Micropatterning strategies to engineer controlled cell and tissue architecture in vitro

The fundamental relationships between tissue structure, patterning and function, as well as how these are disrupted during disease, have puzzled biologists for decades. In the last 15 years, cellular micropatterning has emerged as a powerful tool to explore and dissect the relationship between cell and tissue architecture and resulting function (1, 2). Micropatterning techniques allow users to position cells in areas of a substrate in vitro, enabling the control of cell shape, position, or multi-phase tissue architecture (3, 4). This offers an opportunity to replicate the complex architectural and boundary conditions characteristic of in vivo tissues (1, 5) and therefore dissect the contributions of distinct cellular processes and/or property parameters to the cellular or tissue phenotype (2). For example, cells are responsive to geometrical and mechanical cues present in their microenvironment, and micropatterning provides an opportunity to overcome the lack of such cues when using standard tissue culture (1). Today, a number of micropatterning approaches of varying complexity are available to address problems at both the cellular and tissue scale. Taken together, these provide a powerful platform for asking fundamental questions in cell biology and tissue engineering, including cell survival, proliferation, differentiation, cell migration, cytokinesis, and cell polarity (6, 7).

At a basic level, micropatterning approaches involve controlling cellular attachment, shape and spreading as a function of the engineered spatial properties of the culture surface. While there are a variety of ways to accomplish this (5), all require extracellular matrix (ECM) proteins and polyaminoacids to mediate the attachment and spreading of anchorage-dependent cell types on an underlying substrate. The most commonly employed strategies for micropatterning involve seeding cells on a surface that exhibit areas (patterns) of differential adhesiveness (8). Patterns generated range from a few microns up to hundreds of microns, depending on the application and size of the cells or tissue to be patterned. This practical guide focuses on micropatterning methods that employ differential adhesiveness exclusively; other approaches, including topographical patterns and directed cell delivery to distinct substrate areas, are reviewed elsewhere (9–11). For each method available, we provide an overview of the technique, a description of applications and questions the method can be used to address, and highlight limitations. In particular, we focus on the scales of the patterns achievable with each method, as this is a fundamental aspect of selecting an appropriate patterning strategy for any particular application. Furthermore, as an example, we highlight in greater detail the steps involved in microcontact printing, which is useful for generating both, microsheets and confluent cell sheets containing patterned co-cultures

Micropatterning steps

All micropatterning techniques involve three basic steps as illustrated in Figure 1: (i) Generation of a pattern of controlled surface adhesiveness, (ii) Seeding of a cell population targeted to adhere to the regions of the substrate defined in step (i), and (iii) Washing to remove excess cells and thus generate the pattern. If a multiphase tissue containing multiple cell types is to be patterned, three additional steps are required: (iv) Treatment to render cell repellent substrate regions adhesive, (v) Seeding of a second cell population targeted to adhere to vacant regions of the substrate that do not contain cell type 1 and (vi) Washing again to remove excess cells, thus generating the patterned co-culture.

Micropatterning at different scales

While all micropatterning approaches share a common goal of achieving a controlled spatial architecture within an engineered system, this can be accomplished on different scales, ranging from single cell patterns to tissue level patterns (Figure 2) depending on the technique used and the size of the adhesive regions on the substrate.

Micropatterns at subcellular to single cell resolution (approx. 5-40µm) limit the extent to which cells can spread on the substrate, allowing users to observe the effects of geometrical constraints on single cell phenotype and behavior. A critical requirement here is that the adhesive areas must be less than the optimal spreading area of the particular patterned cell. By limiting the attachment space available to the cell to below its optimal spreading area, cells can only maximize spreading by adjusting cytoskeletal organization and overall shape to replicate that of the patterned island. This results in square, circular or arrow/tear-drop - shaped cells (12). Typically, during such experiments, sub-cellular sized islands are patterned into single-cell array designs that allow many single cells to be assessed within one experiment to enable quantitative analysis of cellular response. This approach has proven pivotal in demonstrating the importance of cell shape and cytoskeletal geometry on force distribution and mechanotransduction (13–15), migratory properties (16, 17) and polarization (17, 18), as well as differentiation potential of individual cells (19, 20). Moreover, maintaining separation between single cells on micropatterned substrates in culture has been used to examine variations in gene expression amongst individual cells (21), or study the production of signalling molecules in immune cells in response to different patterns of attachment molecules (22, 23).

At the multi-cellular scale, micropatterning has been used to pattern single populations of cells into islands with specific dimensions (e.g. islands of diameter 50-200 µm) generating cellular microsheets (24). Here, the geometry of the tissue sheet is specifically controlled, but the shape of the individual cells within the sheet is not. In some instances, the constraints of the tissue geometry will induce specific changes in the shapes of the cells within the sheet. For example, if microsheets are generated on narrow adhesive islands, cells within the sheet will align in a parallel fashion to the direction of the line pattern to enable maximum packing of the cells within the microsheet (25). Patterned microsheets can be generated on solid substrates as well as filter inserts that enable apical-basal polarization using air-liquid or liquid-liquid culture (26, 27). As with single cell patterning, microsheets are usually generated in arrays using controlled spacing to enable a quantitative, statistically relevant analysis of cellular responses. This approach has been used to study the functionality of collective groups of cells where the microtissue defines a simplified connected network. For example, geometrically confined neuronal networks have been used to mimic in vivo architectures (28) and explore, among others, synaptic function (29), and neuronal polarization (30). Similarly, cardiac cells have also been successfully guided by micropatterning approaches (31–33) to generate microtissues with aligned organizations that more accurately reproduce the native architecture of the heart. The same approach has also be adopted for dissecting the impact of paracrine signaling on stem cell colony differentiation and growth by enabling the generation of colonies with specified diameters and spacing (20, 34–38) as well as studying the collective movement of defined groups of cells (24, 39–42).

Finally, at the multicellular scale, micropatterning has been used to generate patterned tissues (confluent cell sheets) containing micro-scale patterns of multiple cell populations/types. Here, the pattern size exceeds that of an individual cell and facilitates connectivity in all directions within the macroscopic cell sheet. The user controls the architecture of the interfaces between different cell populations, but control over cell shape within the sheets is relinquished. This approach builds upon the microsheet patterning approach detailed above (Figure 1); in the first step, a primary cell population is patterned by generating adhesive patterns on a non-adhesive substrate, which can extend between 50µm to hundreds of µm. In a second step, the entire surface is rendered adhesive, thus allowing the attachment of a second population of cells in unoccupied areas. The ability to switch specified regions of the substrate from non-adhesive to adhesive on demand is critical when generating multi-cell population patterns (see below for the details of how this is typically achieved). Since this approach provides control over the intimate juxtaposition of different cell types within the sheet, it enables the generation of architectures reminiscent of tissue structures observed in vivo, thus facilitating the investigation of key physiological processes over time using traditional or high-content microscopy. Recently, this approach has been used to probe signaling mechanisms at the tumor-stromal interface (43), and examine the establishment of tissue boundaries (S Javaherian et al., manuscript in preparation). Importantly, the stability of patterns generated using this method depends on the characteristics of the cell populations being patterned and their interactions, and pattern remodeling and degradation over time, due to cell re-organization within the sheet, is possible (44).

Generating patterns of controlled surface adhesiveness

Lithography-based approaches

Practically all micropatterning approaches that rely on differential adhesion make use of photolithography or soft lithography to fabricate a stamp, or stencil, that enables the user to generate controlled patterns of adhesion on a culture surface (Figure 3). During photolithography, a silicon wafer is spin-coated with a layer of uncrosslinked photoresist, the thickness of which ultimately defines the depth of the features in the stamp or stencil. Photoresist coating is hardened through baking, and geometric features are defined in the photoresist by UV illumination through a mask containing regions (patterns) that either block or enable passage of the UV light. Mask features can be designed with software, and the type of mask (and cost) depends on feature size. Exposure to UV light renders unmasked areas of the photoresist either soluble or insoluble to the subsequent development process (depending on the type of substrate), thereby ultimately revealing the desired pattern on the wafer surface. This technique has a resolution of 1-2µm, and requires a clean room environment for fabrication (28).

Soft lithography is used to generate an elastomeric polymer stamp containing features defined by the topography of a previously patterned wafer (45). In this case, the process becomes more convenient as no clean room is required and multiple stamps can be cast from the same original wafer (28). Different siloxanes and alternative materials can be used for stamp generation, but poly(dimethylsiloxane) (PDMS) is most often the elastomer of choice, as it can be easily separated from the wafer template later in the fabrication process, is inert when in contact with numerous chemicals (3) and adsorbs proteins, which is advantageous for microstamping. Such soft lithography procedures enable the production of microstamps for microcontact printing, microstencils and polymer channels for microfluidic patterning.

In the case of microcontact printing, a microstamp (Figure 3) is generated by casting liquid PDMS (un-crosslinked) onto the wafer generated by photolithography, crosslinking it by baking at temperatures between 50-100 °C (depending on the application), removing the now solid polymer replica that contains topographic features, and then oxygen/air plasma-treating it to aid ink adsorption to its surface. Lastly, the stamp is coated with a protein solution (the ‘ink'). At this point, differential surface adhesion patterns can be generated by transferring the ‘ink' from the patterned surface of the PDMS stamp onto a cell-culture surface. This transfer occurs upon intimate contact between the stamp features and the culture surface. A variety of “inks” have been used to generate patterns of differential adhesiveness on culture surfaces, and choice of ink depends on the experimental requirements. Certain molecules are characterized by a specific packing behavior that results in the formation of self-assembled monolayers (SAMs) (8, 46) when stamped onto metallic surfaces such as gold and platinum. By functionalizing the molecules used to form the SAMs, a user can generate patterned surfaces that facilitate protein physisorption and cell adhesion (47). Of the synthetic inks that resist bioadhesion, polyethylene glycol (PEG) is one of the most effective and most commonly used (28, 48) to “backfill” patterns. Organic inks include ECM proteins (collagen and fibronectin primarily, but also laminin) or cell binding functional groups (such as RGD), and can be directly used for patterning or to functionalize other inks (28, 49). More sophisticated methods also exist, such as thermally responsive polymers (PNIPAAm) that switch from adhesive to non-adhesive (50), and materials that can be stimulated by voltage or light (51). It should be noted that the type of ink is important in determining the stability of the patterns over time.

In microfluidic patterning (Figure 3), patterns generated on the wafer during photolithography result in specific micro-channel networks that can be used to flow and deposit protein solution in the forms of a desired pattern. As with microstamps, a block of elastomeric polymer (such as PDMS) containing channels is molded from a wafer template, crosslinked and removed from the template. Channels are then clamped to the substrate (usually glass), and a series of solutions perfused through the channels to generate substrate patterning. Channels are then removed before cell seeding. In contrast to microcontact printing which divides the culture surface into two regions (adhesive and non-adhesive), microfluidics allows creation of a multi-component surface by choosing different solutions for distinct channels. Additionally, in this patterning strategy, cells can be deposited along with the solution, rather than being added after a protein pattern is generated. Notably, obtaining surfaces with more than two adhesive components (and thus more than 2 cell types to form the pattern) is also possible using microcontact printing; however, this requires either the use of multi-leveled stamps (52), which are very challenging to produce, or sequential patterning of adhesive components onto the same surface, which requires sophisticated manual or automated alignment for accurate superimposition of features reaching accuracies of 10µm and 1µm (53), respectively.

In addition to microstamp and microchannel approaches, microstencil-based approaches (Figure 3) can also be used to generate patterns of differential adhesion (5). Here, an elastomeric PDMS sheet with through-holes of defined geometries is generated and placed on the substrate during the first seeding step. Once cells have adhered, the PDMS sheet is manually removed by peeling from the culture surface, thus leaving only a pattern of the first cell type attached to the underlying substrate. If co-culture is desired, a second cell type can then be seeded. This method is effective with virtually any adherent cell type, with the notable exception of cell populations that are more adhesive to each other than to a patterning surface. A major limitation of this technique, however, is that generating stencils with small feature sizes using soft lithography can be a challenge since the height of any features on the master (and hence the thickness of the stencil) is limited to ∼2-3 times the feature width, and feature damage can occur during the soft-lithography molding process. Stencils with micron-scale features are therefore very thin with properties similar to cling-wrap, which can cause issues when handling. One variation of this method is to mold a hydrogel around a microstamp to generate a gel stencil in situ. This is effective an effective strategy for short-term patterning and when co-culture is not desired (54) or if patterning on porous filters (55).

Non-lithography based-approaches

All of the micropatterning strategies described above require generation of the initial patterned wafer using photolithography in a clean room, a situation which is not always possible for investigators. Furthermore, due to the largely manual and complex nature of the patterning process, these methods have not been adapted for use with 96-well plates in many cases, which hinders high-throughput quantification. To combat these issues, our lab recently developed two extremely inexpensive and easily implemented techniques based on the use of parafilm inserts (26) and liquid tilting (56), as shown in Figure 4. The parafilm insert method is a stencil-based approach that avoids the challenges associated with soft-lithography based stencils by generating parafilm inserts that are manually cut and pressure-stuck to a tissue culture surface, ‘shielding' the underlying area from cell attachment. This approach is extremely straight forward, requiring little or no prior micropatterning experience and yielding large-scale patterns with features in the order of 100µm and more. While this technique is obviously limited to the manual production of patterns with simple geometries (circles, squares etc.), for a number of applications this is sufficient.

The liquid tilting method is a stamp-free approach that generates surfaces with patterns of differential adhesion by tilting volumes of surface blocking solution in well plates. This method allows the generation of a very simple interface between two cell populations in a multi-well plate format, without the need for robotic instrumentation. Again, despite the limitations in pattern geometry, some applications do not require more complex architectures and can instead leverage the throughput of this system.

Selecting a micropatterning method

A critical aspect in micropatterning is the choice of technique for a proposed experiment. Table 1 and 2 summarize some critical considerations when making this decision, along with several options for each stage of the decision making process. Specifically, the single most important point to consider is the scale and complexity of the interactions of interest. Other factors of major importance are the infrastructures available (e.g. access to a clean room, time for optimization), the throughput required, and the complexity of the geometry. Table 2 provides guidelines for selecting a method based on the specific requirements of the user.

Table 1.  Key advantages and disadvantages of different micropatterning strategies.

Table 2.  Summary and decision map of micropatterning options.

Direct ECM microcontact printing for patterning co-culture sheets

Inking the stamp for pattern generation

Critical steps: Once PDMS stamps have been generated, they must be coated in protein solution. This can be accomplished by washing in 100% ethanol, air-drying the stamp surface in the biosafety cabinet and depositing a droplet of the solution on the features of surface of the stamp. Note that droplet volume needs to be adjusted according to the size of the stamp surface. In our experience, coatings should be left in this position for ∼ 1 hour for the protein to deposit onto the stamp. The droplet is then wicked away by pressing a Kimwipe to the edge of the coated stamp surface, which is then washed in PBS and dried again by wicking the liquid. It is of key importance to the success of the patterning process to ensure that the stamp surface be completely dry before inverting the stamp onto the culture area, as any leftover fluid will distort the pattern and may result in the deposition of salt crystal structures onto the tissue culture surface.

Tips for success

• With repeated use of the PDMS stamps, it will become apparent that the stamping quality decreases; this is due to both wear of stamp features and accumulation of impurities on the surface. The latter issue can be resolved by sonicating PDMS stamps in ethanol and storing in purified water.

• Purified and sterile-filtered water may also be used to wash the coated stamp surface as opposed to PBS, which can sometimes create quite extensive salt crystal structures on the tissue culture surface that may impede uniform cell attachment and lead to pattern disruption.

• The coated stamp can be dried off by applying pressurized air flow as well as wicking off of the liquid using a Kimwipe or equivalent.

Ink transfer

Critical steps: In the most straightforward scenario, the molecular layer of ink can be stamp-deposited onto the tissue culture surface by physisorption. This has been reported to be the case for a large variety of adhesion molecules, including both ECM proteins and synthetic peptides equipped with ECM binding sites. For example, fibronectin has been used to successfully pattern various cell types, including endothelial cells, while laminin has proven to be particularly useful in neuronal patterning (30). The resolution of ink patterns that can be transferred is dependent on the physical properties of the elastomeric stamp, which needs to be both soft enough to uniformly contact the surface to be stamped, yet sturdy enough to maintain the geometry of the pattern features to be transferred (hence the common choice of PDMS). More specifically, features must be spaced sufficiently close together to avoid sagging of the stamp between the features and unintentional stamp contact in undesirable locations (the smallest reported feature size transferrable with PDMS is approx. 1 µm) (57, 58). Usually, the efficiency of ink transfer from the stamp to the culture surface is below 100% and large inter-sample variations can be observed. An approximation to near-perfect (i.e. 90% of the surface has robust patterning) ink transfer can be achieved through much practice, keeping the timing of inking and transfer exactly the same, and the use of the same amount of pressure during each ink transfer. This is crucial for obtaining uniform ligand presentation on the patterned surface.

Tips for success

• In our experience, patterned culture surfaces do not need to be utilized immediately; pre-patterning is possible, because a large percentage of the stamped molecules do remain functional over a period of time (usually weeks in our experience).

• During stamping, in order to promote uniform contact between the stamp and the tissue culture surface, small weights (e.g. 10 g) may be placed onto the inverted stamp to ensure constant pressure over time and avoid stamp detachment.

Inactivation of substrate area outside adhesive regions

Critical steps: Ink patterns can be directly used for cellular attachment whenever it consists of cell adhesive moieties. In some cases patterns may require additional functionalization via adsorption of other molecular moieties, the use of biotin-avidin or antibody-antigen interactions, or chemical vapor deposition (3). The difference in adhesiveness created between the patterned and the non-patterned areas manifests as a difference in the speed of cellular attachment in the different substrate regions (i.e. cells will adhere to the pattern much quicker that to the surrounding areas). It is however important to note that i) as microcontact printing is often carried out on tissue culture polystyrene, which is designed to favor cell attachment, the difference in adhesiveness between surface regions may need to be enlarged by coating the non-patterned surface areas with a passivating agent, commonly BSA; and ii) the serum present in tissue culture media contains adhesive proteins which rapidly adsorb to the non-patterned areas, thereby abrogating all differences in adhesiveness. Accordingly, seeding of the first cell population is most effective if performed in serum-free medium, particularly when no passivating agent is used. BSA is typically an effective passivating agent for ‘back-filling' the pattern, despite reportedly only retaining its blocking properties over a very short period (in our experience, hours). So, while it is possible to microstamp and pattern tissue culture polystyrene in advance of the actual experiment to be conducted to split the lengthy micropatterning protocol, we recommend using the pattered surfaces straight away, once the BSA passivation has been performed.

Tips for success

• There is no consensus in the literature as to what concentration of BSA is ideal for blocking purposes and it is recommended to experiment with BSA concentrations that span at least a 10-fold change, in order to determine the most suitable one for the particular cell type in use. Note that patternability can be affected substantially by this choice and, in the particular case of the plate tilting method (56) described earlier, different BSA concentrations result in different pattern curvatures within the well.

Seeding of a first cell population

Critical steps: Irrespective of the scale of patterns to be achieved (single cells, microsheets or patterned co-culture sheets), the cell density for seeding should be selected to obtain confluency over the entire surface area of the well or dish containing the patterns. For example, ARPE19 or Beas2B epithelial cells are seeded on patterns within a 24 well plate at a density of 1×10⁶, which results in a surface completely occupied by not-yet attached cells, with close to no gaps. Cells are left to adhere to the patterned surface for a period of time that is determined by the specific cell type's particular speed of attachment. This incubation step commonly lies within 30 min to 1 hour, but some optimization is usually required to pinpoint it accurately. Subsequently, non-adherent cells (i.e. cells that happened to be located on non-patterned/passivated areas of the culture surface) are washed away using serum free media. At this point, the desired pattern should be clearly visible using any light microscope, and any misplaced cells can be removed by repeated washing.

Tips for success

• Determining the appropriate cell density for seeding can be challenging in cases where there is a dramatic change in aspect ratios of cells between their attached and suspension states. For example, ARPE19 or Beas2B epithelial cells occupy roughly the same space when attached compared to when detached; in contrast, fibroblasts have a high aspect ratio when spread on a surface, but not in suspension. A small number of trials are usually needed to establish the correct cell number for seeding, so that cells will create a monolayer (but will not be over-confluent) after patterning.

• Patterns such as stripes of cells are best washed by tilting the plate and washing in the direction of the stripes; this will prevent pattern disruption or lift-off.

• It is important to wash patterns with pre-warmed medium so as not to shock the cells and produce lift-off from the surface.

Re-activation of substrate adhesiveness

Critical Steps: For co-cultures, a second cell population must be added, so as to fill the gaps between patterned areas. Accordingly, the passivated surface has to be reactivated so as to allow attachment of this second cell population. ECM protein in complete medium is added to the dish or well containing the pattern, resulting in a switch of the BSA passivated regions of the substrate from non-adhesive to adhesive. When a second cell population is then added, these cell-free, now adhesive regions will favors cell attachment. It is important to allow sufficient time to enable protein transfer onto the substrate surface or cell adhesion will be inconsistent.

Tips for success

• While it is important to ensure that sufficient time is allowed for surface re-activation, the first cell population should not be left in ECM protein and complete medium for a prolonged amount of time (e.g. overnight), as cells will spread and migrate beyond their designated pattern areas. This will lead to blurry pattern boundaries after seeding of the second cell type.

Seeding of a second cell population

Critical Steps: As is the case for the first seeding step, the density of the second cell population must be chosen such that the entire surface area of the well or dish (including the regions covered by the first cell type) is occupied with cells. In this step, the time required for attachment is cell-specific and must be optimized. Specifically, the cells must be left long enough to adhere to the cell-free regions of the substrate, but not too long that the second cell population adheres to the first cell population to generate multilayered regions. After this last washing step, a patterned co-culture sheet of organized cell populations is revealed.

Tips for success

• By placing the tissue culture well or dish under the light microscope before and during the washing procedure, misplaced cells (which usually appear more translucent and rounded up) lying on top of more opaque, spread cell underneath, can be identified and removed by additional, gentle washing.

• Usually, some number of cells remain misplaced, even after the washing step; this may however not constitute a problem, as certain cell types spontaneously sort and move into regions of the sheet containing similar cells.

In our experience, some cell combinations cannot be used for microcontact printing or micropatterning in general. This is largely due to their high adhesiveness to each other, which exceeds their adhesiveness to the substrate, thereby rendering washing of misplaced cells of the second population very difficult, if not unattainable. In these cases, gentle washing will not successfully remove the misplaced cells, and more vigorous washes will disrupt the original cell pattern considerably. In the eventuality that only the second cell population exhibits preferential adhesion to other cells compared to the substrate, but not the first, a reverse seeding order can be attempted. Alternatively, cell printing technologies that directly deposit cells in appropriate locations may provide options for patterning such cell combinations.

Challenges for the future

High-throughput quantification

While conventional micropatterning approaches are widely utilized and yield sharp patterns of cellular attachment, these manual processes often only reach medium-throughput at best and novel approaches are directing the field towards high-throughput data acquisition. For example, more subtle control over the spatial distribution of cellular cytoskeleton has recently been achieved by means of micromagnetic patterns in conjunction with magnetic nanoparticles for high-throughput single cell patterning (59). Methods including laser assisted bio-printing (60) and stereomask lithography (61) address this need further while also implementing modulation of multiple cell types within the same array. These approaches have been instrumental in quantifying single cell behavior but do not allow the visualization of the interactions between multiple cells; this can be achieved by extending the areas of attachment to supra-cellular sizes. Micropatterning substrates using robotic technology (4) or high throughput microfluidics are also an option to increase throughput in the future.

Dynamic patterning

Emerging in the field of micropatterning is the concept of selective pattern modification during, and not just before, cell attachment, in order to change patterns of adhesiveness and shape over time. Different approaches can result in these evolving surfaces, such as single cell arrays in which additional space is made available by UV exposure (62), electrochemical stimulation to alter the presentation of ligands available for cell migration (63, 64), or the addition of clickable biotin ligands (4, 65). This modulation of cell attachment space provides control over ligand-receptor interactions, and scalable production of arrays for screening.

Three-dimensional tissues

Within a confluent sheet, micropatterning approaches are successful at imparting architecture to a monolayer of cells. In tissues thicker than a monolayer, we typically observe that any defined spatial constraints are rapidly lost, and the initially defined cellular pattern on the culture surface does not translate to controlled organization in multiple layers of cells. This is analogous to what has been shown for topographical surface cues for cellular organization and indicates that cells require direct contact with the patterning cue in vitro and may suggest that multi-layered tissue architecture in vivo is maintained by ECM cues throughout the tissue structure, as well as soluble factors. The 2D nature of micropatterned cellular sheets represents a major limitation of this approach for tissue engineering, where the assembly of thick multi-component cell constructs is desired. Hydrogel strategies to limit cell spreading within a 3D context are emerging (66, 67). Complex patterned tissues are being generated using sequential layering approaches (e.g. printing), whereby the architecture of each layer is controlled and layers are stacked (68). Also, micro-gels that are guided by magnetic forces to assemble much like Tetris blocks have recently been introduced, in order to achieve higher-order multi-component tissues (3). Competing cues at different scales will likely also add complexity in the 3D situation, as has been observed in 2D (69). Overall, the need for architecturally controlled multilayered tissues represents a great challenge for tissue patterning and architectural engineering in the future.