. 46 Therefore, viscoelasticity is an important feature to distinguish healthy tissues from nonmalignant and malignant tumors. Using clear measurements of viscoelasticity can hence become instrumental in early diagnosis of cancer in patients with the potential to rescue many lives. In this regard, we utilized CFM to measure the elasticity and viscosity of cancer spheroids formed by different cancer cell lines. By confining HeLa and CT26 spheroids while recording the spheroids strain and stress relaxation over time (Figure 8a-c and supplementary video 10), we obtain stress strain relations. These can be explained by a mechanical Maxwell model (Figure 8c) which fits well the relaxation data and allows calculating the elastic and viscous properties of the reconstituted cancer tissue. Our results demonstrate that although the stiffness of CT26 spheroids and HeLa spheroids are not significantly different, CT26 spheroids are considerably more viscous than HeLa spheroids (Figure 8d-e). Supplementary video 10 displays this fact very clearly. It is visible that HeLa cells are moving out of the spheroid and the spheroid is behaving more fluid like compared to the CT26 spheroid, were the superficial cells are strongly adhered to the spheroid and surface tension seems to be higher. Our results are highly consistent with micropipette aspiration measurements from previous studies. 47 These results demonstrate that CFM might be a suitable alternative for complicated systems such as micropipette aspiration.

Discussion
CFM (Confinement Force Microscopy) is a very efficient tool to study cell and organoid behavior under confinement and analyze possible correlations with the external mechanical signals. Due to its ability to adjust the confinement level during the live imaging, it provides the opportunity to study the effect of confinement on very fast moving cells such as immune cells while recording their very first and vital responses to external pressure. CFM enables the dynamic change of the confinement level based on the experimental conditions as well as cell size and spreading area that are observed during the imaging. Providing access to the 3D traction forces of cells and tissues in confinement, CFM is a great tool for study of not only cell and tissue mechanics but also active cellular responses such as mechanotransduction. Due to providing a large range of top gel movement, CFM is suitable for confining biological entities of different scales. We demonstrated successful measurements on a very small scale such as the cell nucleus but also on very large scale such as cancer spheroids and embryos. The bottom and top gels can be made in any desired size and stiffness and can be coated with different chemicals, which makes it suitable for mimicking different in vivo conditions. It can be also combined with other confinement tools such as collagen fibers in between the 2 substrates to increase the complexity of the environment and make it even closer to the natural condition. 40 Due to the large size of substrates compared to the cell and organoid size, we can track the behavior of several hundred cells and multiple organoids simultaneously, which is not possible with other more specialized systems such as an AFM. CFM is user friendly and fast adjustable. The mobile top gel provides the opportunity of exerting periodic compression on the cells and tissues, which is most relevant to their in vivo experiences. The periodic compression can be applied with many arbitrary temporal and local pattern based on the experimental question and to our best knowledge is never done. We can investigate cell and organoid biological response, forces and mechanical properties in different time scales, which based on our experiments can be very different. Comparison of Figure 2d and Figure 5b demonstrates that cells respond differently in a short time scale compared to the longer one. Figure  In short, this system is instrumental to answer many open questions in the field of cell and tissue mechanics while many potential additional aspects could be added to the system. It could be also combined with other biological or biomechanical tools which already exist in this research area.
6 Materials and Methods  If not otherwise stated, all chemicals were purchased from Sigma-Aldrich (Steinheim, Germany). HeLa and CT26 cells were cultured as previously described. 40 Briefly, HeLa cervical cancer cell lines and CT26 cancer cell lines (both kind gifts from Danijela Vignevic, Institut Curie Paris) were stably transformed to express lifeact-GFP and H2B-mCherry, via lentiviral transduction. All cells are confirmed negative for mycoplasma contamination using PCR tests. HeLa, CT26 and C2C12 were cultured in high glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) of fetal bovine serum (FBS, Capricorn, Germany) and a 1% (v/v) of penicillin-streptomycin solution (Gibco, germany), incubated at 37°C with 5% CO2 in a humidified atmosphere. Their culture medium was exchanged every 2-3 days and they were split on reaching the confluency of ¿70%. To split the cells, they were first washed with PBS (modified, without calcium chloride and magnesium chloride) and suspended using trypsin-EDTA (0.05 %), phenol rot (Fisher Scientific, Gibco). Subsequently, they were centrifuged and resuspended in the culture medium.

HoxB8 derived neutrophil culture
HoxB8 cell lines were extracted from bone marrow of Lifeact-EGFP mice as previously described 49 (a kind gift from Johannes Roth, Institute of Immunology, University of Münster) and cultivated in a clear, non-treated polystyrene 6-well plate (Falcon® 6-well Clear Flat Bottom Not Treated Cell Multiwell Culture Plate, with Lid, Sterile, Corning, USA). We used Opti-MEM 1X + GlutaMAX (Gibco, Germany) supplemented with 10% of FBS, 1% of penicillin-streptomycin solution and 30 µM of β-Mercapthoethanol (50 mM, Gibco, Germany) as the base culture medium. Cells were suspended in the base culture medium containing 1% of SCF-supernatant (Stem cells growing factor derived from CHO cells, a kind gift from Johannes Roth) and 10 mM of β-Estradiol. The culture medium was replaced every 2-3 days and cells were split to 2*10 6 cells/3ml (every second day) or to 3*10 5 cells/4ml (every third day) and tested for mycoplasma contamination prior to experiments. To differentiate HoxB8 cells to neutrophils, cell suspension was centrifuged and washed twice with PBS containing 10% FBS to remove β-Estradiol. Subsequently, HoxB8 cells were suspended in 3 ml of base medium medium supplemented with 1% of SCF-supernatant in untreated six-well culture plates. Cells were differentiated for 4 days prior to the experiments.

Cancer spheroid formation
Cell spheroids were made as previously described. 40 In short, HeLa cervical cancer cell lines and CT26 cancer cell lines were washed with PBS (modified, without calcium chloride and magnesium chloride) and suspended using trypsin-EDTA (0.05 %), phenol red. After centrifuge they were resuspended in their culture medium. 48-well plates (Greiner Bio-one) coated with 150 µL of 1% ultra-pure agarose (Invitrogen) that was cooled down for 30 min were used to prepare cancer spheroids. 1 mL of cell suspension containing 1000-2000 cells were added in each well, and the aggregates then formed spontaneously. The aggregates were collected after 2-3 days, transferred to confinement and imaged with a 20x objective.

Fly stocks and Drosophila embryo preparation
We used sqhAX3; sqh-Sqh-GFP (BDSC 57144) fly stock from Bloomington Drosophila Stock Center. Live imaging was performed at 25°C. Embryos were collected on yeasted apple juice plates, dechorionated using bleach for around 2 minutes, and then washed thoroughly with distilled water. Dechorionated embryos were mounted and covered with halocarbon oil series 700 (Halocarbon Products).

Confinement and imaging
To prepare cells for confinement and imaging, they were centrifuged and resuspended in the proper medium (differentiation medium for neutrophils and CO2-independent medium (Gibco, Germany) supplemented with 10% (v/v) of FBS and 1% (v/v) of penicillin-streptomycin solution for HeLa, CT26 and C2C12). Subsequently, they were seeded on the fibronectin-coated PAA gel at the appropriate number (20,000-25,000 of neutrophils, 5,000-10,000 of HeLa and CT26 and 3,000-5,000 of C2C12 per gel of 12 mm diameter) and incubated for at least 30 min. Before imaging 2 ml of the proper medium supplemented with 25 mM HEPES (Millipore, Burlington, USA) was added to the petri-dish. Cancer spheroids and Drosophila embryos were seeded on the uncoated PAA gels. After transferring to confinement cancer spheroids were covered with the CO2-independent medium supplemented with 10% (v/v) of FBS, 1% (v/v) of penicillin-streptomycin solution and 25 mM HEPES while drosophila embryos were covered with halocarbon oil series 700. Within the next desired time period, 4D stacks were captured at the required time intervals (based on the experiment) and desired z plane distance using the spinning disk system (CSU-W1 Yokogawa) combined with a heating chamber set at 37°C for the cells and spheroids and 25°C for Drosophila embryos. In order to investigate the effect of confinement, the glass bottom petri-dish containing the bottom PAA gel and seeded cells was placed in CFM. The position of the bottom gel was recognized using the fluorescent microscopy and subsequently the distance between the bottom and the top gels was adjusted, based on the question and the size of the entity in between, by moving the top gel downward. Cells were imaged with a 60x objective and cancer spheroids and drosophila embryos were imaged using a 20x objective, utilizing the Slidebook software (3I, USA) and analyzed with ImageJ.

Traction force microscopy and stress calculation
After imaging the confined cells for the desired time duration and with the appropriate time interval, the upper layer was moved upwards and 100 ml of 10% sodium dodecyl sulfate (SDS) as a membrane lysis buffer was gently added. Afterwards, the upper layer was moved back to the previous position and the relaxed bottom and top gels were captured as the reference (i.e. null force) images. Traction stress in X-Y (Lateral plane) and Z directions (Normal plane) for single cells were calculated as previously described. 38,48 In short, PAA gel deformations were calculated based on the displacements of the fluorescent beads on the gel surface with respect to the relaxed gel after the SDS addition. Displacements were calculated by an iterative free-form deformation algorithm using the open source software Elastix. 50 To analyse with the Elastix method, Using the cubic B-spline functions, the 3D reference image was iteratively deformed and compared to the respective timepoint of the 3D image of the deformed gel.
The iteration process started with distributing the control points of the Bspline functions over a regular mesh. The control points' positions and, therefore, the applied deformation was adjusted in each iteration to optimize the quality of agreement between both images measured by the advanced Mattes mutual information metric. To do so, the tuning steps followed an adaptive stochastic gradient descent algorithm. 38 The calculation progressed in a three-level pyramid approach from the coarsest to the finest scale and the comparison between images in each iteration was not carried out on the full images but on a randomly sampled subset of each image. For embryo and spheroid experiments the displacement calculation was performed differently. The lateral displacement was derived using a PIV algorithm that compared image sub-windows by an upsampled cross-correlation employing a discrete Fourier Transformation (DFT) method, 48, 51 which uses the phase difference between the DFTs of two images to achieve subpixel resolution. The z-displacement in this approach was determined by a subpixel registration of the PAA gel surface as introduces recently. 48 The final deformation result was inserted into the equation for the Tikhonov regularized inverted elasticity problem for finite thickness substrates. We exploit a Bayesian estimation of the regularization parameter that includes an estimation of the background by explicitly marking regions of the images without cells, following the approach introduced by Huang et al. 52 The equation was solved in the Fourier domain 53, 54 using a custom-made MATLAB (MathWorks) program yielding the 3D traction forces exerted onto the gel's surface. 38 The stress numbers reported here are the average stress of the contact area of the cell with the bottom gel, which is calculated by making the cell mask on the bottom gel using ImageJ.

Statistical analysis and visualization
3D images are made using ImageJ 3D script. Diagrams are sketched by Python, Matlab and GraphPad Prism. Error bars reflect the mean±s.d. Statistical comparison between two groups of data was performed either via a two-tailed t-test or a Mann-Whitney-Wilcoxon test depending on whether the data was qualified as normally distributed at a 5% confidence level according to a Shapiro-Wilk test. In case either or both of the two compared data sets did not display a normal distribution, the Mann-Whitney-Wilcoxon test was employed to evaluate the significance of potential differences between the compared data sets. The data supporting our findings, and the codes utilized, are available from the authors upon reasonable request.