*3.1. Discontinuous Processes*

Discontinuous processes involve the separate fabrication of scaffolds' layers followed by their assembly into 3D structures. These two steps often increase processing times but take advantage of the possibility to use micro/nanofabrication technologies for scaffolds' features creation. This was mainly achieved by replication methods, such as those highlighted in Figure 3, where layers are obtained by replicating the features of master molds. As shown, replication methods can be divided into two main groups based on mold type, namely elastomeric (PDMS) and rigid molds, and two sub-groups, depending on polymers processing (solution/temperature plasticization). We also report suitable assembly techniques for the fabricated layers considering the absence (cell-free)/presence of biological matter (cell-laden). However, we would like to point out herein that layers' fabrication and assembly are not confined to only one set of methods and could be combined properly depending on the application.

Common methods to fabricate two-dimensional (2D) layers involved the deposition of pre-polymer or polymer solutions onto PDMS mold by casting or spin-coating, followed by a consolidation step. For instance, Gallego et al. have presented a multilayer micromolding technique to fabricate and assemble PCL scaffolds [92]. Layers were fabricated via spin-coating of a PCL solution in tetrahydrofuran and dimethylsulfoxide (1:3:6 w/w/w ratio) at 4000 rpm for 1 min. Later, solvent was extracted overnight and <sup>10</sup> <sup>μ</sup>m thick PCL layers, with 45 <sup>×</sup> <sup>45</sup> <sup>μ</sup>m<sup>2</sup> pores were achieved. One of the previously obtained layers was then transferred onto a glass slide and manually stacked to another layer for 3D scaffold building. By following this approach, the authors obtained up to 100 μm thick PCL scaffolds characterized by 81% porosity, which were suitable for studying the effect of pores size and architecture on cell behavior in vitro. A similar approach was used by Sodha et al. for preparing PCL scaffolds with 200 μm circular or star-shaped pores for retinal transplantation [93].


**Figure 3.** Discontinuous processes overview scheme. Left side: two-dimensional (2D) layers' fabrication processes. Right side: three-dimensional (3D) scaffolds' assembly processes. APS: Poly (ester-amide),1:2 poly (1,3-diamino-2-hydroxypropane-co -polyol sebacate); ECM: Extracellular matrix; ES cells: Embryonic stem cells; hMSC: Human mesenchymal stem cell; HUVEC: Human umbilical vein endothelial cell; NIH3T3: Mouse embryo fibroblast cell line; PCGA: Poly (ε-caprolactone–co-glycolic acid); PCL Polycaprolactone; PGS: Poly (glycerol sebacate); PLGA: Polylactic-co-glycolic acid; PLLA: Poly(L-lactic acid); PLT32o: Poly (limonene thioether); Poly(NIPAAm-co-HEMAHex): Poly (N-isopropylacrylamide–co-2-hydroxyethylmethacrylate-6-hydroxyhexanoate); POMaC: Poly (octamethylene maleate (anhydride) citrate; RLC: Rat liver cells; SMC: Smooth muscle cell; SPCL: Starch-polycaprolactone; TE: Tissue engineering.

A valid alternative to spin-coating consists in solution infiltration through a vacuum. Rosellini et al. [94] in fact fabricated a biomimetic myocardial scaffold, based on a simplified model of an original ECM microarchitecture. Several 25 <sup>μ</sup>m thick layers with 100 <sup>×</sup> 500 <sup>μ</sup>m2 rectangular pores were successfully fabricated and thermally assembled to promote layers' merging and achieve a mechanically consistent scaffold.

Freeze-drying has been proven as another effective solution-based consolidation method to increase layers' thickness and obtain additional porosity. For example, He and co-workers [95] have fabricated 2 mm thick cylindrical layers by pipetting a silk fibroin/gelatin solution onto a pre-frozen micropatterned PDMS mold. The frozen system is then freeze-dried for at least one day to extract the residual solvent, preserving the fabricated microstructure. Results showed the possibility to modulate layers porosity, in the 70–90% range, and pores size, from 125 to 225 μm by changing the concentration of the polymer in solution, to control cell behavior. A solution-mediated bonding was used to prepare microstructured scaffolds mimicking the liver lobule architecture for liver TE purposes. The use of freeze-drying for polymeric layers' setting was also explored by Wang et al. [37], who fabricated porous scaffolds for vascular TE purposes. The authors used a microfluidic molding method to obtain 500 μm thick chitosan/gelatin layers (100 μm microstructures thickness) pipetting a 1:1 solution between a PDMS mold covered by a glass slide. The final layer was achieved by cooling and freeze-drying. An interesting aspect of their approach was that, before scaffolds' assembly, the layers were seeded with human umbilical vein endothelial cells (HUVECs) or smooth muscle cells (SMCs) with bonding promoted in this case by the cell/cell and cell/ECM interactions. Morphological and histological analysis demonstrated the possibility to create a complete branching vascular network and direct SMCs growth into fiber-like bundles inside the microstructured channels. A similar approach was implemented by He et al., who fabricated agarose/collagen layers by solvent casting and thermal gelation [96]. These layers were seeded with HUVECs/collagen suspension, disposed inside an alignment mold, and bonded with the aid of a thin layer of agarose to obtain a fully perfusable 3D construct.

To explore the advantages of combining layers and cells, Son et al. [97] have presented an evolution of the aforementioned methods using cell-laden solutions and a solution cross-linking assembly method to fabricate a 3D construct which mimics the hepatic liver lobule with sinusoids. To accomplish this purpose, a cell-laden alginate suspension was casted on a plasma-cleaned PDMS mold. Then, the system was incubated into a humidifier with a cross-linking reagent to induce gelation and achieve 8 × 8.7 mm layers with thicknesses up to 200 μm. The authors fabricated a PDMS chamber for layer stacking and used a small amount of alginate solution and cross-linker at layers' edges for bonding. The results show that layers maintained their structure during cell proliferation, while the manipulation techniques did not result in cell loss. Furthermore, cells show high viability because scaffolds' lateral and central pores ensure oxygen and nutrients' transport in the entire 3D structure. HepG2 cell-loaded constructs exhibited increased hepatic secretion and, when used in combination with mouse embryo fibroblast cell line (NIH3T3), allowed for studying cells interactions in 3D co-culture experiments. This approach was also used to test different porous structures, namely hexagonal pores with size in the 100–500 μm range, and by using collagen as layers' material [98]. A patterned cellulose filter substrate was used for collagen layer manipulation and the scaffold was assembled by alternating cell-free and HUVECs-laden collagen sheets to study cells' migration and scaffold vascularization [98].

Solution-based layers' fabrication was also implemented by using pre-polymer mixtures, which can be consolidated by UV radiation, as reported by Zhang and co-workers, for the microfabrication of the AngioChip scaffold [99]. Layers were fabricated from a mixture of poly (ethylene glycol) dimethyl ether (PEGDM) and poly (octamethylene maleate (anhydride) citrate) (POMaC), that was injected in a patterned PDMS prior to UV cross-linking and solution consolidation. Later, the as-obtained layers (5 <sup>×</sup> 3.1 mm<sup>2</sup> surface and 150–300 <sup>μ</sup>m thickness) were demolded, stacked, and bonded by an additional UV treatment. The key feature of this micro-construct is the presence of a built-in endothelialized branched network, suitable to assess cardiac and hepatic tissues' responses to drugs delivered through the internal vasculature. For example, the generation of an angiogenic stimulus (thymosin β4) in vitro allowed endothelial cells' migration through the scaffold micro-holes as a first step of blood vessel formation in vitro. AngioChip also enabled fast anastomosis in vivo and tissue remodeling during the first week.

Processing biomaterials and bioactive molecules from organic solvent solutions require the removal of solvent residues from the final scaffolds, as these residues could be toxic for cells and tissues. In this context, previous researchers have also documented that PDMS could be used as a mold to produce micro-patterned layers from thermally plasticized polymers [100]. Yang et al. [101] have in fact presented several protocols to fabricate PLGA layers (120 μm wide pores and 60 μm thick) by PDMS micro-embossing at a temperature close to the PLGA glass transition temperature. The final porous scaffolds were obtained by stacking layers with the help of an alignment mold followed by compressed CO2 bonding for 1 h. This solvent-free approach was successfully applied to cell-seeded PLGA layers, demonstrating that CO2 bonding ensured proper human MSCs viability and functions [101]. Later, Xie and co-workers also demonstrated the possibility of bonding PLGA layers using N2, which resulted in enhanced embryonic stem (ES) cells' viability with respect to CO2 [102].

Although we have explained the motivations that have directed several researchers to choose elastomeric PDMS mold for layers' fabrication, features distortion during the process may be a critical issue. This problem arises because PDMS may swell and deform in contact with a broad range of organic solvents [103] or during compression [101]. As shown in Figure 3, rigid molds are the suitable alternative to overcome this limitation and fabricate layers for TE applications through replication techniques from solution and thermal processing.

Regarding solution-based processes, the first examples we introduce are those presented by Papenburg et al., who fabricated layers of different biocompatible polymers through solution casting/phase separation on a silicon mold [104]. Morphological analysis evidenced 80% porosity and high pore interconnection, low closed isolated pores, and a minor dense outer layer. However, this process leads to films with micropattern dimensions differing to the mold pattern because of film shrinking during the solvent extraction process [104]. Manual stacking and residual solvent bonding enabled the achievement of 3D scaffolds. In a further work, settled layers were seeded with C2C12 pre-myoblasts cells and rolled up to form a hollow cylinder without bonding to evaluate the effect of static and dynamic culture conditions on nutrient transport and cell behavior in vitro [104].

Recently, Liu et al. proposed an electrodeposition process for the preparation of rat liver cell (RLC-18)-laden alginate layers for an in vitro liver application [105]. The process involves the casting of a solution onto a rigid mold, fabricated through photolithographic techniques, with an architecture mimicking the hepatic lobule morphology. Then, the solution was electrodeposited for 15 s to obtain 300 μm thick cell-laden hydrogel layers, whose cells remain viable during all the microfabrication steps and proliferate over time. Two layers were subsequently stacked in an appropriate mold, similar to the process described in Reference [97], to obtain a 3D scaffold.

As for the "Angiochip" device [99], cell-free scaffolds for vascular TE purposes represent interesting examples of modules fabricated by solution consolidation [106,107]. In the work by Ye and co-workers, a modular strategy was proposed to build a slowly degradable poly(ester-amide),1:2 poly (1,3-diamino-2-hydroxypropane-co -polyol sebacate) (APS) bilayer scaffold connected to a microfluidic base through a rapidly degradable porous poly (glycerol sebacate) (PGS) module fabricated by an acrylic template [106]. As-obtained four-layer scaffolds increased the 3D permeability to oxygen and nutrients in vitro and degraded in vivo with a rate suitable to enhance scaffold vascularization. The fabrication of layer-by-layer heart scaffolds by photo-cross-linkable poly (limonene thioether) (PLT32o) prepolymer was reported by Fisher et al., with the aim to provide long in vivo half-life [107]. Layers with rectangular micropores (362 <sup>×</sup> 564 <sup>μ</sup>m2) were obtained by replica molding (REM) of polycarbonate molds and were assembled to form 3D scaffolds with elastomeric mechanical behavior and were able to retain structural integrity until one month in vivo.

Micro-embossing in rigid molds is the last discontinuous process described in this section. This process was widely used by Ryu and co-workers, who fabricated silicon molds to realize patterned layers with interconnecting structures made of thermoplastic materials such as PLGA, poly (p-dioxanone), and Monocryl® [108]. Morphological analyses showing the possibility of embossing structures of different aspect ratios were presented and discussed. Technological points of interest for the process, mainly mold-microstructures detachment and modulation of polymers bulk properties were also addressed. Porous scaffolds were fabricated by layers' stacking and bonding using a novel solvent vapor-mediated assembly process. Briefly, two layers were placed in an assembly chamber at a pre-defined temperature followed by a solvent vapor injection. Layers bonding was then achieved bringing the layers in contact under pressure. By this approach, it was possible to preserve layers' features and eventually incorporate bioactive molecules. As a result, 60 μm thick scaffolds with rectangular pores (20 × 30 μm) were achieved and tested as a 3D platform for single-cells' culture and characterization. In another work, Lima et al. [109] produced PCL and starch-polycaprolactone (SPCL) thicker layers (500 μm) with 300 μm circular pores and 300 thick pillars using a stainless-steel mold. Layers were manually stacked and bonded by using a PCL solution in chloroform, finally achieving 1.5 mm thick scaffolds with 88% porosity for in vitro bone TE.

#### *3.2. Continuous Processes*

Scaffolds' fabrication has evolved significantly by continuous processes due to the impressive evolution in the fields of materials science, cells engineering, and AM materials/cells processing platforms. AM are bottom-up processes where the basic components are assembled layer-by-layer to make objects from 3D model data. For example, the common workflow starts with the 3D virtual reconstruction of the defect to regenerate and can end with a patient-specific scaffold implantation to the site of injury [110]. To date, several AM systems available in market are capable of performing multiple operations simultaneously in the same work, e.g., extruding a synthetic polymer strand from a nozzle and embedding a cell-laden hydrogel in a predefined position. In addition, other important features of AM are scaffolds' reproducibility and consistency, as well as the possibility to create complex shaped 3D structures that are necessary for patient-specific treatments.

Regarding the application fields, AM techniques have still proven versatile and of great impact in regenerating several tissues. Indeed, the level of control offered by these techniques is a key technological aspect to increase our knowledge regarding biophysical and biochemical cues governing tissues' formation and functions. Through this section, we will show relevant results published in recent literature about AM scaffolds, pointing out advantages of the implemented manufacture technique and promising results.

Bone is a dynamic tissue characterized by heterogeneous and anisotropic structures and compositions that are required to support biomechanical and biological bone functions. The hierarchical structure of bone is composed of nanostructures made of organic (e.g., collagen) fibers and inorganic (HA) crystals that form the macroscopic cortical and cancellous bone structures passing through a series of intermediate microstructures, like lamellae, osteons, and harvesian channels. Scaffolds for bone regeneration must mimic bone morphology and structure. Concomitantly, these scaffolds must promote bone deposition (ostoconductive) and must be capable of delivering growth factors, such as BMPs and TGFs, to promote recruited cells' osteogenic differentiation (osteoinductive).

Advances in bone scaffolds' fabrication by AM processes have tried to replicate bone biological and biomechanical complexities. An example of this biomimetic approach is proposed by Kang and co-workers, who developed an innovative AM platform named "integrated tissue–organ printer" (ITOP) [110]. The ITOP is equipped with multi-cartridges capable of printing concomitantly synthetic polymers and cell-laden hydrogels with a resolution down to 2 μm for biomaterials and down to 50 μm for cells (Figure 4a). These features were used to fabricate a calvarial bone construct (8 mm diameter × 1.2 mm thickness) made of a PCL and tricalcium phosphate (TCP) nanoparticles blend and stem cells-loaded hydrogels, embedded in predefined positions (Figure 4b). After 10 days of in vitro osteogenic culture, the bioprinted bone is implanted in a calvarial bone defect region to study maturation up to five months. Histological (Figure 4c) and immunohistological images clearly show new bone formation even in the defect central portion; moreover, the presence of blood vessels demonstrates the absence of tissue necrosis confirming regeneration effectiveness. These promising results suggested the potential utility of printed living tissue constructs in translational applications.

Other recent examples have demonstrated, in vivo, successful calvarial bone regeneration using printed scaffolds made of hydroxyapatite (HA) or PCL/PLGA/HA composite, respectively [111,112]. Furthermore, the advantage of printing techniques to process multiple bioinks in a single scaffold was used to bioactivate the scaffold with BMP-2 peptide or μ-RiboNucleic Acid (μ-RNA) conjugates to enhance stem cells' osteoinduction to stimulate in vivo bone formation.

The regeneration of interface tissues, as osteocartilagenous anatomical regions, requires scaffolds displaying compositional and structural complexity that are only achievable with AM processes. In this context, an interesting fabrication approach is presented by Mekhileri and co-workers [113]. The authors have combined a commercial printer (BioScaffolder) with a custom-made device capable of handling pre-loaded μ-tissues (Figure 4d). The fabricated polymer strands are about 225 μm with a maximum resolution of 25 μm. μ-tissues could be positioned in scaffolds' pores once the fabrication process is finished or could be integrated during the fabrication process (inset of Figure 4d),

demonstrating the possibility to fabricate large hybrid constructs with a predetermined architecture and mechanical stability. μ-tissues were produced with dimension of 700 μm to 1.4 mm, without undifferentiated or necrotic cells in the central regions at 28 days of in vitro culture and the chosen dimension was 1 mm for the integration into scaffolds due to design and handling considerations. Using this approach, the authors presented a proof of concept scaffold for joint resurfacing purposes (Figure 4e,f), in which two different natural hydrogels' microspheres were used to simulate the biphasic bone and cartilage portions. The process enabled the manipulation and positioning of the μ-tissues inside the scaffold (Figure 4g), while adjacent μ-tissues fusion is observed at 35 days of in vitro culture in chondrogenic differentiation media (Figure 4h).

A wide range of materials was used for AM purposes in this field, with encouraging results. For example, Gao and co-workers [114] have synthesized a strong copolymer hydrogel with large stretchability (up to 860%) and high compressive strength (up to 8.4 MPa). The material had a rapid thermoreversible sol-gel transition behavior that makes it suitable for graded scaffold printing. Furthermore, this gradient hydrogel scaffold printed with TGF β1 and β-tricalciumphosphate for chondral and bone layers respectively, promotes simultaneous regeneration of cartilage and subchondral bone in a rat model [114]. In another work, Deng and co-workers [115] used 3D printing process to prepare lithium (Li)- and silicon (Si)-containing scaffolds to study the effect of ions' release on osteochondral tissue repair in rabbits. The release of Li and Si ions synergistically exerted a positive effect on cartilage through the activation of hypoxia-inducible factor (HIF-1α) pathway and preservation of chondrocytes from an osteoarthritic environment. Concomitantly, Li and Si ions released from the scaffold improve subchondral bone reconstruction through activating Wnt signal pathways.

The versatility of AM techniques in terms of materials choice and structure design enabled the use of additive manufactured scaffolds in other important fields, such as cardiac and nerve tissues' regeneration. One of the most interesting works concerns a scaffold for cardiac remodeling after myocardial infarction, which is proposed by Yang and co-workers [116]. This device was fabricated by employing the fused deposition modeling (FDM) technology, whose typical resolution is of hundreds of microns [117], to obtain a stacked construction of PGS/PCL blend with regular crisscrossed strands and interconnected micropores (Figure 4i). The PGS/PCL scaffolds exhibited improved elasticity and toughness, if compared to raw PCL and PGS scaffolds respectively, and mechanical properties similar to heart tissue. Moreover, the PGS/PCL mixture was filled with NaCl particles with the goal to leach them out to generate an additional interconnected microporosity for oxygen and nutrients' transport and neovascularization. The study was conducted to first assess the in vitro and in vivo scaffolds' behavior, demonstrating an interesting therapeutic effect in rodents with respect to scaffold-free and PCL or PGS scaffolds implanted after myocardial infarction (Figure 4j), and later to study an annular-shaped scaffold whose results indicate a promising application for preventing ventricular dilation (Figure 4k). Moreover, those 3D-printed PGS/PCL scaffolds possess interesting shape-memory properties after rolling, folding, and compression. This feature holds promise for minimal invasiveness delivery via, for example, a catheter or mini-thoracotomy, in case of future surgical translation.

Another interesting example in this field is that of Boffito and co-workers [118], who have used a custom-made AM equipment to fabricate polyurethane (PU) scaffolds seeded with human cardiac progenitor cells (CPCs). PU scaffolds grafted with laminin-1 supported CPCs differentiation in cardiomyocytes while preliminary in vivo subcutaneous implantation experiments evidenced a minimal inflammatory response and adequate angiogenesis, suggesting their future use as implantable patches for myocardial TE.

Regarding the neural TE field, here we reported the results of the study of Koffler and co-workers [119], who have developed a "microscale continuous projection printing method" (μCPP) (Figure 4l) to fabricate, in a very short time (less than 2 s), a 2 mm-thick biomimetic scaffold for spinal cord injury repair (Figure 4m,n). Materials used for fabrication were mixtures of PEG and gelatin methacrylate. This material, in fact retained its structure over four weeks in vivo and exhibited an acceptable inflammatory response. The chosen material was then processed to obtain scaffolds

mimicking the spinal cord structure (Figure 4m,n) and which were seeded with neural progenitor cells (NPCs) before implantation. After six weeks in vivo, injured host axons regenerate into 3D biomimetic scaffolds and synapse onto NPCs implanted into the device (Figure 4o). Furthermore, implanted NPCs extend axons out of the scaffold and into the host spinal cord below the injury to restore synaptic transmission and significantly improve spinal cord functionality.

The advantage of NPCs-laden 3D-printed biocompatible scaffold on nerve tissue repair is also highlighted in Reference [120], where clusters of induced pluripotent stem cell (iPSC)-derived spinal NPCs and oligodendrocyte progenitor cells (OPCs) are placed in precise positions within 3D-printed hydrogel scaffolds during assembly. A combination of transplanted neuronal and glial cells enhance functional axonal connections' formation across areas of the damaged central nervous system. Finally, the combination of cells and growth factor therapies, such as scaffolds releasing neurotrophin-3 growth factor [121], may represent a possible further step towards complete nerve tissue repair.

**Figure 4.** (**a**) Integrated tissue–organ printer (ITOP) system components and materials. (**b**) Photograph of the printed calvarial bone construct. (**c**) Histological image of the printed calvarial construct after in vivo implantation. (**d**) Image of the bioscaffolder + micro-tissue injection system (inset: working

concept overview of the micro-tissue injection system) used for the preparation of the osteochondral joint resurfacing device. (**e**) Computer-aided design (CAD) images and (**f**) optical image of an assembled hemispherical construct. (**g**) Image of μ-tissues in 3D printed PCL fibers and (**h**) resulting 4- ,6-diamidino-2-phenylindole (DAPI) (blue) and Aggrecan (purple) antibodies staining of the construct showing cells distribution and μ-tissues fusion at 35 days of in vitro chondrogenic culture. (**i**) Fused deposition modeling (FDM) machine overview and materials for the elastic cardiac patch fabrication. (**j**) Illustration of the scaffold implantation site after induced myocardial infarction in rats. (**k**) Representative Masson's trichrome stained heart section four weeks after implantation. Black boxes denote higher magnification area of the left panel. Black arrows indicate the annular-shaped PGS-PCL scaffolds. Scale bars: 2.0 mm. (**l**) Microscale continuous projection printing (μCPP) system used to fabricate PEG–gelatin methacrylate scaffolds loaded with neural progenitor cells (NPCs) for nerve regeneration. (**m**) Spinal cord structure evidencing fascicles regions (motor systems are shown in green and sensory systems are shown in blue) and (**n**) corresponding scaffold. (**o**) Image of the NPCs-loaded scaffold after four weeks in vivo showing channels filled with green fluorescent protein (GFP)-expressing NPCs. (**a**–**c**) Reproduced with permission from Reference [110] (Kang, Nature Biotechnology; published by Springer Nature, 2016). (**d**–**h**) Reproduced with permission from Reference [113] (Mekhileri, Biofabrication; published by IOP Publishing, 2018). (**i**–**k**) Reproduced with permission from Reference [116] (Yang, Advanced Healthcare Materials; published by John Wiley and Sons, 2019). (**l**–**o**) Reproduced with permission from Reference [119] (Koffler, Nature Medicine; published by Springer Nature, 2019).
