Design and bioprinting for tissue interfaces

Tissue interfaces include complex gradient structures formed by transitioning of biochemical and mechanical properties in micro-scale. This characteristic allows the communication and synchronistic functioning of two adjacent but distinct tissues. It is particularly challenging to restore the function of these complex structures by transplantation of scaffolds exclusively produced by conventional tissue engineering methods. Three-dimensional (3D) bioprinting technology has opened an unprecedented approach for precise and graded patterning of chemical, biological and mechanical cues in a single construct mimicking natural tissue interfaces. This paper reviews and highlights biochemical and biomechanical design for 3D bioprinting of various tissue interfaces, including cartilage-bone, muscle-tendon, tendon/ligament-bone, skin, and neuro-vascular/muscular interfaces. Future directions and translational challenges are also provided at the end of the paper.


Introduction
Tissue interfaces have a critical role in dynamic interaction between two distinct tissues. They have unique structures organized in a highly ordered matrix with spatially graded biochemical cues and cell types Jiang 2005, Thomopoulos et al 2012). However, these structures have limited self-repair capacity and any damage at a tissue interface may not be fully recovered Jiang 2005, Greenspoon et al 2015). The transplantation of auto-, allo-, or synthetic grafts is currently in clinical use for the treatment of the tissue interfaces. However, healing of the damaged tissue interfaces with those grafts is unsatisfactory primarily due to the lack of the gradient structure formation in a shorter time (Betz 2002, Jia and Sun 2015, Pujji et al 2017. Injectable hydrogels or cell-sheets have been used as a non-invasive approach, but it is challenging to restore such complicated structures with required mechanical and biological performances similar to native interfaces (Liu et al 2017, Tu et al 2019, Kim et al 2019b, Xu et al 2020. Several investigations were conducted to determine the role of each biological, physical, and mechanical factor to restore the function of tissue interfaces with appropriate biochemical and biomechanical designs for scaffolds (McConnell et al 2017, Calejo et al 2020. Biofabrication of multi-layered tissue scaffolds is one of the appealing ideas to mimic the physicochemical properties of their native counterparts by the way of setting the organization and composition of the matrix with multiple biomaterials and cell types (Kang et al 2018, Liverani and Boccaccini 2018, Costello et al 2019, Fuchs et al 2019. However, fabricating multiple scaffolds separately and integrating them in the damaged interface area cannot be sufficient to repair the tissue interface function due to the lack of cell-to-cell signaling between independent scaffolds.
Three-dimensional (3D) printing has opened up a new perspective and offers an excellent opportunity for interface tissue engineering (Lee and Wu 2012, Bose et al 2018, Moroni et al 2018, Calejo et al 2020, Ramos and Moroni 2020. It is a pioneering method in which bioengineered constructs are fabricated in a precise manner by layer-by-layer deposition of two or more biomaterials spatiotemporally with the aid of computer-aided design in a tunable and hierarchical architecture. The process aims to fabricate constructs with the structural and compositional heterogeneity, and the functionality of host tissue or organs (Murphy andAtala 2014, Hafezi et al 2015). To tailor toward biomimicry, a more recent approach called '3D bioprinting' which uses biomaterials, cells and bioactive molecules for the printing, is proposed (Groll et al 2018. Considering the aforementioned complex structure of tissue interfaces, bioprinting could be a promising technique to tackle the challenges of biofabrication of intricate and multifunctional tissue interface structures .
A systematic biomaterial design with a suitable bioprinting process is necessary to achieve the biomimicry of tissue interfaces. This review article covers the construction of tissue interfaces with specific biomaterial design using 3D bioprinting strategy. First, 3D bioprinting methods used for interface tissue engineering are introduced. Tissue interfaces including cartilage-bone, muscle-tendon, tendon/ligamentbone, skin, and neuro-vascular/muscular interfaces are described to enlighten a perspective for design of the structures and their bioprinting. Finally, perspectives on interface tissue bioprinting are discussed to shed light on the requirements and challenges for the future direction of the treatment of tissue interfaces.

Inkjet-based bioprinting
The first generations of inkjet bioprinters were based on modified commercial printers traditionally used for ink printing on papers. The inks and papers were replaced by biomaterials, cells and the substrate (Singh et al 2010, Murphy andAtala 2014). Inkjet printing, also referred to as drop-on-demand or drop-by-drop printing, is one of the most common bioprinting techniques. Schematic representation for working principle of the ink-jet bioprinting is given in figure 1(A). Basically, nanoliter liquid droplets of biomaterials with or without cells are deposited on a moving substrate with a high precision (Derby 2008). The final diameter of drops, and consequently, the resolution of the printed structure are influenced by the physical properties of the ink, such as viscosity, surface tension, or process-dependent parameters like nozzle diameter, feed rate, and layer-thickness. Controlled positioning of the jet on the substrate requires a compromise between applying force and the distance between the nozzle and substrate (Singh et al 2010. Since different bioinks, including multiple cells and biochemical materials, could be simultaneously loaded and spatially deposited during the process, it has a high potential to be utilized for complex and multifunctional and gradient structures like tissue interfaces (Gurkan et al 2014, Daly andKelly 2019). This bioprinting technique has been used for the fabrication of multiple tissue interfaces such as osteochondral (Daly and Kelly 2019), muscletendon (Phillippi et al 2008, Laternser et al 2018, tendon-bone (Ker et al 2011), and skin (Min et al 2018, Kim et al 2019a. Researchers also took advantage of inkjet bioprinting techniques individually or as a hybrid setup together with other bioprinting methods for the construction of tissue interfaces (Laternser et al 2018, Daly andKelly 2019). The main drawback of inkjet printing is that only low viscosity bioinks could be used by this method. Although some efforts have been made to address this limitation by postcrosslinking of the bioinks after deposition, the delay in crosslinking of hydrogels at each layer is a major concern in the bioprinting of large structures (Xu et al 2013. Moreover, problems regarding the clogging of the nozzle limit the cell density of bioinks to be utilized in this approach (Pedde et al 2017).

Stereolithography (SLA)
This technique is based on light-irradiation (mainly in the ultraviolet range) on photo-cross-linkable polymer liquids by user-defined patterns. Each layer is formed by the projection of a computer-controlled light pattern onto the surface of a collector on the z-axis as schematically represented in figure 1(B) (Melchels et al 2010, Knowlton et al 2015. Digital light projection (DLP) is the most commonly used method that can cure a whole layer of a construct at once in a short period of time (Lu et al 2006, Melchels et al 2010. Two-photon polymerization is used as another approach which facilitates the construction of fine structures at a sub-micron resolution (Nguyen and Narayan 2017). Compared to inkjet bioprinting, encapsulated cells in SLA-based techniques are not exposed to the shear stress of the nozzle, and high cell densities can be used. In addition, DLP based SLA bioprinting typically results in high cell viability. Moreover, the crosslinking density of each layer can be changed by the amount of radiation energy and the time of exposure (Pedde et al 2017). Hence, the layer-by-layer fabrication principle of SLA with the capability of micro-scale transition both in size and the biomaterial composition at each layer offers an excellent opportunity to construct graded structures for tissue interfaces with high spatial precision. For example, it has been utilized in biofabrication of osteochondral tissue interfaces and tendon-bone interfaces (Gurkan et al 2014, Aisenbrey et al 2018, Zhou et al 2019. Despite the advantages mentioned above, high energy radiation and the toxicity of some photoinitiators used within the hydrogel may damage the cells (Pedde et al 2017).

Laser-assisted bioprinting (LAB)
LAB is a non-contact additive patterning process based on a laser-induced forward transfer (Delaporte and Alloncle 2016, Pedde et al 2017). It has been utilized with bioinks of different biological compositions, such as biopolymers, DNA, peptides, and nanotubes (Delaporte and Alloncle 2016, Serra and Piqué 2019). Compared to the other bioprinting strategies, this technique is less common; however, a wide range of materials could be used with a high resolution-deposition in a micro-scale (Murphy and Atala 2014, Delaporte and Alloncle 2016). LAB does not possess a nozzle in the process as represented in figure 1(C), which eliminating the problems associated with clogging and high shear stress on the nozzle tip. Therefore, the constructs with higher cell densities are achieved with relatively high cell viability (more than 95%) (Hopp et al 2005, Malda et al 2013, Knowlton et al 2015, Ozbolat and Hospodiuk 2016. The ability to bioprint a wide range of bioinks with different viscosities, multiple cell types, and biomaterial compositions allows the construction of multiphase structures of tissue interfaces. It was used for the fabrication of skin tissue interface (Koch et al 2012, Murphy and Atala 2014, Pedde et al 2017. The costly process with complex hardware, the dependency of the structural resolution on the process parameters, the requirement of fast gelation kinetics of polymers for high configurational resolution, and the probability of existing metallic residual absorbents have limited the widespread applications of this method.

Extrusion bioprinting (EB)
EB is currently the most promising and commonly utilized bioprinting methodology. In general, extrusion printing refers to a continuous flow of material from a reservoir through a nozzle mounted on a computer-controlled collector in a layer-by-layer fashion as schematically represented in figure 1(D).
Material feeding mechanisms are based on either pneumatic or mechanical (piston or screw-based) systems (Placone andEngler 2018, Luo et al 2019). Spreading of extruded filaments and subsequent negative effects on shape fidelity are challenging for EB, requiring high viscosity or fast gelation kinetics of the extruded inks to fabricate structures with acceptable geometrical precision (Moroni et al 2018). Novel methods have been developed to overcome the limitations arising from high viscosity requirements, including utilization of sacrificial supporting mediums, co-axial nozzle extrusion systems, and implementation of pre-or post-crosslinking strategies for soft hydrogels (Malda et al 2013, Hinton et al 2015, Jin et al 2017, Afghah et al 2020. EB can be used with multi-head and/or multi-nozzle designs to fabricate seamless multifunctional structures (Datta et al 2017, Zhang et al 2017). The capability to switch the dispensing systems of multi-head printers individually, including the myriad of biomaterials from soft hydrogels to stiff polymers in a controlled manner, enables the fabrication of heterogeneous structures mimicking natural hierarchical structures like tissue interfaces (Merceron et al 2015, Ahangar et al 2019. As one of the most affordable and simplest bioprinting techniques, EB has been used for the fabrication of various types of tissue interfaces including osteochondral (Zhou et al 2019, Kilian et al 2020, muscle-tendon (Merceron et al 2015), tendon-bone (Park et al 2018), skin (Baltazar et al 2020), and neurovascular unit (NVU) (Han et al 2019) interfaces. More examples of the utilization EB for the construction of various tissue interfaces are given under subsections below. It is important to note that printing temperature, viscosity, rheological properties, cross-linking mechanism, and the shear-thinning/thickening behavior of inks dictate the suitability of cell encapsulation and printability via the EB method (Murphy and Atala 2014). Low resolution and low printing speed are the most common drawbacks of extrusion-based bioprinting, along with low cellular densities due to exposure of cells to high shear forces during deposition through the nozzle (Datta et al 2017, Zhang et al 2017, Ahangar et al 2019.

Direct cell-printing method
Contrary to all the bioprinting strategies described above, the direct cell printing method is a scaffoldfree approach (Moldovan et al 2017, Ozler et al 2017, Ayan et al 2020a, 2020b. Fulfilling several materials requirements such as biocompatibility, high shear stress cells being exposed to, possessing a small window of rheological properties, and postprocessing of biomaterials have led to the development of a bioprinting method with no dependency on biomaterials (Skardal and Atala 2015). In this context, various tissue engineering methods have been developed using cellular secreted extracellular matrix (ECM), which requires a long process (Moldovan 2018). Kenzan method and aspirationassisted freeform bioprinting approaches have been demonstrated for cell spheroids as building blocks and facilitated to overcome the cellular instability and diffusion of directly-bioprinted cells (Moldovan et al 2017, Ozler et al 2017, Ayan et al 2020a, 2020b. Kenzan method was used by the Regenova bioprinting system (Cyfuse Biomedical K. K., Japan). In principle, cell spheroids are aspirated and then skewered on an array of steel microneedles that work as temporal support. These microneedles would assist the diffusion of cells without affecting cell viability and ECM production (Arai et al 2018). This method relies on cell-cell interactions to form and grow spheroids that lead to functional tissue formation, in which the structure will be removed finally (Aguilar et al 2019). Aspiration-assisted freeform bioprinting is based on picking and lifting the spheroids into the air by aspiration forces and placing them onto a sacrificial hydrogel, where the spheroids are incubated until their fuse (Ayan et al 2020a). The control over the selection of spheroids with different cell types has enabled the construction of heterogeneous tissue structures like tubular structures, mini-liver tissue constructs, and osteochondral tissue interfaces (Kizawa et al 2017, Grogan et al 2020, Ayan et al 2020b. However, the picking and placing spheroids one at a time would take a long time (LaBarge et al 2019). Technical issues such as microneedle distance, spheroid size, and cell distribution for heterogeneous spheroids are important for maintaining a balance between cell interactions and their ECM production (Moldovan et al 2017). The direct-cell printing method is suitable for using stem cells and fabricating large tissue constructs.

Cartilage-bone (osteochondral) tissue interface
The osteochondral interface is the transitional region between viscoelastic cartilage and stiff subchondral bone tissue at the joint (Hoemann et al 2012). The smooth transition of mechanical and physicochemical properties between these two tissues plays a critical role in proper load bearing and movement at the joints. Cartilage is an avascular connective tissue comprised of distinct superficial, middle and deep zones (Yang and Temenoff 2009). In the superficial zone, chondrocytes, mostly in flattened shape morphology, are aligned parallel to the articular surface inside thin collagen fibrils. On the other hand, the middle zone has mostly round-shaped chondrocytes, which are not entirely aligned to the articular surface and embedded within larger collagen fibrils. Chondrocytes in the deep zone are aligned perpendicular to the articular surface within the larger-sized collagen fibers. Gradual increase in the proteoglycan content from the superficial zone to the deep zone is another prominent feature of cartilage tissue. The increase in water content is directly proportional to the proteoglycan content of the translation zones Temenoff 2009, Bayrak andYilgor Huri 2018). The cartilage connects to the subchonral bone from deep zone which is connected to the calcified cartilage zone with the extension of collagen fibers in a wavy tidemark (Hoemann et al 2012). Hypertrophic chondrocytes are embedded in collagen type X and mineralized matrix with hydroxyapatite (HAp) in the calcified zone. Collagen fibers provide unique viscoelastic properties to transfer the loads to the underlying subchondral bone and help articulation at the joints. The compressive modulus at each zone changes remarkably from 0.079 MPa at the superficial zone to 5.76 GPa at the subchondral bone. The calcified cartilage zone with a compressive modulus of 320 MPa is not as strong as a subchondral bone even though it is mineralized .
The defect in the osteochondral interface mostly occurs between articular (hyaline) cartilage and subchondral bone, which results in movement disabilities. Chondrocyte implantation, osteochondral grafts, or bone marrow stimulations are currently used for treating these types of defects, but the injuries could re-occur with a high possibility (Roberts et al 2002, Prado et al 2016, Sadlik et al 2017. The inflammatory responses due to the lesion at the cartilage cause damage to the subchondral bone (Li et al 2013). Thus, tissue engineering works on the development of the novel scaffolds induces regeneration of osteochondral interfaces from the superficial cartilage zone to the subchondral bone region. It is difficult to construct such an interface between biologically, physically, and mechanically different cartilage and subchondral bone using conventional tissue engineering methods (Calejo et al 2020). Although each zone have been constructed separately using single scaffolds, they mostly failed to integrate and function synchronically (Seo et al 2014, Calejo et al 2020. The success of the osteochondral interface treatment strongly depends on the regeneration of the calcified cartilage zone. 3D printing technologies have enabled the construction of different scaffolds for osteochondral interface by spatiotemporal positioning of different biomaterials with or without appropriate progenitor cells (Gao et Table 1 summarizes the 3D bioprinted single-phase constructs for the osteochondral interface engineering using different biomaterials and 3D printing technologies. Bioactive ceramics are one of the most promising biomaterials for single-phase scaffolds for the regeneration of osteochondral interfaces by fulfilling the requirements of both chondrogenesis and osteogenesis (figure 2(A)). Extrusion-based printing is often used to construct porous ceramic structures by using homogeneously dispersed powders in sacrificial hydrogels such as Pluronic-F127 and alginate, followed by sintering at high temperatures. The chondrogenic and osteogenic bi-lineage activities of the ceramics are due to the doped a group of elements such as silicon (Si), Lithium (Li), Molybdenum (Mo), Strontium (Sr), Copper (Cu), Calcium (Ca), Manganese (Mn) and Magnesium (Mg).
Biocompatible thermoplastic materials such as polycaprolactone (PCL), poly(lactic acid) (PLA), and poly(lactic-co-glycolic acid) (PLGA) are often employed to recapitulate the natural stiffness and strength of osteochondral interfaces. These types of polymers can be constructed by environmentally and cell-friendly thermally assisted extrusion and melt electrospinning writing approaches ( In single phase constructs, these polymers are usually functionalized with biochemical cues to enhance the cell attachment, and osteogenic or chondrogenic differentiation (Guo et al 2018, Koo et al 2018. The direct cell-printing method also showed favorable results for constructing single-phase scaffolds for osteochondral tissue interfaces ( Rational gradient scaffold designs in terms of architecture, composition, and physiochemical properties are proposed to guide cells to provide specific features required at each zone. These types of multiphasic scaffolds have been also constructed for this purpose to induce both osteogenesis and chondrogenesis for osteochondral tissue interfaces (table 2). The type of these structures can be divided into three subgroups. The first type of structure can be bioprinted by changing the material or cell composition by targeting the layers facing the subchondral bone or cartilage regions ( (2019) 3D printed PCL scaffold coated with self-assembling peptide hydrogel (SAPH) Extrusion printing SAPH coating promotes cell attachment, proliferation, and osteogenic differentiation and maintains chondrocyte phenotype and regeneration.
Li et al (2019) 3D printed PCL microchambers loaded with GelMA encapsulated with MSCs and chondrocytes Extrusion printing and inkjet bioprinting PCL provided mechanical reinforcement to MSCs and chondrocytes-laden GelMA that was printed into microchambers via inkjet bioprinter. Also, it guided their growth and fusion and the development of stratified cartilage tissue with a depth-dependent collagen fiber architecture.
Daly and Kelly (2019) Cartilage-dECM/GelMA/exosome Stereolithography By the use of the exosome of MSCs, cartilage mitochondrial function was restored, and cartilage regeneration was facilitated. Poly(ethylene glycol). Arginine-alanine-leucine-arginine. The third type is bioprinted by the combination of both aformentioned approaches which use different type of bioactive molecules and gradient porous architecture together in the multi-phase structure. For these kinds of structures, more than one bioprinting techniques are usually used at the same time. For example, mechanically strong biocompatible and biodegradable polymers such as PCL and PLA are fabricated in gradient porous architecture with appropriate mechanical stiffness (Bas et al 2018). Then, cell-laden bioinks are infiltrated into a PCL framework by a dual 3D printing platform with multiple-head extrusion or inkjet-based printing approaches simultaneously (Kosik-Kozioł et al 2020).

Muscle-tendon interface
Tendon is a connective tissue that connects muscles to bones . It functions in a synchronized manner with muscle to transfer the applied force to bone as a mechanical bridge. Tendon integrates with muscle at a delicate point which has three distinct units: an elastic muscle that is predominantly made up of multinucleated myofibers, a tendon that mainly consists of collagen fibers with discontinuous fibroblast (FB) cells in a stiff structure, and a myotendinous junction (MTJ); a linkage of the spatial distribution of these biologically and mechanically distinct tissues to each other in a complex structure (Merceron et al 2015. MTJ is responsible for controlling the movements and positions of the joints (Bayrak and Yilgor Huri 2018). The morphology and thickness of MTJ differ depending on age, gender, and body mass index (Ovalie 1987). Paxillin, integrin, vinculin, and talin proteins at the MTJ act as a focal adhesion connecting actin filaments of myofibers at the elastic muscle side to collagen and laminin proteins of the stiff tendon side (Finni 2006, Yang and Temenoff 2009, Merceron et al 2015, Bayrak and Yilgor Huri 2018. The re-establishment of a functional MTJ strongly depends on the formation of aligned myofiber bundles at the muscle side and extensive collagen fibers deposition at the tendon side. 3D printing technology has been utilized with promising strategies to construct the distinct units of MTJ. Most of the studies focused on the formation of organized and aligned muscle cells in the structure aiming to induce myogenic differentiation and the formation of aligned myofiber bundles. The cell orientation has been guided with geometrical (microgrooves, micro/nanopatterned surfaces) or micro-topological polymeric (PLA, PVA) (Kim et al 2018b, 2020b cues included into the bioinks. It was also reported that cell alignment and maturation of muscle fibers can be regulated by mechanical or electrical (graphene) stimulations or treatment with several growth factors such as VEGF, insulin-like growth factor-1, and growth hormone-1.
The high survival rate of muscle cells is important for the myofiber formation (Kang et al 2016, Kim et al 2018a. Thus, microchannels resembling vessels are commonly introduced in the 3D printed structures for the transportation of sufficient nutrients, oxygen, and metabolites. Considering the aforementioned requirements for a functional MTJ, an integrated 3D organ printing system with multiple dispensing heads was utilized to construct a functional MTJ (Merceron et al 2015). Four different ink compositions including thermoplastic polyurethane (PU) and PCL as the scaffolding component, and two different bioinks composed of hyaluronic acid, gelatin, and fibrinogen and individually loaded with C2C12 myoblast and NIH/3T3 FB cells were loaded to different print heads to target muscle and tendon formation. First, PU and PCL were 3D printed by an extrusion printer to support the structure and to guide the alignment of cells and myofiber formation. Then, the bioink composed of C2C12 myoblast was co-localized between the PU fibers aimed to develop the muscle part, while NIH/3T3 containing bioink was deposited between the PCL fibers to target tendon formation in the same layer, as demonstrated in figure 3. After crosslinking of fibrinogen with thrombin, the sacrificial gelatin was removed to provide a porous structure. An integrated structure was formed with aligned C2C12 myoblast cells and the secretion of desmin and collagen type I by NIH/3T3 FB cells at the interface region. The elastic modulus of the muscle-targeted side was measured as 0.39 ± 0.05 MPa, while the tendon-targeted region exhibited a stiff structure with 46.67 ± 2.67 MPa. The successfully integrated construct also showed an intermediate elastic modulus of 1.03 ± 0.14 MPa at the interface.
In another study by Laternser et al, a muscletendon-like structure was bioprinted in a multiwell plate with the aim of high throughput drug screening application for the treatment of muscletendon interfaces (Laternser et al 2018). Two roundshaped dumbbell-like posts were produced by molding polypropylene, which was inserted into the polystyrene-coated multi-well plate. To mimic a tendon-like tissue, GelMA-polyethylene glycol dimethacrylate-based ink and tenocytes were printed around the posts in alternating layers via extrusion and inkjet printing, respectively. Then, GelMA and myocytes were bioprinted in alternating layers in between the posts by extrusion and inkjet printing to produce muscle-like tissue. The muscle cells aligned and rapidly induced myofiber differentiation and tendon formation in the final 3D structure. Moreover, muscle functionality was supported by contracting myofibers via electrical stimulation.
It is worth mentioning that current 3D printing approaches demonstrated that the geometrical, biological, and mechanical cues guide the cells to myofiber differentiation for muscle to tendon interfaces. The attempts are also made for the development of clinically relevant tissue models for drug discovery studies. Further studies are necessary for the development of more effective 3D printing strategies with novel biomaterials to construct functional MTJ structures.

Ligament/tendon-bone interfaces
Ligaments are another connective tissue that connects two bones . Although tendons and ligaments are found in different locations of the bone and exhibit different functions, they have similar molecular compositions. The structure of the tendon-bone and ligament-bone is formed by the gradual transition of four different zones: tendon/ligament, nonmineralized and mineralized fibrocartilage, and bone (Benjamin et al 2006). Each zone has different cell types and ECM components. The tendon/ligament zone consists of interspaced and elongated FB cells with organized collagen fibers having 70%-80% of the dry weight of ECM (Rumian et al 2007). Elastin, proteoglycans, glycoproteins, and water are the other elements of the ECM, but ligaments comprise of higher concentration of elastin than tendons (Yang and Temenoff 2009). The nonmineralized fibrocartilage zone has a high concentration of collagen type II and III, a small amount of collagen type I and X, and proteoglycans. In addition, it has decorin and aggrecan, which provide flexibility at the interface (Smith et al 2012). Mineralized fibrocartilage is composed of collagen type II and X and aggrecan. A highly dense mineral content, about 69%, together with collagen type I, osteoblasts, osteocytes, and osteoclasts are found on the bone side (Bayrak and Yilgor Huri 2018). Bone Figure 3. Schematic representation of 3D printing of MTJ with the integrated organ printing system. Four different dispensing heads were loaded with different inks including thermoplastic polymers of PU and PCL, and C2C12 myoblast and NIH/3T3 fibroblast cells loaded bioinks. First PU and PCL fibers were printed with a 10% overlap area at the interface region. Then, the bioink composed of C2C12 myoblast was deposited within the PU gap region while other bioink with NIH/3T3 fibroblast cells were dispensed between the PCL polymeric patterns. The C2C12 myoblast cells were labeled green fluorescence dye (DiO), and fibroblast cells were labeled red fluorescence (Dil). Confocal microscopy images of (A) the 3D construct and (B) zoomed image for the interface region on day 1. (C) The image of the construct on day 7 shows the cell migration and integration of the structure (Adapted from Merceron et al (2015). © IOP Publishing Ltd. All rights reserved). as a calcified tissue is much tougher than tendon and ligament, with ten times higher elastic modulus (Benjamin et al 2006).
Fabrication of scaffold for ligament/tendon-bone interfaces requires a complex design possessing a delicate soft to hard tissue transition. Coating or embedding the growth factors to the 3D printed structures and their controlled release is one of the approaches to promote regional regeneration by inducing endogenous differentiation of progenitor cells (Gurkan et al 2014, Perikamana et al 2018, Tarafder et al 2019. For example, three different bioink compositions were prepared by loading connective tissue growth factor (CTGF), CTGF and TGFβ3, and BMP-2 into PCL ( figure 4(A)). Before loading, these factors were encapsulated in PLGA. These bioinks were printed layer by layer via an extrusion-based device by targeting tendon, fibrocartilage interface zone, and bone regions (Tarafder et al 2019). The 3D-printed scaffolds with flexible structure (figure 4(Aii), (iv)) successfully was implanted between tendon and bone of the rat rotator cuff (figure 4(Aii)), and promoted regional differentiation of endogenous stem/progenitor cells and enhanced tendon-to-bone interface healing.
An inkjet printer was used to fabricate the tendonbone interface by depositing nanoliter droplets in micro-scale transition in another study (Ker et al 2011, Gurkan et al 2014. Figure 4(B) represents the bioink composition for a non-mineralized and mineralized fibrocartilage tissue interface region. The GelMA hydrogel composed of hMSCs and TGF-β1 was used to construct non-mineralized fibrocartilage, while the GelMA hydrogel with hMSCs and BMP-2 was used for the mineralized fibrocartilage (figure 4(Biii)) (Gurkan et al 2014). In vitro results showed the upregulation of osteogenic and chondrogenic-related gene expressions in the nonmineralized and mineralized fibrocartilage targeting regions of the 3D bioprinted structure.
It is clear that a bioinspired design and systematic and precise bioprinting of the biomaterials could allow the maturation of functional structures to recapitulate the ligament/tendon to bone interface or build functional structures for personalized drug discovery studies.

Skin tissue interface
Skin is the first barrier organ that protects our body from microorganisms, burns, and different light sources. As the body's largest tissue, it comprises three main compartments of the epidermis, dermis, and subcutaneous tissues (Kolarsick et al 2011). The epidermis layer is composed primarily of keratinocytes (KCs), melanocytes (MCs), Langerhans, and T-cells (Velasco et al 2018. The dermis consists of FBs surrounded by collagen and fibronectin in ECM and endothelial and stem cells . These layers differ based on their functions due to the expressed major protein types. The basal membrane separates and connects the epidermis and dermis and provides mechanical support. In addition, the basal membrane provides paracrine and autocrine signals to control the growth and differentiation of KCs. Laminin and collagen IV are the main proteins of the basal membrane forming different polymeric networks. Fibrillin, collagen V and VII are also involved in the formation of organized polymeric structure in the basal membrane of skin tissue (Masunaga et al 1996, Francois et al 2015.
Numerous studies have been performed to bioengineer functional structures to substitute with injured skin by employing bioprinting techniques. The general idea is to control the deposition of FB and KCs in an optimum cell density with maximum viability to mimic physiological structure (Koch et al 2012, Cubo et al 2016, Rimann et al 2016, Kim et al 2017, Pourchet et al 2017, Min et al 2018. The differentiation of KCs into corneocytes which leads to the formation of stratum corneum and the formation of an epidermis-like structure, is commonly monitored to evaluate the functionality of the design construct.
Since skin is a soft tissue, less viscous hydrogels and liquid cell suspensions are mostly utilized to fabricate dermal skin resembling native mechanical properties. Inkjet, droplet-like microextrusion, and LAB strategies or their combination are more appropriate for those kinds of materials and commonly pursued for precise and controlled construction. Table 3 summarizes the utilization of biomaterials, bioprinting strategies, and design for the regeneration of skin tissue interface. The formation of the laminin layer between the fibroblasts and keratinocytes was observed, which presented the structure's functionality.
Koch et al (2012) Human immortalized keratinocytes encapsulated in collagen and a mixture of blood plasma and alginate (Continued.) Layer-by-layer bioprinting of matrix components with cell suspensions where an appropriate crosslinking is applied in each layer is the most common approach used to obtain functional tissue (Koch et al 2012, Rimann et al 2016, Min et al 2018. Initially, dermal component FB cells are bioprinted in alternating layers with hydrogel, and then KCs are seeded on top of the construct to form the epidermis layer. An air-liquid interface culture is utilized to facilitate the differentiation and maturation of KCs. To improve the functionality of the skin tissue substitutes, vascularization and skin pigmentation would be important to be considered (Min et al 2018). For example, an inkjet bioprinter with multiple independent microvalves was used to dispense cell suspensions (MCs, KCs), and hydrogel precursors or growth factors to construct the full thickness skin model. First, a FB dermal equivalent was fabricated as follows: A square-shaped collagen layer with a 12 mm length was fabricated, followed by bioprinting of three layers of collagen rim and a squareshaped FB suspension at a 7 mm dimension. On top of the dermal construct, a complete collagen layer was deposited, and afterward, two layers of collagen rim were fabricated and served as a container for further bioprinting of MCs and KCs to prevent them from being washed away inside the culture medium. Finally, the dermal compartment was cultured in FB media for one day, and the media was then removed. MC suspension was deposited on top of the structure with two different configurations, one in a single spot model with a 2 mm diameter located at the center and an area model in a square-shaped pattern with a 6 mm length. The structures were cultured for four days in MC culture media. The epidermal layer was finalized by depositing KCs in the middle of the collagen rim and cultured for one day in KC media. To induce the differentiation of KCs, the structure was kept in an air-liquid interface containing KC stratification media and cultured for four days. Their results showed no contraction after nine days of culture. Visual investigations revealed light pigmentation on day 4 of incubation and dark pigmentation after airliquid interface culture in the area model, indicating the interaction between MCs and KCs and their proliferation and differentiation. Histological data also showed the signs of pigmentation and melanin granules inside the epidermis layer for both MC area and spot models and at the epidermal-dermal junction.
In conclusion, their skin equivalent showed a distinct dermal and epidermal layer with a basal layer at the interface, and no sign of contraction was observed.
Vascularization between the dermis and hypodermis compartments is an important factor for communication with the epidermis, which helps the maturation of the epidermis. Introducing the microvessel-forming cells to the construct can assist in the formation of perfusable microchannels (Kim et al 2019a, Baltazar et al 2020. It is a laborious and complex fabrication procedure to mature the distinct compartments of skin tissue with vessels. For example, Kim and his coworkers modified their printing setup to develop a perfusable vascularized complex skin construct comprising not only the epidermis and dermis but also hypodermis components representing a complex skin substitute. The developed skin construct showed a better resemblance to native skin tissue compared to mere dermal and epidermal substitutes (Kim et al 2019a). Their hybrid hypoderm and dermal bioinks were composed of adipose-derived decellularized ECM (dECM) mixed with fibrinogen, and skin-derived dECM with fibrinogen, respectively. Gelatin-based bioink was used to bioengineer vascular channels inside their skin model. The printing process was as follows: first, a transwell system was fabricated via extrusion printing of PCL struts followed by extruding sacrificial gelatin hydrogel in between the PCL pores to prevent the diffusion of further 3D printed bioinks. Then, they printed PCL mesh and created a perfusion system that was covered by an adipose-derived dECM bioink embedded with human adipocytes and pre-adipocytes for the hypodermis layer. Vascular channels were then created on top of the hypodermis layer with human umbilical vein endothelial cells and closed by the extrusion of a PCL layer on top of them. The dermal layer was extruded using human dermal fibrablasts (HDF)encapsulated skin-dECM with fibrinogen. After that, the hydrogels were crosslinked to achieve an intact construct. Following this step, the structure was cultured for seven days in FB growth and pre-adipocyte differentiation media while the channel was supplied with endothelial growth medium with the aid of a peristaltic pump. Subsequently, the epidermal compartment was created via inkjet bioprinting of human epidermal keratinocytes. A schematic representation of their fabrication process is illustrated in figure 5. They created a full-thickness of skin substitute that PCL extruded mesh could overcome the hydrogel contraction during incubation. The custommade PCL transwell system with the desired size also facilitated feeding multiple mediums to resemble more realistic in vivo conditions. It consisted of four sublayers of stratum corneum, stratum granulosum, stratum spinosum, and stratum basale, with a cell density of dermal layer almost close to human skin.
Considering the complex and multimaterial skin tissue comprising several layers and cell types, current biofabrication techniques yet cannot meet the functionality criteria of skin regarding its interface, and more developments regarding architectural design, cell types, and biomaterials are required.

Neurovascular and neuromuscular interfaces
Neural tissues exist in the two general nervous systems, i.e., the central nervous system (CNS) and peripheral nervous system. The neural and vascular tissue interface is provided by the physicochemical and functional coupling of neuronal, vascular, and ECM components at the NVU. NVU, a cellular and molecular interface between CNS and the circulatory system, is responsible for the regulation of microscopic blood flow locally, to provide the demands for oxygen and nutrients of specific brain sections depending on the activity changes, which is called hyperaemia (McCarty 2009, Muoio et al 2014. NVU is a multi-cell compartment including three main components of neural such as neurons, astrocytes, pericytes, microglia, myocytes, vascular such as endothelial cells (ECs), pericytes, and vascular smooth muscle cells, and ECM components (Muoio et al 2014, McConnell et al 2017, Lougiaki et al 2019. Through these different neural and glial cells, NVU detects the needs of neuronal supply and initiates required responses by providing vasodilation or vasoconstriction signals to meet the demands (Muoio et al 2014). ECM plays a role in the delivery of signals to the neural and vascular components and inducing cellular behavior (McCarty 2009). In general, the total interaction of neurovascular components with biochemical and mechanical cues from ECM facilitates the required stimuli to help neurogenesis and angiogenesis (Potjewyd et al 2018).
Blood-brain barrier (BBB) membrane, which acts as a semipermeable interface between circulating blood and the CNS to form the NVU, plays a vital role in maintaining the functions, restricting the passage of cells, particles, and large molecules, and tightly regulating the movement of molecules between blood and the brain (Daneman and Prat 2015, de la Vega et al 2019). Junctional protein complexes found between adjacent ECs of BBB are involved in the tight regulation of BBB permeability, extracellular signal transfer to the surrounding cells, and maintaining the integrity of BBB (Potjewyd et al 2018).
The neuromuscular junction (NMJ) is a synaptic interface specialized for transmitting nerve impulses to muscle fibers and triggering contraction (Gonzalez-Freire et al 2014). Pre-synaptic, intrasynaptic, and post-synaptic parts constitute the NMJ, which includes motor nerve terminal, synaptic basal lamina, and muscle fiber and muscle membrane, respectively (Hughes et al 2006). Electrical impulses from neurons are transmitted to muscles through chemical transmitters called acetylcholine. NMJ also has several channels, transmitters, specialized protein types, and micro-structures (Hughes et al 2006, Wu et al 2010. Collagen IV, laminin, fibronectin, and  Kim et al (2020a) entactin are the other proteins forming the molecular signature of NMJ (Hughes et al 2006). Due to the multicomponent intrinsic structures of NVU and NMJ, the bioprinting technique could be a promising approach to fabricating a heterogeneous construct with different materials and cell types from both neural and vascular sources. Table 4 summarizes the 3D bioprinting strategies, from selecting biomaterials to designing structures for NVU and NMJ and their effect on the final structure's functionality.
Neural stem cells (NSCs) and ECs are utilized as precursor cells for the construction of NVU (Han et al 2019). The presence of ECs in the neuroregeneration of CNS is crucial due to their secretion of adhesive molecules, which help the integrity of NVU. Co-culture of NSCs and ECs would improve the self-renewal ability and fate determination of NSCs. Another advantage is the stimulation of angiogenesis of ECs by NSCs. In this respect, Han et al co-cultured NSCs and ECs on a hyaluronan-grafted chitosan (CS-HA) substrate, which rapidly formed self-organized 3D co-spheroids of larger size. After forming co-spheroids, they were embedded in a gelatin-based hydrogel with 3,4hydroxyphenylpropionic acid (HPA) conjugate. An EB method was employed to fabricate tubular structures. Fibroblast growth factor 2 was added to hydrogel to mimic the neural environment prior to gelation. HPA hydrogels were formed in the presence of hydrogen peroxide and horseradish peroxide. The final hydrogel showed angiogenesis potential with a higher differentiation tendency of NSCs towards glial cells compared to other 2D culture conditions. In this study, it was possible to produce a viable mini-NVU by providing crosstalk of NSCs and ECs in 3D cospheroids embedded in gelatin-based hydrogel and bioprinted capillary-like structure.
The control of the internal porosity of the structure is another approach to tuning oxygen and nutrient flow to mimic the vascularization in a complex in vivo brain condition (Lozano et al 2015). A brainlike structure was biofabricated using a peptidemodified biopolymer Gellan Gum-RGD (RGD-GG) bioink embedded with primary cortical neural cells representing the formation of glial and neural cells as a NVU in vitro model. They developed a facile handheld reactive bathless bioprinting process with a syringe connected to a coaxial needle-tip to fabricate 3D layered architectures. The modified hydrogel supported the primary cortical adhesion and differentiation, and both neural and glial cells were viable with a proper morphology and could form interconnected neural networks. The scaffold provided internal porosity for oxygen and nutrient flow as well. Finally, they validated their model by printing a brain-like layered structure with interconnected and, at the same time, distinct layers with glial and neuron cells similar to NVU. A recent study was performed by Kim et al in which a neural, skeletal muscle construct was bioprinted with the improved formation of NMJ and capability of rapid integration with the host neural network in vitro (Kim et al 2020a) They used hMPCs and human neural stem cells (hNSCs) with the aim of pre-formation of NMJ, and hence, the muscle maturation and integration of the biofabricated construct with the host neural tissue were enhanced. Pre-formation of NMJ on myofibers resulted in the long-term survival of the bioengineered construct and increased cell differentiation. They used the EB strategy with three syringes, including a thermoplastic polymeric anchor as a mechanical reinforcing agent, cell-laden hydrogel, and an acellular sacrificial material. Hydrogel for both cell-laden and acellular sacrificial material was comprised of fibrinogen, gelatin, hyaluronic acid, and glycerol in Dulbecco's modified Eagle medium (DMEM). PCL was employed as a mechanical supporting material, which was printed in a parallel pattern like a box structure. Cell-laden bioink and gelatin-based sacrificial ink were printed in parallel fibers, and then sacrificial ink was removed after the printing process to create empty microchannels (figure 6). Afterward, the hydrogel was crosslinked by thrombin and characterized in vitro. Their results revealed cell survival after seven days of culture, and hNSCs were differentiated into glial and neural cells. Human MPCs also differentiated into myofibers and formed NMJs integrated with the host tissue. NMJs were formed between newly formed myofibers and the host nerve tissue.
In another study of the same group, a 3D muscle of a rodent model for tibialis anterior muscle defect was fabricated using the same multi-dispensing printing setup with a similar scaffold design (Kim et al 2018a). Their final bioprinted constructs represented high cell viability, and myofibers were formed and aligned along the printed patterns in vitro. Moreover, after two weeks of in vivo implantation, myofibers formation, vascularization, and neural integration with the host tissue were observed. After 4 and 8 weeks of implantation, the structural evaluation showed continuous myofiber formation and maturation by hMPCs. Finally, the formation and maturation of NMJ were identified at week 8 of implantation, showing the potential of the EB technique for a short-term repair of NMJ.
It is worth mentioning that due to the complexity of the neural tissue and cellular sensitivity to the surrounding environment, printing studies in this field are mainly concentrated on the fabrication of scaffolds with the potential to guide neural cells or vascularization rather than focusing on a thorough investigation of the tissue interface in constructed structures. An emerging approach in 3D bioprinting of neural tissue is the utilization of multi-cells and/or induced pluripotent stem cells (Lee et al 2009, De la Vega et al 2018, Joung et al 2018, and multi-materials (Gu et al 2016, Zhu et al 2016, Zhou et al 2018, de la Vega et al 2019 to act either as an environment resembling the neural tissue or a differentiation guide for stem cells. It should be noted that meeting the printability criteria and simultaneously fulfilling concerns regarding biomimicry and mechanical properties is yet a challenge for bioprinting of neural tissue engineering.
In addition, BBB as an important part of the neural tissue interface is considered in several studies (Cho et al 2015, Zhu et al 2016, Wang et al 2017. Microfluidic platforms are commonly employed for the fabrication of BBB, which has the benefit of 3D printing technology to construct microfluidic channels or organ-on-a-chip formation (Booth and Kim 2012, Griep et

Conclusions, challenges, and future perspective
Tissue interfaces are one of the most complex and vital parts of the body. In case of irreversible damage, the treatment is challenging due to their inherent complexities with gradient nature of biochemical and mechanical properties, and multiple cell types composition within a micro-scale. 3D printing approaches provide a solution by placing tissuespecific cues, biomaterials, and cell types at high resolutions in a hierarchical and complex architecture, bringing promising prospects for the regeneration of tissue interfaces.
The establishment of a successful 3D biofabrication for anisotropic tissue interface structures is based on critical steps to be sequentially followed, including pre-design, processing, and post-processing. The basic knowledge about the physicochemical features of the tissue interface would enable the proper selection of cell types/biomaterials within an appropriate pre-design. Each tissue interface has specific features requiring the development of novel bioinks to achieve their complexity. The biochemical composition of the bioinks can be fine-tuned with the addition of growth factors, bioactive molecules, molecular cues or varying numbers of cells and cell and matrix types to reach the natural properties of tissue interfaces as well as assist better integration with the host tissue. The performance of the printing process depends on polymer or bioink properties. The developed bioink should meet printability requirements, including viscoelasticity, gelation time, and crosslinking mechanism. In addition, the printing systems should also allow for customized printing patterns to enable microscale material deposition in a proper and desired geometry. Each bioprinting approach has advantages and disadvantages regarding bioink features, printability, and precision. Even though some could generate constructs in a nanoscale resolution, the final structure may not provide a suitable microenvironment for cell viability and functionality or meet the targeted physical characteristics of the tissue interface. Moreover, integrating subsequent layers comprised of materials with different physiochemical properties could be challenging. For example, the soft and rigid material connection could be challenging. From a mechanical point of view, the difference in stiffness could be up to a few orders of magnitudes. In this context, a sharp transition from soft to hard tissue reduces the effectiveness of the fabricated structure. Consequently, under mechanical forces stress concentration at the interface would cause insufficient mechanical stability and eventually failure of the integrated structure. Furthermore, different degradation kinetics of materials must be considered in this interfacial region. The recent improvements in bioprinting processes and bioink development have addressed some of these issues that positively contributed to the further advancement of interface tissue engineering. Modifying physical structures, such as pore size, fiber diameter, and bioink types, has enabled engineering specific tissue analogs with different functionalities at different tissue zones. The recent development of bioprinting technologies using multi-material and multi-extrusion/printing systems provided the construction of structures with controlled micro-architecture of tissue interfaces.
The post-processing step is also critical, which determines the functionality of the biofabricated structure, including viability, proliferation, differentiation of cells, and remodeling of the construct. While biofabrication technologies facilitate controlled implementation and realization of physicochemical properties and geometrical characteristics of the tissue interfaces, the traditional cell culture methods need to be modified to allow multiple cell type co-culturing systems. A lack of balance between proliferation and differentiation among different cell types may hinder the successful maturation of tissue interfaces in vitro. Developing a dynamic environment and culturing conditions would allow the postbiofabrication phase and a successful translational stage.
The initial steps have been accomplished in developing interface tissue models for drug discovery studies or small model organs. We envision its clinical applications are forthcoming. Due to the potential issues related to ethical regulations and social barriers, animal models are mostly used to predict the body's response to the engineered viable biomimetic tissue interface structures but they may not be biologically and mechanically biocompatible with the human body. Since tissue interfaces are mechanically unstable, the disintegration of the implant and failure of the integration with the host tissue is highly possible. In addition, the transition of the clinical infections to the host tissue may cause severe problems. However, continued developments may help to overcome the challenges associated with transplantation, and facilitate the rapid integration of bioprinted tissue constructs with the living body.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.