**3D Bioprinting Strategies for the Regeneration of Functional Tubular Tissues and Organs**

#### **Hun-Jin Jeong <sup>1</sup> , Hyoryung Nam <sup>2</sup> , Jinah Jang 2,3,4,5,\* and Seung-Jae Lee 1,6,\***


Received: 29 February 2020; Accepted: 30 March 2020; Published: 31 March 2020

**Abstract:** It is difficult to fabricate tubular-shaped tissues and organs (e.g., trachea, blood vessel, and esophagus tissue) with traditional biofabrication techniques (e.g., electrospinning, cell-sheet engineering, and mold-casting) because these have complicated multiple processes. In addition, the tubular-shaped tissues and organs have their own design with target-specific mechanical and biological properties. Therefore, the customized geometrical and physiological environment is required as one of the most critical factors for functional tissue regeneration. 3D bioprinting technology has been receiving attention for the fabrication of patient-tailored and complex-shaped free-form architecture with high reproducibility and versatility. Printable biocomposite inks that can facilitate to build tissue constructs with polymeric frameworks and biochemical microenvironmental cues are also being actively developed for the reconstruction of functional tissue. In this review, we delineated the state-of-the-art of 3D bioprinting techniques specifically for tubular tissue and organ regeneration. In addition, this review described biocomposite inks, such as natural and synthetic polymers. Several described engineering approaches using 3D bioprinting techniques and biocomposite inks may offer beneficial characteristics for the physiological mimicry of human tubular tissues and organs.

**Keywords:** 3D bioprinting; biocomposite ink; tubular tissue; tubular organ

#### **1. Introduction**

Tubular tissues and organs exist with various forms and functions in the gastrointestinal (esophagus, intestines), respiratory (trachea), vascular (veins, arteries), and urinary (bladder, urethra) systems [1]. These tubular tissues have various diseases and malfunctions requiring appropriate therapeutic interventions, such as donor tissue transplantation, autologous implant, and replacement with a synthetic prosthesis. Autologous transplantation is considered as one of the best therapeutic methods; however, in the case of the trachea and esophagus tissue with little redundancy and non-existent autologous tissues, therapeutic approaches using donor tissues or synthetic prosthesis are required [2–4]. Donor tissue transplantation is an ideal option, but there remains a disparity between the number of the appropriate donors and the high demand for the therapeutic use of donor tissue [5–7]. In addition, finding a suitable donor tissue is not easy since most of the tubular tissues

are associated with poor prognosis after surgery [4,8]. For these reasons, many tissue engineering approaches have been researched for the manufacture of suitable tubular tissues and organs.

Successful bladder tissue engineering using tissue-engineered hollow spherical biodegradable structures was first reported in 1996 [9]. Since then, many studies have been reported to create artificial functional tubular tissue, such as tracheal tissues. However, these tubular tissues generally have similar morphological features, but the fabrication of tubular tissues requires high-level microfabrication-techniques due to their complex hierarchical macro- and micro-structure containing the different cell types and extracellular matrix (ECM) [10–13].

Various microfabrication techniques for tissue engineering, such as electrospinning [14–19], cell-sheet engineering [20–22], and mold-casting [23–26], have been widely studied to make complex multi-layered architecture for artificial functional tubular tissues. These approaches would use additional substrates (e.g., rotating rod, sacrificial mold) to create the tubular architecture, as well as requiring complex and multiple manufacturing processes. Besides, these approaches are not only insufficient to create the tubular structure with a target-specific mechanical property but also restrict shape-freedom due to the technical limitations.

3D bioprinting technique is emerging as an alternative to overcome the limitations of fabrication in terms of building tubular tissues and organs [27,28]. 3D bioprinting has also been utilized for higher-complexity structures with printable biocomposite inks containing the living cells and natural and synthetic polymers [29]. The greatest benefit of 3D bioprinting technology in tubular tissue engineering is the ability to fabricate tubular structures with multi-layer free-form constructs, as well as allowing the placement of biomaterials in a cell and printable ink containing the biochemical microenvironmental cues [10]. In particular, this technology has unlimited possibilities that are feasible for producing complex tissues and organs. Additionally, it can be applied anatomically and clinically since it should facilitate the manufacture of patient-tailored 3D structures [30–32].

Therefore, this review dealt with the state-of-the-art of 3D bioprinting technologies, various biocomposite inks, and their applications to tubular tissue engineering focused on a blood vessel, trachea, and esophagus tissue regeneration.

#### **2. 3D Bioprinting Techniques**

To achieve the building of 3D-engineered human tissue and organ analogs, it is necessary to use accurate and well-controlled fabrication methods for suitable biomaterials and living cells. Due to these functional 3D fabrication requirements, four types of 3D bioprinting techniques have been developed based on the principle of releasing the printable biomaterials from the printing head.

#### *2.1. Extrusion-Based Bioprinting Systems*

The principles of extrusion-based bioprinting are dispensing biomaterials through the nozzle by physical force (e.g., pneumatic pressure, piston, or metal screw) and pneumatic pressure (Figure 1a). The extrusion head moves in the x, y, and z directions under the instruction of the CAD-CAM software to produce a 3D architecture by staked biomaterial onto the substrate. Even if this technique has a lower accuracy than the other 3D bioprinting methods (ink-jet, laser-based), it can be capable of various biomaterials, such as cell-laden bioink, cell-spheroids, hydrogels, and high-viscosity thermoplastic polymers. This bioprinting method allows the extrusion of an extensive range of viscous materials (6–30 <sup>×</sup> <sup>10</sup><sup>7</sup> mPa·s), and the resolution of the extruding in the exit of the nozzle is in the range of 100 µm–millimeter [33,34]. Among them, thermoplastic polymer, such as polycaprolactone (PCL) [35–37], poly (lactide-co-glycolic-acid, PLGA) [38,39], poly (L-lactic acid, PLLA) [40,41], have been widely applied to fabricate hard-tissue and sturdy supporting constructs. The distinct advantage of this bioprinting technique is that it can be installed in a multi-head system, allowing the simultaneous use of one or more biomaterials, such as synthetic polymer and cell-laden bioink. Therefore, given the ability to quickly manufacturing complex 3D tissue structures that morphologically and biologically mimic the human body, the extrusion technique is regarded as a promising clinical approach [30].

include the acrylics and epoxies. However, to apply the tissue-engineered approaches, the biocompatible photocurable resin needs to contain propylene fumarate (PPF) and trimethylene

challenging due to cytotoxicity of the photocurable resins and high cost for system installation.

**Figure 1.** Schematic illustration of the (**a**) extrusion-based; (**b**) ink-jet; (**c**) assistant laser; and (**d**) stereolithography-based 3D bioprinting systems. **Figure 1.** Schematic illustration of the (**a**) extrusion-based; (**b**) ink-jet; (**c**) assistant laser; and (**d**) stereolithography-based 3D bioprinting systems.

#### **3. Printable Biocomposite Inks for Various 3D Bioprinting Techniques** *2.2. Ink-Jet Bioprinting Systems*

Printable biocomposite inks are generally classified as natural, synthetic, and functional polymers. Natural polymers (e.g., collagen, gelatin, alginate, ECM-based ink) have been widely used in the field of the tissue-engineering and have been considered promising biomaterials with similar components of native tissue or organs in the human body [52–54]. In particular, protein-based natural polymers, such as collagen, gelatin, and ECM-based ink, have a remarkable capacity to help regenerate the epithelial layer, which is essential for creating the functional tubular tissue. Alginate bioink has an inferior biological activity compared to protein-based natural polymers. However, it has been widely utilized as bioprinting material to build the tubular constructs due to the easily controllable printability and excellent biocompatibility. Ink-jet bioprinting makes use of mechanical pulses, such as piezoelectric and thermal, to manufacture small-sized bioink droplets (spot size resolution of around 50–75 µm) and is referred to as the drop-on-demand method (Figure 1b). In the case of the piezoelectric ink-jet, break piezoelectrical materials are made by generating acoustic waves using an actuator to create droplets [42]. The thermal ink-jet uses the generated small air bubbles by applying electrical heat within a heated printhead to make droplets. These air bubbles are able to control the material being extruded from the exit of the print nozzle [43]. Cell viability in ink-jet bioprinting may differ depending on the applied mechanical pulses.

Ideally, synthetic polymers support the structure of the 3D printed target-tissue and degrade completely after implantation without side effects. Also, synthetic polymers have to pass the verification of strict criteria to be applied in clinical settings. In this section, we have described representative biocompatible synthetic polymers, such as thermoplastic polycaprolactone (PCL), polyethylene glycol (PEG), and polylactic acid (PLA), which have been recognized by the Food and Drug Administration (FDA) or widely applied in the field of tubular tissue-engineering [55–58]. In addition, we previously reported the combinations of natural polymer and synthetic polymer or nanocellulose that have been tried for enhancing mechanical properties and positive biochemical The main advantage of the ink-jet technique is that it enables to print with picoliter-volume droplets to build micro-structures because it can control the desired ejected droplet size as a variation of ultrasound parameters, such as amplitude, pulse, and time. However, there are drawbacks, including the need for low-viscosity material (3.5–12 mPa·s) to avoid clogging, and bioprinting material should be quickly gelated in post-print for 3D build constructs. In addition, the mechanical property of the post-printed construct has weak solidity, and the drying of a printed droplet on the substrate during bioprinting is a problem to be solved.

#### factors [59]. *2.3. Laser-Assisted Bioprinting*

Laser-assisted bioprinting uses a laser source to fabricate at high precision onto substrates. There are two separate approaches: laser-guided direct writing and laser-induced forward transfer (LIFT) [44,45]. Compared to other printing approaches, laser-assisted printing was rarely used in the old days; however, it has recently become increasingly popular as the 3D printing method for microfabrication.

This technique consists of a focusing system (to align and focus laser), an absorbing layer (ribbon), a pulsed laser beam (to induce the transfer of bioink), and a substrate for the bioink layer (Figure 1c) [46]. In brief, the laser pulse is induced on the absorbing layer to create a high-pressure bubble from the bioink layer and then drop the bioink onto the substrate. This bioprinting method can also use bioink with the high-viscosity and high-concentration of the cell (1–300 mPa·s) since it is a nozzle-free system [47]. Therefore, the formation of delicate shaping and arrangement of the cell patterning can be achieved while bioprinting without affecting cell viability. However, the fast gelation of bioink is essential for achieving a highly precise shape. For this reason, the laser-assisted bioprinting is a time-consuming process because of the relatively low flow rate for the crosslinking of the material. Therefore, the challenge of building 3D constructs of clinically relevant size remains [48].

#### *2.4. Stereolithography-Based Bioprinting*

Stereolithography (SLS or SL) is a form of 3D bioprinting technology using a laser source (ultraviolet, infrared radiation) for building 3D structures. An SLS bioprinting system consists of the light source, a reservoir with the liquid photocurable resin, an elevator system, and a digital mirror device (DMD) [49]. This technique was developed commercially in 1986 by Chuck Hill and then widely used for creating prototypes and complex production parts (Figure 1d). It can allow relatively rapid manufacturing time since the whole layers of the 2D slicing pattern of the 3D constructs are irradiated in a photopolymer reservoir. Also, this system can fabricate the sub-micron structure with highly precise 3D shapes (~1.2 µm) because a laser source can be focused on a small spot in the photocurable resin [50].

As this stereolithography bioprinting technology has been applied in the field of tissue engineering, various materials have been developed to contain bioink and cells with photo-initiators. In particular, commonly used photocurable materials in stereolithography bioprinting technique include the acrylics and epoxies. However, to apply the tissue-engineered approaches, the biocompatible photocurable resin needs to contain propylene fumarate (PPF) and trimethylene carbonate (TMC). These biocompatible photocurable materials have been widely used to manufacture with complex architecture and sacrificial mold [51]. However, SLS 3D bioprinting is still challenging due to cytotoxicity of the photocurable resins and high cost for system installation.

#### **3. Printable Biocomposite Inks for Various 3D Bioprinting Techniques**

Printable biocomposite inks are generally classified as natural, synthetic, and functional polymers. Natural polymers (e.g., collagen, gelatin, alginate, ECM-based ink) have been widely used in the field of the tissue-engineering and have been considered promising biomaterials with similar components of native tissue or organs in the human body [52–54]. In particular, protein-based natural polymers, such as collagen, gelatin, and ECM-based ink, have a remarkable capacity to help regenerate the epithelial layer, which is essential for creating the functional tubular tissue. Alginate bioink has an inferior biological activity compared to protein-based natural polymers. However, it has been widely utilized as bioprinting material to build the tubular constructs due to the easily controllable printability and excellent biocompatibility.

Ideally, synthetic polymers support the structure of the 3D printed target-tissue and degrade completely after implantation without side effects. Also, synthetic polymers have to pass the verification of strict criteria to be applied in clinical settings. In this section, we have described representative biocompatible synthetic polymers, such as thermoplastic polycaprolactone (PCL), polyethylene glycol (PEG), and polylactic acid (PLA), which have been recognized by the Food and Drug Administration (FDA) or widely applied in the field of tubular tissue-engineering [55–58]. In addition, we previously reported the combinations of natural polymer and synthetic polymer or nanocellulose that have been tried for enhancing mechanical properties and positive biochemical factors [59].

#### *3.1. Natural Polymers*

#### 3.1.1. Collagen

Collagen is a representative natural polymer applied to bioprinting, and it consists of proline and glycine, with a triple-helix arrangement of polypeptides, as the most abundant protein in the human body [52,60]. In human organs (e.g., skin, bone, cartilage, vessel) and connective tissues, various collagen types exist, such as collagen I, III, and IV [61]. Among them, collagen type I is the most abundant and also the most commonly used in 3D bioprinting [62]. Collagen has characteristics of little cross-species immunological reaction and low toxicity, as well as allowing enhanced cell attachment and proliferation due to the presence of asparagine-glycine-aspartic acid (RGD) residues [11]. For this reason, collagen-based bioinks are regarded as highly promising biomimetic materials.

The main advantage of the collagen-based bioink is that it enables the embedding of living cells with ECM components and biochemical materials. However, because it has a crosslinking property, the use of a crosslinker or gelation process by temperature is essential for the construction of the 3D structure. In addition, the mechanical strength and bioprinting property of the collagen-based bioinks are dependent on viscosity due to collagen contents. For this control, many studies have combined collagen with biocomposite materials, such as fibrin [63,64], alginate [65], chitosan [60], and agarose [66], for improving printability and mechanical properties. Collagen-based bioink is undoubtedly an excellent biomaterial, but there remains scope for the improvement for use as bioprinting materials.

#### 3.1.2. Gelatin

Gelatin is a type of protein obtained from collagen as a partially hydrolyzed form, and it is a biodegradable and biocompatible natural polymer [67]. Likewise, as with collagen, gelatin-based bioink can enhance cell attachment and proliferation because it has an RGD sequence with abundant integrin-binding motifs. Gelatin is dissolved in water to maintain the thermo-sensitivity property, but it reversibly forms a low-viscosity soluble state at human body temperature [68]. Because of these limitations in maintaining the form, using only the gelatin-based bioink as a printable biomaterial is not suitable to build sturdy 3D tissue structures. Therefore, many studies have attempted the development of the printable gelatin-based composite ink mixed with other polymer materials, such as PCL [69,70], chitosan hydrogel [71], hyaluronic acid [72], fibrin [73], alginate [74,75], and silk [76,77], for improving structure's stability.

Gelatin methacrylate (GelMA) has been widely used as an advanced bioink that modifies photocrosslinkable polymers [78]. The mixture of GelMA with photoinitiator (e.g., 2-hydroxy-1-(4-(hydroxyethoxy) phenyl)-2-methyl-1-propanone (Irgacure 2959)) undergoes rapid crosslinking after extrusion through exposure to UV light (360–480 nm in wavelength). This crosslinkable property enables the structural stabilization after bioprinting. Also, GelMA has excellent biological characteristics of cell adhesion, biodegradability, and cell migration because it involves collagen and gelatin components, such as integrin-binding motifs and RGD. Due to these promising properties, many studies have applied the combined material for improving the desired quality.

#### 3.1.3. dECM Ink

dECM-based ink has been regarded as a promising material for 3D bioprinting [79]. dECM-based ink is fabricated by the decellularization process of the target-tissue. It has an inherent component of tissue-specific microenvironment cues, such as proteoglycans, glycoprotein, and collagenous protein. To date, various dECM-based inks have been reported for target-specific tissues, such as derived skin [80], bone [39], vessel [59], liver [81], kidney [82], and so on. Each derived tissue has different printability properties, but all have the distinguishing feature of temperature-responsive gelation under the physiological environment [83].

#### 3.1.4. Alginate

Alginic acid, also called alginate, is an anionic polysaccharide distributed in the cell walls of brown algae [84]. It has a hydrophilic property and forms a viscous gel when hydrated. Alginate hydrogel has been applied as a wound dressing material because it has good biocompatibility and is structurally similar to natural ECM with a bioinert property [85]. In addition, alginate has been extensively used as an ink to fabricate the 3D structure in the field of tissue engineering because it can robustly form a cell-compatible hydrogel by instantly polymerizing using multivalent cations (e.g., Ca2+, Ba2+). Given their facilitation of tissue formation, hydrogel inks have been modified for a variety of tissue-engineered approaches, such as bone [86], cartilage [87], and vascular tissue [88]. In addition, because the alginate has no cell-adhesive site, the bioactive component enables the addition of the signal trigger, such as RGD, for cell viability and differentiation [89].

#### *3.2. Synthetic Polymer*

#### 3.2.1. Polycaprolactone

PCL is one of the aliphatic polyesters; it is the most frequently used biomaterial for 3D bioprinting in the field of tissue engineering. PCL has superior printability due to its low melting temperature and glass-transition temperature. In addition, it is well known as a clinically applicable biomaterial approved by the FDA as a biocompatible and biodegradable polymer [90].

The degradation rate of biomaterials must be carefully considered before the fabrication of the target-specific tissue-engineered structure. If using the 3D scaffold with quickly degradation materials, there is a possibility of the mechanical property rapidly degrading after implantation in the body. In this regard, PCL has a great benefit as it can control the degradation rate by blending of different ratios of the polymer and copolymers [91,92]. The degradation mechanism of the PCL has a bulk erosion process by hydrolysis, and, in this process, PCL does not release toxic components [93,94]. Because of these convenient advantages, PCL is actively utilized as various bioprinting materials.

#### 3.2.2. Polylactic Acid

Polylactide or polylactic acid (PLA) is most widely used for creating tissue-engineered architecture [95]. Also, PLA has been approved by the USA FDA for human clinical applications. The PLA has been used as a biomaterial for frequency 3D bioprinting because of its readily available thermoplastic properties [96]. Although there are differences depending on molecular weight (MW), PLA has relatively high mechanical properties, with an approximate tensile modulus of 3 GPa and tensile strength of 50–70 MPa [97]. The MW has a significant effect on biodegradability, but high-MW PLA is likely to cause inflammation and infection in vivo [98]. Therefore, before 3D bioprinting, the MW property must be considered for the mechanical properties of the target tissue.

#### 3.2.3. Polyglycolic Acid

Polyglycolic acid (PGA) is a thermoplastic material with a high melting point and glass transition temperature, and it is more acidic and hydrophilic than PLA [99]. In addition, it is used as a surgical suture fiber because of its high mechanical strength and biocompatibility [100]. In the field of tissue engineering, solvent casting and compression molding are used to create PGA-based porous scaffolds [56]. However, PGA requires precise control as it is highly sensitive to degradation. Additionally, glycolic acid produced during the biodegradation process can be absorbed into the body, but the increased acid concentrations in the surrounding tissues may cause tissue damage.

#### *3.3. Functional Polymer*

As mentioned above, generally, biocomposite inks are classified as natural- and synthetic-based polymers, and these have been attempted to be used for complex and cell-compatible 3D constructs as tissue-engineering approaches [101]. Among them, hydrogel-type inks (i.e., alginate, collagen, dECM ink) have been considered as attractive materials because these can provide an optimized environment to a living cell. However, to be suitable for 3D bioprinting, these hydrate materials require adequate rheological properties to keep shape during bioprinting and must have cross-linking abilities, allowing to retain the 3D structure fidelity after bioprinting. Recently, the importance of versatile bioink materials in the field of tissue engineering has led to the development of functional polymers with improved biocompatibility, rheological behavior, and mechanical properties [102]. In this section, functional polymers that improve the bioprinting stability and fidelity when combined with nanocellulose biomaterials are introduced.

Nanocellulose refers to cellulosic nanomaterials, including cellulose nanocrystals (CNC) and cellulose nanofibrils (CNF) [103–105]. Gary Chinga-Carrasco et al. developed printable ink with bagasse, which is an underutilized agro-industrial residue [106]. This functional polymer has demonstrated non-cytotoxicity, stable bioprinting property, and shape fidelity, as well as potential that a low-value agro-industrial residue (bagasse) can be converted into a high-value product as disposable bioinks for 3D bioprinting. However, further evaluation is required for clinical applications. Kajsa Markstedt et al. developed functional bioink that combines the outstanding shear thinning properties of nanofibrillated cellulose (NFC) with the alginate [105]. The nanocellulose-based bioink enables fidelity bioprinting of 2D structure as well as 3D construct, which is anatomically the shape of a human ear and sheep meniscus. Also, nanocellulose-based bioink exhibits excellent cell viability. Therefore, these functional polymers using cellulose nanofibrils have shown promising potential as 3D bioprinting materials.

#### **4. Recent Design Approaches for Engineering Tubular Structures**

Tubular tissues and organs, such as the gastrointestinal tract, urinary tract, and respiratory tract, exist everywhere in the human body, serving major functions, including distributing fluids and air through the organs [107]. Almost all of the tubular tissues have a multilayer cellular structure from the innermost to the outermost, and the inner structure has an endothelium cell layer [6]. There are no existent pioneering fabrication techniques to fabricate tubular-shaped tissues and organs in the field of tissue engineering [5]. However, various fabrication methodologies have been suggested for constructing tubular structures with mimicking the inherent multilayer cellular constructs.

Traditional methods, such as casting, cell sheet assembly, and dip coating, have been attempted to create the tubular structures. The casting method creates the tubular structure by the biomaterials filling in the sacrificial mold and then demolding after appropriate chemical processes, such as gelation or crosslinking (Figure 2a). This method was proposed in 1986 by Weinberg and Bell, who made artificial vascular structures using the collagen-containing fibroblast and smooth muscle cells [23]. Since then, cell sheet assembly technology has been reported for reproducing hierarchical multi-layered cellular structures (Figure 2b) [20–22]. This method has been facilitated for creating the multilayer tubular structure by rolling on the rod using the stacked monolayer fabricated by biological functional materials containing extracellular matrix and target-specific cell components. The dip-coating method can also produce multiple tubular structures using rods by repeatedly dipping in the hydrogel and cross-linker agent (Figure 2c) [108–110]. These traditional methods have shown the promising ability to mimicking the cellular arrangement of native tubular tissues. However, there has been an unmet challenge in implementing physiological and mechanical properties suitable to tissue-specific complex environments. Also, these methods have unavoidable hurdles for fabricating shape-free forms and controllable structures. To overcome these challenges, hybrid-type technology, combining the traditional method with 3D bioprinting, has been tried to fabricate a free-form tubular structure [111]. Hybrid-type approaches have shown the possibility of creating a free-form tubular construct, although the structure has not been directly printed.

range of biomaterials. In this article, we summarized and classified several categories of the

rod supporting-, support bath-, direct bioprinting) (Table 1).

**Figure 2.** Schematic illustrations of the traditional methods of (**a**) casting; (**b**) cell sheet assembly; (**c**) dip coating and the extrusion-based 3D (**d**) co-axial; (**e**) kenzan method; (**f**) rod supporting; (**g**) support bath-based; and (**h**) direct bioprinting for fabricating tubular structures. **Figure 2.** Schematic illustrations of the traditional methods of (**a**) casting; (**b**) cell sheet assembly; (**c**) dip coating and the extrusion-based 3D (**d**) co-axial; (**e**) kenzan method; (**f**) rod supporting; (**g**) support bath-based; and (**h**) direct bioprinting for fabricating tubular structures.

The co-axial bioprinting method is capable of creating a tubular structure using a core–shell nozzle that is capable of extruding two or more biomaterials (Figure 2d). Several research teams have been employing this method to print complex tubular structures with biocomposite inks. Gao et al. developed the printable hybrid bioink containing a mixture of vascular tissue-derived decellularized extracellular matrix (VdECM), alginate, and human umbilical vein endothelial cells [59]. Subsequently, they fabricated a perfusable multilayer blood vessel by co-axial bioprinting [89]. Yongxiang Luo et al. printed a 3D porous scaffold with regular macropores and a network of a controllable hollow structure as an embedded vasculature-like system using co-axial bioprinting [114]. This method can allow not only the building of a hollow construct with functional biological 3D bioprinting technology has been emerging as a promising approach to facilitate complex structures and spatial cell positioning in tubular tissue engineering [112,113]. Among the various 3D bioprinting techniques (e.g., extrusion-based, ink-jet, laser-assisted, stereolithography-based 3D bioprinting), extrusion-based 3D bioprinting has been one of the most utilized for creating the tubular structure because it is relatively convenient for the installation of the system and availability of a wide range of biomaterials. In this article, we summarized and classified several categories of the extrusion-based 3D bioprinting for building a multilayer tubular structure (co-axial-, kenzan method-, rod supporting-, support bath-, direct bioprinting) (Table 1).

components but is also capable of fabricating permeable vascular-embedded 3D constructs. Moreover, it can manufacture small-diameter vascular structures with endothelial and smooth muscle layers, as well as being able to print long-length warping vascular structures with a minimal amount of time. However, this method has limitations in terms of making the anatomical bifurcate structure and stacking hierarchical constructs. Kenzan method bioprinting was invented by Koich Nakayama. It can create a high-density cellular structure by locating cell spheroids on a fine needle array (Figure 2e) [115–118]. The main principle of this method uses the natural and intrinsic feature of cell-to-cell self-aggregation. Recently, this research group fabricated the esophagus-like tubular structure without scaffold using the multicellular spheroids that maturated during several periods in the bioreactor to create the rigid organoids. The co-axial bioprinting method is capable of creating a tubular structure using a core–shell nozzle that is capable of extruding two or more biomaterials (Figure 2d). Several research teams have been employing this method to print complex tubular structures with biocomposite inks. Gao et al. developed the printable hybrid bioink containing a mixture of vascular tissue-derived decellularized extracellular matrix (VdECM), alginate, and human umbilical vein endothelial cells [59]. Subsequently, they fabricated a perfusable multilayer blood vessel by co-axial bioprinting [88]. Yongxiang Luo et al. printed a 3D porous scaffold with regular macropores and a network of a controllable hollow structure as an embedded vasculature-like system using co-axial bioprinting [114]. This method can allow not only the building of a hollow construct with functional biological components but is also capable of fabricating permeable vascular-embedded 3D constructs. Moreover, it can manufacture small-diameter vascular structures with endothelial and smooth muscle layers, as well as being able to print long-length warping vascular structures with a minimal amount of time. However, this method has limitations in terms of making the anatomical bifurcate structure and stacking hierarchical constructs.



Kenzan method bioprinting was invented by Koich Nakayama. It can create a high-density cellular structure by locating cell spheroids on a fine needle array (Figure 2e) [115–118]. The main principle of this method uses the natural and intrinsic feature of cell-to-cell self-aggregation. Recently, this research group fabricated the esophagus-like tubular structure without scaffold using the multicellular spheroids that maturated during several periods in the bioreactor to create the rigid organoids.

The rod supporting bioprinting method produces a hollow construct by dispensing printable biomaterials on a rotating rod (Figure 2f). The rotating rod is provided as temporary support to the printed biomaterial for keeping a 3D shape and is removed when the printed structure is considered to be self-supporting. Sang-Woo Bae et al. printed the artificial tracheal structure with a synthetic polymer (i.e., PCL) and cell-laden bioink (epithelial cells and bone-marrow stem cells) by sequential extruding on the rotating rod [119]. Qing Gao et al. fabricated a hydrogel-based vascular structure with multilevel fluidic channels using a combination with co-axial bioprinting [31]. The main advantage of this bioprinting method is the manufacturing ability of the self-supporting multi-layer hollow structure with target-specific cell components though sequential bioprinting using above two or more printable biomaterials, such as polymer-based and cell-laden bioink. However, there is a disadvantage in that the figuration of the printed structure is dependent on the rotating rod shape.

The support bath-based bioprinting method refers to using a thermo-sensitivity gel bath or sacrificial materials for supporting the bioprinting biomaterials (Figure 2g). This method obtains a self-supporting structure, keeping the reversible condition for removing the biomaterials after bioprinting in the gel bath. Thomas J. Hinton et al. introduced this bioprinting method as the freeform reversible embedding of suspended hydrogel (FRESH), which uses this technique to print the biomimetic section of the human right coronary arterial tree with alginate-based bioink [120]. The research team, Tal Dvir et al. printed the endothelial cell-laden hydrogel to create the blood vessel-embedded cardiac tissue of the rabbit scale in the supporting bath with an aqueous solution containing sodium alginate, xanthan gum, calcium carbonate [121]. This method enables 3D bioprinting of the hydrate biomaterials, including alginate, collagen, and fibrin. Also, it can create complex 3D anatomical architectures, including branched coronary arteries, embedded vascular system organoids, etc. These results have demonstrated the potential of the approach for engineering personalized tissues and the bioprinting of patent-tailored biochemical microenvironment.

Direct bioprinting can be used to stack the printable biomaterials layer-by-layer for building the 3D structure (Figure 2h). In order to use this bioprinting technique, biomaterials must have sufficient solidity properties to maintain the 3D structure. For bioprinting hydrate bioink, in particular, it is essential to consider the rheological properties of materials during the bioprinting process. Anthony Atala et al. used direct bioprinting to build a 3D urethra tubular scaffold with a polymeric framework and cell-laden fibrin hydrogel [122]. The polymeric framework consisting of the PCL and PLCL (polylactic acid-co-ε-caprolactone) stably supports the printed hydrogel bioink for the desirable fabrication of the 3D tubular structure. Recently, Yifei Jin et al. developed self-supporting hydrogel by mixing laponite nanoclay and then successfully printed sturdy architecture without supporting the gel bath and polymeric structures [123].

#### **5. Application of the 3D Printed Tubular-Organs with Various Biocomposite Inks**

#### *5.1. Esophagus*

The esophagus is one of the gastrointestinal tracts and a 20–25 cm hollow structure, connecting the oropharynx and the stomach [124–126]. It allows the transport of food to the stomach by peristalsis and contractions of the muscle layer. Because the esophagus has a complex hierarchical structure (mucosa, submucosa, muscle layers), this must be considered when fabricated using engineering approaches [127–129].

Every year, 5000 to 10,000 patients are diagnosed with an esophageal disease requiring partial repair or full-thickness circumferential replacement, such as esophageal cancer, malignancy, congenital long-gap atresia, and esophageal achalasia. The strategy of the esophageal treatment is normally a gastric pull-up or autotransplantation using intestine or skin. In the case of autograft, using the gastrointestinal tract is an unavoidable strategy to achieve circumferential full-thickness repair since there is no substitute for esophagus tissue. However, although the autograft might allow the transfer of liquids or solid matter, complete restoration of the native tissue is compromised. Therefore, several approaches using 3D bioprinting technology have been researched to achieve the esophageal substitution, which replicates primary histological features of hierarchical cellular structures (Table 2).


**Table 2.** The 3D bioprinting technique and biocomposite ink for the esophageal tubular structure.

Yosuke Takeoka et al. developed a scaffold-free biomimetic structure for the regeneration of the esophagus using the kenzan method bioprinting (Figure 3) [115]. This team used the maturated cell spheroids of the normal human dermal fibroblasts (NHDFs), human esophageal smooth muscle cells (HESMCs), human bone marrow-derived mesenchymal stem cells (MSCs), and human umbilical vein endothelial cells (HUVECs) to print the tubular multicellular structures. Mechanical and histochemical assessment of the printed esophagus-like tubular structure had been done with the content ratio of those cell sources. The high proportion of mesenchymal stem cell groups tended to give greater mechanical strength as well as the expressed α-smooth muscle actin and vascular endothelial growth factor (VEGF) on immunohistochemistry. After bioprinting of esophagus-like scaffold-free tubular structures with demonstrated multicellular proportion, it was matured in bioreactor and then transplanted into rats as esophageal grafts. The esophageal grafts were implanted between the stomach and esophagus with a silicone tube. Results showed the grafts were maintained in vivo for 30 days, and the epithelium extended and covered the inner lumen and was able to pass food as well. The epithelialization of the inner surface of the esophageal lumen should be considered as the key regenerative factor because it must be done postoperatively with non-sterilized solid matter. In this respect, this research result has been promising as a potential substitute for esophageal transplantation using bioprinting. *Bioengineering* **2020**, *7*, x FOR PEER REVIEW 11 of 24 Cell-seeded (hMSCs) tubular frameworks were maturated in the customized bioreactor system, and shear stress of the 0.1 dyne/cm<sup>2</sup> flow-induced with a pattern of 1 min/2 min for engagement/resting was applied to invigorate the frameworks. In comparison results of the histological analysis in the circumferential esophageal defects in a rat model from bioreactor cultivation and the omentumcultured groups, both the groups showed over 80% mucosal regeneration without a fistula. The follow-up study by this research group suggested the further extended bioreactor culture system that could apply the different mechanical stimuli and biochemical reagents at the inner lumen and outside of the scaffold [16]. Among the mechanical stimuli, in particular, the intermittent shear flow by hydrostatic pressure and the shear stress by flow media were relevant to improving efficacy for differentiation of the epithelial and muscle lineage compared to steady shear flow.

**Figure 3.** Schematic illustration of the esophageal tubular structure by kenzan method bioprinting [115]. **Figure 3.** Schematic illustration of the esophageal tubular structure by kenzan method bioprinting [115].

**Table 2.** The 3D bioprinting technique and biocomposite ink for the esophageal tubular structure. **3D Bioprinting Technique Biocomposite Ink Reference** Kenzan method bioprinting Cell spheroids with human dermal fibroblasts, human esophageal smooth muscle cells, human bone marrow-derived mesenchymal stem cells, human umbilical vein endothelial cells [115] Rod supporting bioprinting and electrospinning Polyurethane (PU), polycaprolactone (PCL) [15] Rod supporting bioprinting and Polycaprolactone (PCL) [14] In Gul Kim et al. employed both techniques of the supporting rod bioprinting and electrospinning to build the enhanced tubular structure with two-layer [15]. That study evaluated the 3D printed esophageal graft and the effect of bioreactor cultivation on cell maturity for muscle regeneration and epithelialization. To fabricate the tubular framework, the membrane was manufactured by electrospun polyurethane (PU) on the rotating rod (diameter: 2 mm), and then to improve the mechanical stability, the PCL strand was squeezed using the extrusion-based system. Cell-seeded (hMSCs) tubular frameworks were maturated in the customized bioreactor system, and shear stress of the 0.1 dyne/cm<sup>2</sup> flow-induced with a pattern of 1 min/2 min for engagement/resting was applied to invigorate the frameworks. In comparison results of the histological analysis in the circumferential esophageal defects in a rat model from bioreactor cultivation and the omentum-cultured groups, both the groups

reinforcing ring on the rod, a thin PCL layer was formed by electrospinning to form a nano-structured

electrospinning

showed over 80% mucosal regeneration without a fistula. The follow-up study by this research group suggested the further extended bioreactor culture system that could apply the different mechanical stimuli and biochemical reagents at the inner lumen and outside of the scaffold [16]. Among the mechanical stimuli, in particular, the intermittent shear flow by hydrostatic pressure and the shear stress by flow media were relevant to improving efficacy for differentiation of the epithelial and muscle lineage compared to steady shear flow.

Similarly, Eun-Jae Chung et al. utilized both supporting rod bioprinting and electrospinning and developed an esophageal scaffold reinforced by a 3D-printed PCL ring [14]. After bioprinting the reinforcing ring on the rod, a thin PCL layer was formed by electrospinning to form a nano-structured tubular structure. The printed tubular structures were wrapped into the omentum of rats for 2 weeks and then orthotopically transplanted for a circumferential esophageal defect. When macroscopically observed in the in vivo study, no fistulas, necrosis of the anastomosis, or abscess formation was found in the surrounding of the operating sections.

Maohua Lin et al. used direct bioprinting to a fabricated esophageal tubular stent with spiral patterns that applied the optimized design by computational simulation [130]. The printed esophageal tubular stent consisted of a mixed biodegradable polymer of medical-grade thermal polyurethane (TPU) and PLA in optimum proportion to achieve appropriate mechanical stiffness and flexibility. The group of the tubular stent with 10% PLA was investigated as a remarkable condition in the anti-migration force, self-expansion force, and human esophagus epithelial cell viability.

#### *5.2. Blood Vessel*

Blood vessels are components of the circulatory systems with hollow tube structures in the tissue and organs [131]. These transport blood cells, oxygen, and nutrients throughout the body and receive the CO<sup>2</sup> and waste from the metabolic activity of the peripheral cells and tissues. Blood vessels are divided into arteries, veins, and capillaries according to their structural characteristic and biological functions. In general, the artery and vein walls consist of three layers: tunica intima (squamous endothelium), tunica media (smooth muscle cells), and tunica adventitia (fibrous collagen) [132,133].

Generally, blood vessel disorder refers to the hardening, enlargement, and narrowing of arteries and veins. These health problems trigger arterial diseases, which can cause death, such as coronary artery heart disease, cardiovascular disease, peripheral artery disease. Worldwide annual mortalities related to cardiovascular disease are expected to rise to 23.3 million by 2030 [134]. To date, revascularization strategies have included the stent, surgical bypass grafting, and angioplasty. Also, commercialized off-the-shelf alternatives, such as polytetrafluoroethylene (PTFE), gore-tex, and dacron, are also being proved clinically effective when replacing large-diameter vessels (≥6 mm). When using a small-diameter (≤6 mm) vascular graft, however, it has caused a thrombosis event with the closing lumen and the lack of long-term patency as well as intimal hyperplasia. Considering the limitations of current vascular grafts, tissue-engineered vascular graft (TEVG) has been developed using 3D bioprinting technology. In particular, for the clinical applications of the TEVG, anti-thrombosis and long-term patency overcome the essential issues. Recently, to achieve this goal, several interesting studies have reported generating a tubular structure with a biochemical component capable of physiological remodeling (Table 3).

Gao et al. successfully fabricated the tubular bio-blood-vessel (BBV) with hybrid bioink (a mixture of VdECM and alginate) using the versatile 3D co-axial bioprinting method (Figure 4) [59]. The VdECM/alginate hybrid bioink containing the atorvastatin-loaded PLGA microspheres (APMS) and endothelial progenitor cells (EPCs) provides a favorable environment to promote the proliferation and neovascularization. When bioink encapsulating APMS/EPCs is used with a core–shell nozzle, the inner shell is filled with CaCl<sup>2</sup> solution (CPF127) for ionically crosslinking by releasing the calcium ion. The co-axial cell-printed tubular structure has been estimated in an ischemia model in nude mouse hind limb. It has induced an increased rate of neovascularization and the remarkable regeneration of ischemic limbs. The noteworthy point is that this research has achieved the creation of tubular structures with a broad range of diameters by controlling the core–shell nozzle. In addition, functional

encapsulation of the cell/drug-laden bioink has shown potential for expansion as the printed BBV with the carrier that enables anti-thrombosis and long-term patency.


**Table 3.** The 3D bioprinting technique and biocomposite ink for the vascular tubular structure.

**Figure 4.** Schematic illustration and structural images of the vascular tubular structure by co-axial bioprinting [59]**. Figure 4.** Schematic illustration and structural images of the vascular tubular structure by co-axial bioprinting [59].

**Table 3.** The 3D bioprinting technique and biocomposite ink for the vascular tubular structure. **3D Bioprinting Technique Biocomposite Ink Reference**  Co-axial bioprinting Vascular-tissue-derived decellularized extracellular matrix (VdECM) with alginate [59] Rod supporting bioprinting Fibrinogen and gelatin [135] Rod supporting bioprinting Sebastian Freeman et al. developed fibrin-based vascular constructs using rod supporting bioprinting [135]. The printable bioink consists of the fibrinogen with gelatin to achieve the desired shear-thinning property for self-standing. Unprintable fibrinogen was used as a printable biomaterial by blending the favorable rheological properties with heat-treated gelatin. During two months of the cultures after bioprinting, the burst pressure of the tubular structure reached 1110 mm Hg, and the

bioprinting Alginate and gelatin slurry support bath [120]

Direct bioprinting Pluronic 127 and gelatin methacrylate (GelMA) [138] Quing Gao et al. fabricated hydrogel-based perfusable vascular structure with multilevel fluidic channels in tubular tissue approaches by using the 3D bioprinting that combined with co-axial and rod supporting bioprinting [31]. Partially cross-linked hollow alginate containing the fibroblasts and

Photocrosslinkable bioelastomer prepolymers ink (dimethyl itaconate. 1,8-ictanediol and triethyl citrate) and carbomer gel bath

[137]

Co-axial bioprinting and

Support bath-based

Co-axial printing and support bath-based bioprinting

remarkable improvement of the tensile mechanical properties was achieved in both the circumferential and axial elastic moduli.

Sang Jin Lee et al. combined rod supporting bioprinting and electrospinning for mechanical robustness and to build multi-layered structures using synthetic polymers (e.g., PCL) [136]. To induce potent angiogenic activity, the printed tubular structure was coated with polydopamine (PDA) and vascular endothelial growth factor (VEGF) on the surface. The coated-PDA layer enhanced the ability of the hydrophilicity; it also remarkably increased the vascular cell proliferation and angiogenic differentiation during in vivo/in vitro.

Quing Gao et al. fabricated hydrogel-based perfusable vascular structure with multilevel fluidic channels in tubular tissue approaches by using the 3D bioprinting that combined with co-axial and rod supporting bioprinting [31]. Partially cross-linked hollow alginate containing the fibroblasts and smooth muscle cells were extruded through a two co-axial nozzle and then printed along with a rotating supporting rod. The printed tubular structure exhibited a sufficiently strong mechanical strength (ultimate strength: 0.148 MPa) for the implantation due to the fusion of adjacent crosslinking reaction. Encapsulation of the fibroblast in the tubular structure showed over 90% survival within 1 week in vivo. This research has shown the ability to directly fabricating a perfusable vessel-like structure by cell-laden biomaterials through a coupled co-axial bioprinting and rod supporting method.

Thomas J. Hinton et al. 3D bioprinted a more complex structure than the perfusable arterial tree with alginate bioink and embedded it in the gelatin slurry support bath using a support bath-based bioprinting technology [120]. Perfusion structures mimicking a portion of the right arterial tree obtained through MRI data were printed in multiple branches with 3D tortuosity. Houman Savoji et al., similarly, printed vascular tubes (using core–shell nozzle) via freeform reversible embedding of photocrosslinkable bioelastomer prepolymers within a carbomer hydrogel bath by co-axial bioprinting and support bath-based bioprinting [137]. This tubular tissue-engineered approach to create a further advanced tubular structure made the significant achievement of mechanical robustness and recreated complex 3D anatomical architectures.

Kolesky et al. fabricated embedded vasculature constructs, repleted with multiple types of cells and an extracellular matrix (ECM), using direct bioprinting [138]. An aqueous fugitive ink composed of pluronic 127 was used for easy printing and removing under mild conditions to create vascular channels. In addition, gelatin methacrylate (GelMA) was used as a bulk matrix and cell carrier. After infilling and photopolymerizing the GelMA matrix on the fugitive pluronic 127 ink, the fugitive ink was removed by cooling the printed constructs below 4 ◦C, yielding open channels to fabricate the embedded vasculature constructs. Using this 3D bioprinting process, the potential of 3D vascular embedded constructs with human neonatal dermal fibroblasts (HNDFs) and human umbilical vein endothelial cells (HUVECs) was convincingly demonstrated. This result showed the development possibility for remodeling heterogeneous tissue constructs containing vasculature and multiple cell types.

#### *5.3. Trachea*

The trachea is a tubular structure in which the lower respiratory tract begins and refers to the pathway that begins immediately running between the larynx and the bronchi. The trachea is a composite tubular structure consisting of epithelium, basement membrane, connective tissue, smooth muscle, and cartilaginous layer. The tubular shape is about 2 cm in diameter and 11 cm in length with a flat posterior. The trachea acts as an airway to enter and exit the air during respiration. In addition, when debris, such as dust, enters into the trachea with air, it functions to move and remove the debris using ciliary movement and mucus [139,140].

Tracheas are becoming increasingly damaged due to severe environmental pollution. In addition, damage to the trachea has become a serious problem due to the increased use of ventilators for the treatment of patients [141]. In order to solve this problem, the transplantation of donor tissue from a deceased person to an injured organ has been reported. However, not only is it difficult to obtain donor

tissue but even if it is obtained, there is a disadvantage that it can be transplanted through a complex pretreatment process over a long period of time [142]. The method of treating organ damage depends on the extent and length of the involvement of the site of injury. End-to-end anastomosis is a common treatment for circumferential injuries. However, end-to-end anastomosis has disadvantages, such as continuous endotracheal intubation, rupture, or stenosis of anastomosis after surgery. In addition, end-to-end anastomosis cannot be applied if more than 50% of the trachea needs to be excised. Such cases are difficult to treat clinically [143,144]. Tissue engineering is an appropriate approach to solve these problems; in addition, recent advances in 3D bioprinting technology have enabled the production of more sophisticated and systematic artificial structures (Table 4) [145,146].


**Table 4.** The 3D bioprinting techniques and biocomposite inks for the tracheal tubular structure.

Daisuke Taniguchi et al. developed an artificial trachea using kenzan method bioprinting [147]. They assessed the circumferential tracheal replacement using scaffold-free trachea-like grafts generated from spheroids consisting of several types of cells—chondrocytes, endothelial cells, and mesenchymal stem cells—to build 3D structures. This artificial trachea from spheroids was matured in a bioreactor and transplanted into a rat. In the transplantation, they used silicone stents to prevent collapse. As a result, chondrogenesis and vasculogenesis could be observed in this artificial trachea.

Manchen Gao et al. printed a biodegradable reticular PCL scaffold with similar morphology to the rabbits' native trachea by direct bioprinting [148]. Chondrocytes were cultured in this 3D scaffold and conducted into the subcutaneous of nude mice. The scaffold showed the successful reconstruction and the proper supporting force to maintain the lumen as well as presenting remarkable cartilaginous properties both in vitro and in vivo.

Cheng-Tien Hsieh et al. fabricated a tissue-engineered trachea with structural similarity to the native trachea from water-based biocomposite ink at low temperature using direct bioprinting [149].

In that research, two kinds of water-based biodegradable polyurethanes with different physicochemical properties were used as biocomposite ink. The human MSCs were seeded into this tracheal construct, and then the construct was implanted in nude mice. After 6 weeks, the results showed dynamic compression moduli of the scaffolds that were 0.3–0.8 MPa under the force of 0.1–0.8 N, which was similar to the native trachea. It also confirmed gas-tightness by airflow test at positive and negative air pressures. Moreover, MSCs seeded in the tracheal scaffolds were grown into cartilage-like tissue. It expressed chondrogenic potential and secreted glycosaminoglycans (GAGs) and collagen after 14 days in vitro culture without any exogenous growth factors, such as bioactive factors or small molecular drugs. They showed that the tracheal scaffold of biocomposite inks and 3D bioprinting techniques might be used to fabricate personalized artificial tracheas for clinical applications. It also showed the possibility of incorporating exogenous growth factors in the water-based biocomposite ink to enhance the chondrogenesis of the MSCs.

Jae Yeon Lee et al. developed an artificial tracheal structure PCL framework by extrusion bioprinting and silicone band by direct bioprinting and rod-supporting bioprinting (Figure 5) [150]. In particular, the states of the PCL extrusion were precisely controlled to create dotted circular patterns so that the bellows framework had about 300 µm pores in the wall except for groove parts. Then, they used a rod supporting bioprinting to print ring-shaped bands into the outer grooves of the PCL framework using a medical-grade silicone elastomer. The PCL framework was put around the rotating rods and rotated. They proved the potential of this artificial scaffold to be applied immediately in emergencies. *Bioengineering* **2020**, *7*, x FOR PEER REVIEW 16 of 24

**Figure 5.** Structural images of the tracheal tubular structure by direct and rod-supporting bioprinting [150]. **Figure 5.** Structural images of the tracheal tubular structure by direct and rod-supporting bioprinting [150].

#### **6. Future Perspectives and Concluding Remark 6. Future Perspectives and Concluding Remark**

The biofabricated tissue-engineered tubular constructs require particular features of targetspecific mechanical properties, anatomical accuracy, autoimmune acceptance, long-term patency, and similar cell arrangement for creating and mimicking of native tissues. This represents a significant technical challenge, and, to date, clinically meaningful tubular structure bioprinting approaches have been reported, utilizing versatile additive manufacturing techniques and biocomposite inks. The biofabricated tissue-engineered tubular constructs require particular features of target-specific mechanical properties, anatomical accuracy, autoimmune acceptance, long-term patency, and similar cell arrangement for creating and mimicking of native tissues. This represents a significant technical challenge, and, to date, clinically meaningful tubular structure bioprinting approaches have been reported, utilizing versatile additive manufacturing techniques and biocomposite inks.

Various 3D bioprinting methodologies have emerged in the field of tissue engineering, and advanced technologies have been secured through rapid technological developments. Among them, the extrusion-based bioprinting system has been actively used to fabricate tubular structures with advantages of ease construction and flexibility in the use of various biomaterials. The fabrication of multi-layered tubular structures using supporting rods has been actively used, and approaches with more histologically close multi-cellular components using cell-spheroids have been developed. Coaxial bioprinting has also established itself as a promising approach that allows easy fabrication of freely adjustable perfusable tubular structures. In addition, this technology can be loaded with a variety of functional drugs, as well as cells that help bioenvironment cues, which are used in expandability platforms, such as vascular-embedded organoids and drug-screening devices. Support bath-based bioprinting has the advantage of being able to produce free- and multi-branched selfsupporting forms using hydrate materials. Despite the breakthrough in 3D bioprinting technology, artificial tubular structures of the Various 3D bioprinting methodologies have emerged in the field of tissue engineering, and advanced technologies have been secured through rapid technological developments. Among them, the extrusion-based bioprinting system has been actively used to fabricate tubular structures with advantages of ease construction and flexibility in the use of various biomaterials. The fabrication of multi-layered tubular structures using supporting rods has been actively used, and approaches with more histologically close multi-cellular components using cell-spheroids have been developed. Co-axial bioprinting has also established itself as a promising approach that allows easy fabrication of freely adjustable perfusable tubular structures. In addition, this technology can be loaded with a variety of functional drugs, as well as cells that help bioenvironment cues, which are used in expandability platforms, such as vascular-embedded organoids and drug-screening devices. Support bath-based bioprinting has the advantage of being able to produce free- and multi-branched self-supporting forms using hydrate materials.

esophagus, blood vessels, and trachea still face challenges for application as clinical substitutes [151– 153]. This is because these tissues are exposed to high clinical needs, such as contraction, expansion by peristalsis, and blood pressure. In particular, the esophagus and trachea inevitably contact with external contaminants, such as liquid, food, and air, after insertion of the artificial tubular structure, thus hindering the growth of the functional endothelial layer. In addition, in the case of blood vessels, the absence of the endothelium layer may induce thrombus and stenosis. Therefore, pre-maturation, such as omentum culture, of bioreactor is considered a significant factor in the growth of artificial tissues before surgical approaches. In addition, many researchers are working to rebuild tubular organs, such as the stomach, intestine [154], bladder [9], and urethra [122], as well as the mentioned ones in this review. Thus, future developments of artificial tubular tissue should simultaneously entail the promising benefits provided by 3D bioprinting as well as the development of functional biocomposite inks and optimal cell culture techniques for target-tissues. **Author Contributions:** investigation, H.-J.J. and H.N.; writing—original draft preparation, H.-J.J. and H.N.; writing—review and editing, H.-J.J. and H.N.; supervision, J.J. and S.-J.L.; All authors have read and agreed to the published version of the manuscript. Despite the breakthrough in 3D bioprinting technology, artificial tubular structures of the esophagus, blood vessels, and trachea still face challenges for application as clinical substitutes [151–153]. This is because these tissues are exposed to high clinical needs, such as contraction, expansion by peristalsis, and blood pressure. In particular, the esophagus and trachea inevitably contact with external contaminants, such as liquid, food, and air, after insertion of the artificial tubular structure, thus hindering the growth of the functional endothelial layer. In addition, in the case of blood vessels, the absence of the endothelium layer may induce thrombus and stenosis. Therefore, pre-maturation, such as omentum culture, of bioreactor is considered a significant factor in the growth of artificial tissues before surgical approaches. In addition, many researchers are working to rebuild tubular organs, such as the stomach, intestine [154], bladder [9], and urethra [122], as well as the mentioned ones in this review. Thus, future developments of artificial tubular tissue should simultaneously entail the promising benefits provided by 3D bioprinting as well as the development of functional biocomposite inks and optimal cell culture techniques for target-tissues.

**Author Contributions:** Investigation, H.-J.J. and H.N.; writing—original draft preparation, H.-J.J. and H.N.; writing—review and editing, H.-J.J. and H.N.; supervision, J.J. and S.-J.L.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by National Research Foundation of Korea (NRF) grant funded by the Ministry of Education, Science, and Technology (NRF 2016R1D1A1B01006658) and by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2015R1A6A3A04059015) and by the MSI T(Ministry of Science and ICT), Korea, under the ICT Consilience Creative program (IITP-2019-2011-1-00783) supervised by the IITP (Institute for Information and communications Technology Planning and Evaluation).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Advantages of Additive Manufacturing for Biomedical Applications of Polyhydroxyalkanoates**

**Alberto Giubilini <sup>1</sup> , Federica Bondioli <sup>2</sup> , Massimo Messori <sup>3</sup> , Gustav Nyström 4,5 and Gilberto Siqueira 4,\***


**Abstract:** In recent years, biopolymers have been attracting the attention of researchers and specialists from different fields, including biotechnology, material science, engineering, and medicine. The reason is the possibility of combining sustainability with scientific and technological progress. This is an extremely broad research topic, and a distinction has to be made among different classes and types of biopolymers. Polyhydroxyalkanoate (PHA) is a particular family of polyesters, synthetized by microorganisms under unbalanced growth conditions, making them both bio-based and biodegradable polymers with a thermoplastic behavior. Recently, PHAs were used more intensively in biomedical applications because of their tunable mechanical properties, cytocompatibility, adhesion for cells, and controllable biodegradability. Similarly, the 3D-printing technologies show increasing potential in this particular field of application, due to their advantages in tailor-made design, rapid prototyping, and manufacturing of complex structures. In this review, first, the synthesis and the production of PHAs are described, and different production techniques of medical implants are compared. Then, an overview is given on the most recent and relevant medical applications of PHA for drug delivery, vessel stenting, and tissue engineering. A special focus is reserved for the innovations brought by the introduction of additive manufacturing in this field, as compared to the traditional techniques. All of these advances are expected to have important scientific and commercial applications in the near future.

**Keywords:** polyhydroxyalkanoates; scaffolds; biomedicine; additive manufacturing; 3D printing; drug delivery; vessel stenting; tissue engineering

#### **1. Introduction**

The term "biopolymer" is nowadays very common and widely spread in different fields of application. However, it is sometimes improperly used, due to the fact that there is not a brief and comprehensive definition of this word. To clarify the meaning of "biopolymer", it is important to define the concepts of "bio-based" and "biodegradable", and if the former is strictly connected with the origin of the material, at the opposite, the latter is related to its end-of-life.

A material can be defined as bio-based if it derives in whole or in part from biomass resources, i.e., organic materials that are renewable [1].

A material can be properly defined as biodegradable if it can be used as a carbon source by microorganisms and converted safely into CO2, biomass and water [2]. Besides, if the material undergoes a biodegradation and a physical disintegration level of at least 90%, in less than six months, then it can also be defined as "compostable" [3].

**Citation:** Giubilini, A.; Bondioli, F.; Messori, M.; Nyström, G.; Siqueira, G. Advantages of Additive Manufacturing for Biomedical Applications of Polyhydroxyalkanoates. *Bioengineering* **2021**, *8*, 29. https://doi.org/10.3390/ bioengineering8020029

Received: 11 November 2020 Accepted: 16 February 2021 Published: 23 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Hence, the family of biopolymers can be divided into three main groups: 1. Biopolymers coming from renewable resources but not being biodegradable, e.g.,

90%, in less than six months, then it can also be defined as "compostable" [3]. Hence, the family of biopolymers can be divided into three main groups:

*Bioengineering* **2021**, *8*, 29 2 of 31

1. Biopolymers coming from renewable resources but not being biodegradable, e.g., bio-based polyethylene terephthalate (bio-PET), bio-based polypropylene (bio-PP), and bio-based polyethylene (bio-PE); bio-based polyethylene terephthalate (bio-PET), bio-based polypropylene (bio-PP), and bio-based polyethylene (bio-PE); 2. Biopolymers coming from not-renewable resources but being biodegradable, e.g.,

if the material undergoes a biodegradation and a physical disintegration level of at least

	- 3. Biopolymers coming from renewable resources and being biodegradable, e.g., polyhydroxyalkanoate (PHA), poly(lactic acid) (PLA), and polybutylene succinate (PBS). hydroxyalkanoate (PHA), poly(lactic acid) (PLA), and polybutylene succinate (PBS). In this review, the PHA family is taken into consideration, and particular interest is

In this review, the PHA family is taken into consideration, and particular interest is reserved to its application in the biomedical field. Since the beginning of the twenty-first century, an increasing number of scientific studies and clinical trials have been published about PHA medical devices for different final applications, such as tissue engineering, drug delivery, or as vascular stents [4]. Therefore, first this review is aimed to present and discuss the results obtained with PHA and traditional techniques, like solvent casting, phase separation, salt leaching, or electrospinning. Furthermore, a great importance is given to the introduction of additive manufacturing in this research field, and particularly to the innovations and advantages introduced by 3D printing, which allowed us to overcome some of the greatest limitations of traditional approaches. For example, thanks to additive manufacturing, it was possible to obtain a finer control over the porosity, a true development of the devices in all three dimensions, and even the reproduction of complex structures, which are able to mimic natural tissues and which are highly tailored to the physical requirements of each patient [5,6]. Finally, this review is concluded with a discussion, in the authors' opinion, of the most likely future biomedical perspectives for this promising class of biopolymer, and of the new targets that can be achieved, thanks to 3D printing, in a new way of considering medicine, with a high customization of medical care. In Figure 1 shows a schematic representation of the overall topic and structure of this review work. reserved to its application in the biomedical field. Since the beginning of the twenty-first century, an increasing number of scientific studies and clinical trials have been published about PHA medical devices for different final applications, such as tissue engineering, drug delivery, or as vascular stents [4]. Therefore, first this review is aimed to present and discuss the results obtained with PHA and traditional techniques, like solvent casting, phase separation, salt leaching, or electrospinning. Furthermore, a great importance is given to the introduction of additive manufacturing in this research field, and particularly to the innovations and advantages introduced by 3D printing, which allowed us to overcome some of the greatest limitations of traditional approaches. For example, thanks to additive manufacturing, it was possible to obtain a finer control over the porosity, a true development of the devices in all three dimensions, and even the reproduction of complex structures, which are able to mimic natural tissues and which are highly tailored to the physical requirements of each patient [5,6]. Finally, this review is concluded with a discussion, in the authors' opinion, of the most likely future biomedical perspectives for this promising class of biopolymer, and of the new targets that can be achieved, thanks to 3D printing, in a new way of considering medicine, with a high customization of medical care. In Figure 1 shows a schematic representation of the overall topic and structure of this review work.

**Figure 1.** Schematic representation of the production, technological transformation, and biomedical applications of polyhydroxyalkanoate (PHA)-based devices. **Figure 1.** Schematic representation of the production, technological transformation, and biomedical applications of polyhydroxyalkanoate (PHA)-based devices.

The methodology carried out for the analysis of the literature started with searching published reviews on two of the most widespread databases, i.e., Scopus and ScienceDirect. Keywords selected for the literature search included PHA, additive manufacturing, biomedical application, biopolymer medical device, and PHA biosynthesis. These reviews were scanned, all parts related to PHA were highlighted, and the cited original research articles were acquired. After that, all references' abstracts were examined, and a first category clustering was performed according to this filtering system: (1) PHA production; (2) traditional PHA medical devices (solvent casting, salt leaching, thermally induced phase separation, non-solvent induced phase separation, emulsification, and electrospinning); (3) innovative PHA medical devices (Direct Ink Writing, Fused Deposition Modeling, Selective Laser Sintering, and Computer Aided Wet-Spinning). Afterwards, a second classification was implemented, to order all the references in accordance with the final medical application: (1) drug delivery; (2) vessel stenting; (3) bone tissue engineering, and (4) cartilage tissue engineering. Eventually, a combination of the two former groups was completed, and this synthesis was used as starting point for the manuscript development. The methodology carried out for the analysis of the literature started with searching published reviews on two of the most widespread databases, i.e., Scopus and ScienceDirect. Keywords selected for the literature search included PHA, additive manufacturing, biomedical application, biopolymer medical device, and PHA biosynthesis. These reviews were scanned, all parts related to PHA were highlighted, and the cited original research articles were acquired. After that, all references' abstracts were examined, and a first category clustering was performed according to this filtering system: (1) PHA production; (2) traditional PHA medical devices (solvent casting, salt leaching, thermally induced phase separation, non-solvent induced phase separation, emulsification, and electrospinning); (3) innovative PHA medical devices (Direct Ink Writing, Fused Deposition Modeling, Selective Laser Sintering, and Computer Aided Wet-Spinning). Afterwards, a second classification was implemented, to order all the references in accordance with the final medical application: (1) drug delivery; (2) vessel stenting; (3) bone tissue engineering, and (4) cartilage tissue engineering. Eventually, a combination of the two former groups was completed, and this synthesis was used as starting point for the manuscript development.

#### **2. PHA: Biosynthesis and Properties 2. PHA: Biosynthesis and Properties**

Due to the global awareness of the environmental impact of fossil-based polymers [7], the main goal of plastic industry is nowadays to tackle plastic pollution and its sociopolitical and economic challenges by developing new materials that can combine the advantages of traditional plastics with a sustainable production and disposal. In this research field, biopolymers play a central role due to their great benefits, such as carbon footprint reduction, saving of fossil resources and landfill decrease [8]. Due to the global awareness of the environmental impact of fossil-based polymers [7], the main goal of plastic industry is nowadays to tackle plastic pollution and its sociopolitical and economic challenges by developing new materials that can combine the advantages of traditional plastics with a sustainable production and disposal. In this research field, biopolymers play a central role due to their great benefits, such as carbon footprint reduction, saving of fossil resources and landfill decrease [8].

PHA is a large family of thermoplastic aliphatic polyesters mainly produced by prokaryotic organisms, such as bacteria, most prevalently Gram-negative [9], and archaea under conditions of nutrient depletion and in the presence of an excess of carbon source [10]. It is noteworthy to consider that, although only at a preliminary scientific research level, the production of PHAs from plants was achieved [11]. The general structure of PHAs is reported in Figure 2, where m can be equal or greater than one and R can be a hydrogen atom or an alkyl substituent, depending on the type of PHA [12]. Maurice Lemoigne, a French microbiologist, was the first researcher who identified the synthesis of PHAs from bacteria in 1926 by using a culture of *Bacillus megaterium* to isolate poly(3-hydroxybutyrate) (PHB) [13]. PHA is a large family of thermoplastic aliphatic polyesters mainly produced by prokaryotic organisms, such as bacteria, most prevalently Gram-negative [9], and archaea under conditions of nutrient depletion and in the presence of an excess of carbon source [10]. It is noteworthy to consider that, although only at a preliminary scientific research level, the production of PHAs from plants was achieved [11]. The general structure of PHAs is reported in Figure 2, where m can be equal or greater than one and R can be a hydrogen atom or an alkyl substituent, depending on the type of PHA [12]. Maurice Lemoigne, a French microbiologist, was the first researcher who identified the synthesis of PHAs from bacteria in 1926 by using a culture of *Bacillus megaterium* to isolate poly(3 hydroxybutyrate) (PHB) [13].

**Figure 2.** General chemical structure of polyhydroxyalkanoates (PHAs); "m" varies from 1 to 4 and "n" ranges from 100 to 30,000; R denotes a hydrogen atom or an alkyl side chain [12]. **Figure 2.** General chemical structure of polyhydroxyalkanoates (PHAs); "m" varies from 1 to 4 and "n" ranges from 100 to 30,000; R denotes a hydrogen atom or an alkyl side chain [12].

As several biopolymers belong to the PHA family, their classification is important, and they can be sorted depending on their chain-length monomeric composition, according to the number of carbon atoms per monomer, and here a great importance is played by the composition of the monomer side chain R [14]: As several biopolymers belong to the PHA family, their classification is important, and they can be sorted depending on their chain-length monomeric composition, according to the number of carbon atoms per monomer, and here a great importance is played by the composition of the monomer side chain R [14]:


Generally, scl-PHAs, containing mainly 3-hydroxybutyrate (3HB) or 3-hydroxyvalerate (3HV) units, have a higher degree of crystallinity, a higher glass transition temperature, and

a higher molecular mass compared to mcl-PHAs [15–17], containing 3-hydroxyhexanoate (3HH), 3-hydroxyoctanoate (3HO), 3-hydroxydecanoate (3HD), or 3-hydroxydodecanoate (3HHD) monomers. temperature, and a higher molecular mass compared to mcl-PHAs [15–17], containing 3 hydroxyhexanoate (3HH), 3-hydroxyoctanoate (3HO), 3-hydroxydecanoate (3HD), or 3 hydroxydodecanoate (3HHD) monomers.

Generally, scl-PHAs, containing mainly 3-hydroxybutyrate (3HB) or 3-hydroxyvalerate (3HV) units, have a higher degree of crystallinity, a higher glass transition

*Bioengineering* **2021**, *8*, 29 4 of 31

Another possible distinction can be made between homopolymer, of which the most famous and widespread example is PHB, and copolymers, such as poly(3-hydroxybutyrate-*co*-3-hydroxyvalerate) (PHBV), poly(3-hydroxybutyrate-*co*-4-hydroxybutyrate) (P3HB-4HB), or poly(3-hydroxybutyrate-*co*-3-hydroxyhexanoate) (PHBH). In this latter case, also the monomer arrangement can define a further method of classification. In fact, the difference between block copolymers and random copolymers is due to the ordered succession of similar monomers, unlike a random distribution, distinctive of the second type of copolymers [18]. The physical blending or the chemical copolymerization allow us to obtain a final material with tuned properties, which directly depend on the structures of the singleconstituent monomers [10]. For example, PHB has a high crystallinity and brittleness, which can be reduced by introducing a new monomer unit, such as 3HV or 3HH [19]. The molar composition ratio of the copolymers is a key factor to tune the final properties, such as elongation at break and degree of crystallinity, which increase with the increase of 3HV [20] or 3HH [21] molar content in the structure. Figure 3 shows a schematic representation and categorization of the PHA family according to the chain-length and the composition of the structural units. Another possible distinction can be made between homopolymer, of which the most famous and widespread example is PHB, and copolymers, such as poly(3-hydroxybutyrate-*co*-3-hydroxyvalerate) (PHBV), poly(3-hydroxybutyrate-*co*-4-hydroxybutyrate) (P3HB-4HB), or poly(3-hydroxybutyrate-*co*-3-hydroxyhexanoate) (PHBH). In this latter case, also the monomer arrangement can define a further method of classification. In fact, the difference between block copolymers and random copolymers is due to the ordered succession of similar monomers, unlike a random distribution, distinctive of the second type of copolymers [18]. The physical blending or the chemical copolymerization allow us to obtain a final material with tuned properties, which directly depend on the structures of the single-constituent monomers [10]. For example, PHB has a high crystallinity and brittleness, which can be reduced by introducing a new monomer unit, such as 3HV or 3HH [19]. The molar composition ratio of the copolymers is a key factor to tune the final properties, such as elongation at break and degree of crystallinity, which increase with the increase of 3HV [20] or 3HH [21] molar content in the structure. Figure 3 shows a schematic representation and categorization of the PHA family according to the chain-length and the composition of the structural units.

**Figure 3.** PHAs classification depending on the chain length and the chemical structure of the monomers. PHBV—poly(3 hydroxybutyrate-*co*-3-hydroxyvalerate); P(3HB-4HB)—poly(3-hydroxybutyrate-*co*-4-hydroxybutyrate); PHB—poly(3-hydroxybutyrate); P4HB—poly(4-hydroxybutyrate); P(3HO-3HD-3HDD)—poly(3-hydroxyoctanoate-*co*-3-hydroxydecanoate-*co*-3-hydroxydodecanoate); P(3HO-3HH)—poly(3-hydroxyoctanoate-*co*-3-hydroxyhexanoate); PHO—poly(3-hydroxyoctanoate); PHH—poly(3-hydroxyhexanoate); PHD—poly(3-hydroxydecanoate). **Figure 3.** PHAs classification depending on the chain length and the chemical structure of the monomers. PHBV—poly(3-hydroxybutyrate-*co*-3-hydroxyvalerate); P(3HB-4HB)—poly(3-hydroxybutyrate-*co*-4-hydroxybutyrate); PHB—poly(3-hydroxybutyrate); P4HB—poly(4-hydroxybutyrate); P(3HO-3HD-3HDD)—poly(3-hydroxyoctanoate*co*-3-hydroxydecanoate-*co*-3-hydroxydodecanoate); P(3HO-3HH)—poly(3-hydroxyoctanoate-*co*-3-hydroxyhexanoate); PHO—poly(3-hydroxyoctanoate); PHH—poly(3-hydroxyhexanoate); PHD—poly(3-hydroxydecanoate).

As already reported, PHAs are bio-based polymers, whose origin derives from bacterial and archaeal fermentation. Some microorganisms, when they are subjected to an environmental stress, such as a depletion of essential nutrients, can start a conversion of the carbon sources in hydroxyalkanoate units, such as carbon and energy reserve, which are further polymerized into PHA granules through a biosynthetic pathway and stored in the bacterial cell cytoplasm [22]. The average size of the PHA granules is approximately As already reported, PHAs are bio-based polymers, whose origin derives from bacterial and archaeal fermentation. Some microorganisms, when they are subjected to an environmental stress, such as a depletion of essential nutrients, can start a conversion of the carbon sources in hydroxyalkanoate units, such as carbon and energy reserve, which are further polymerized into PHA granules through a biosynthetic pathway and stored in the bacterial cell cytoplasm [22]. The average size of the PHA granules is approximately 0.2–0.5 µm [23,24]. In Figure 4, a transmission electron micrograph of *Rhodovulum visakhapatnamense* cells containing PHA granules is reported.

*patnamense* cells containing PHA granules is reported.

*patnamense* cells containing PHA granules is reported.

*Bioengineering* **2021**, *8*, 29 5 of 31

**Figure 4.** TEM image of *Rhodovulum visakhapatnamense* accumulating intracellular PHA granules, appearing as whitish and bright areas (adapted from Reference [25]). **Figure 4.** TEM image of *Rhodovulum visakhapatnamense* accumulating intracellular PHA granules, appearing as whitish and bright areas (adapted from Reference [25]). **Figure 4.** TEM image of *Rhodovulum visakhapatnamense* accumulating intracellular PHA granules, appearing as whitish and bright areas (adapted from Reference [25]).

0.2–0.5 μm [23,24]. In Figure 4, a transmission electron micrograph of *Rhodovulum visakha-*

0.2–0.5 μm [23,24]. In Figure 4, a transmission electron micrograph of *Rhodovulum visakha-*

The biosynthetic pathway of PHB consists of three enzymatic reactions catalyzed by three different enzymes: phbA, phbB, and phbC. The first reaction is a condensation of two acetyl coenzyme A (acetyl-CoA) molecules into acetoacetyl-CoA by β-ketoacyl-CoA thiolase (encoded by phbA). The second reaction is the reduction of acetoacetyl-CoA to (R)-3-hydroxybutyryl-CoA by an NADPH-dependent acetoacetyl-CoA dehydrogenase (encoded by phbB). Lastly, the (R)-3-hydroxybutyryl-CoA monomers are polymerized into PHB by PHB polymerase (encoded by phbC) [26,27]. The scheme in Figure 5 synthesizes the fundamental enzymatic biosynthetic pathway. The biosynthetic pathway of PHB consists of three enzymatic reactions catalyzed by three different enzymes: phbA, phbB, and phbC. The first reaction is a condensation of two acetyl coenzyme A (acetyl-CoA) molecules into acetoacetyl-CoA by β-ketoacyl-CoA thiolase (encoded by phbA). The second reaction is the reduction of acetoacetyl-CoA to (R)-3-hydroxybutyryl-CoA by an NADPH-dependent acetoacetyl-CoA dehydrogenase (encoded by phbB). Lastly, the (R)-3-hydroxybutyryl-CoA monomers are polymerized into PHB by PHB polymerase (encoded by phbC) [26,27]. The scheme in Figure 5 synthesizes the fundamental enzymatic biosynthetic pathway. The biosynthetic pathway of PHB consists of three enzymatic reactions catalyzed by three different enzymes: phbA, phbB, and phbC. The first reaction is a condensation of two acetyl coenzyme A (acetyl-CoA) molecules into acetoacetyl-CoA by β-ketoacyl-CoA thiolase (encoded by phbA). The second reaction is the reduction of acetoacetyl-CoA to (R)-3-hydroxybutyryl-CoA by an NADPH-dependent acetoacetyl-CoA dehydrogenase (encoded by phbB). Lastly, the (R)-3-hydroxybutyryl-CoA monomers are polymerized into PHB by PHB polymerase (encoded by phbC) [26,27]. The scheme in Figure 5 synthesizes the fundamental enzymatic biosynthetic pathway.

**Figure 5.** Biosynthetic pathway of poly(3-hydroxybutyrate) production within the bacterial cytoplasm. PHB is synthesized by the successive action of three enzymes: β-ketoacyl-CoA thiolase (phbA), acetoacetyl-CoA dehydrogenase (phbB), and PHB polymerase (phbC) in a three-step pathway. **Figure 5.** Biosynthetic pathway of poly(3-hydroxybutyrate) production within the bacterial cytoplasm. PHB is synthesized by the successive action of three enzymes: β-ketoacyl-CoA thiolase (phbA), acetoacetyl-CoA dehydrogenase (phbB), and PHB polymerase (phbC) in a three-step pathway. **Figure 5.** Biosynthetic pathway of poly(3-hydroxybutyrate) production within the bacterial cytoplasm. PHB is synthesized by the successive action of three enzymes: β-ketoacyl-CoA thiolase (phbA), acetoacetyl-CoA dehydrogenase (phbB), and PHB polymerase (phbC) in a three-step pathway.

Nowadays, the number of bacteria that is able to produce PHA is remarkable, i.e., more than eighty different genera [22]. The most commonly used bacteria species able to produce PHAs belong to the genera of *Alcaligenes*, *Azotobacter*, *Bacillus*, *Cupriavidus*, *Chromobacterium*, *Delftia*, *Pseudomonas*, *Ralstonia*, and *Staphylococcus* [28]. Different microorganisms own different polymerase enzymes, and this leads to the fact that every single microorganism is capable of producing small differences in the final biopolymer [29]. For example, *Ralstonia* bacteria have a particular polymerase enzyme that prioritizes the synthesis of scl-PHA [30]; on the opposite, *Pseudomonas* bacteria produce mcl-PHA [31]. Moreover, the PHA production yield can vary significantly from 0.25 g/L, using, for example, terephthalic acid as carbon source for *Pseudomonas putida* GO16, to 51.2 g/L, using commercial glycerol as carbon source for *Cupriavidus necator* DSM 545 [10].

Carbon is at the basis of organic chemistry and the fundamental element for all biomasses. There are different possible carbon sources that can be used to feed the microorganisms during PHA production and they can be classified in three different substrate groups: carbohydrates (e.g., sucrose, lactose, starch, or lignocellulose) [32–34], triacylglycerols (e.g., animal fats or plant oils) [35,36], and hydrocarbons. The last group is not economically significant since only few species of bacteria are capable to synthetize PHAs from this source and the process tends to have a low efficiency [37]. Apart from the carbon, other chemical compounds are required such as nitrogen sources, and some of the most used are (NH4)2SO4, NH4Cl, or NH4NO<sup>3</sup> [22]. Variation in carbon to nitrogen ratios led to a different amount of PHA concentration in bacterial cells [38], and most of the studies showed that limiting nitrogen concentration while increasing carbon substrates had a positive effect on the PHA production rate [39,40]. Since the biosynthesis process ends with the storage of PHA granules into the cell cytoplasm, a further crucial step is required, the extraction of the PHAs granules from the bacterial cell. The approaches for biopolymer recovery can be different, and they are here synthesized:


The choice of the most suitable recovery method depends on several factors such as the microbial strain, the type of PHA and the required purity grade of the final product. Specifically, the purity of the polymer has a critical importance for biomedical applications. In fact, biological active contaminants, such as endotoxins, can cause undesired

immunological responses. For example, the US Food and Drug Administration (FDA) regulations limited the endotoxin content of medical devices to 20 USP endotoxin units per device, and to 2.15 in case of devices associated with the cerebrospinal fluid [49]. So far, different approaches have been suggested, but there is still room for improvement and innovation on this particular aspect. Burniol-Figols et al. evaluated an innovative PHA purification through dilute aqueous ammonia digestion (purity 86 ± 0.8%), and they compared it with reference processes, such as dissolution in chloroform and precipitation in methanol (purity 99 ± 0.2%), or also acid-mediated digestion with H2SO4, followed by a treatment with NaOCl and subsequent washing with water and centrifugation (purity 98 ± 2.6%) [50]. Moreover, more environmentally friendly purification processes were proposed like the use of dimethyl carbonate for extraction, followed by a purification step with 1-butanol via reflux. After this purification, the overall purity increased from 91.2 ± 0.1% to 98.0 ± 0.1% [51]. Wampfler et al. investigated another possible purification step, particularly experimented for biomedical applications, which implies the filtration through a column filled with activated charcoal (0.5 mL of charcoal per mL of solution to be filtered). The authors stated that endotoxins were almost completely eliminated by this method, removing polymeric impurities with a molecular weight below 10 kDa, as well as the colored impurities [52].

In terms of process development, there are three main steps for industrial PHA production, first the process has to be optimized at laboratory-scale level, and then it is performed in bioreactor and eventually in pilot plant scale with 100–300 L fermenters [53]. After obtaining a globally recognized result at laboratory scale, in the last decades, the industrial PHA market is still gradually increasing, along with the number of independent companies that are investing on PHA production. However, the final result is far from achieved, if we consider, for example, that, in terms of global production capacity, PHA is about 30,000 tons, which is almost ten times less than bio-PE, and almost 20 times less than bio-PET [54]. For successful industrial scale-up PHA production, the influence of oxygen mass transfer and proper agitation are the most important aspects. Therefore, the scale-up strategies need to be based on keeping one of these parameters constant, with respect to the optimized laboratory-scale setup: volumetric oxygen transfer coefficient (KLa), volumetric power consumption (P/V), impeller tip speed of agitator (Vs), and mixing time (tm) or dissolved oxygen (DO) concentration [55]. To date, worldwide, only a few examples of PHA producers (e.g., Danimer Scientific and Newlight Technologies) have the production capacity to establish collaborations with owners of world-renowned brands in the fields of furniture and food and beverage packaging. This collaboration allows us to boost their economy and lead to a global PHA market growth.

Concurrently, scientific research and technological innovation are engaged for enhancing PHA production efficiency, by optimizing the biosynthesis mechanisms, valorizing cheap and renewable nutrient substrates, and engineering some new bacterial strains or also mixed microbial cultures (MMCs), which do not require sterile conditions and have a wider metabolic potential than single strain [56].

The great structural variety inside the PHA family is reflected in a wide spectrum of physical properties of PHAs, varying from a stiffer behavior, comparable to polystyrene for PHB, to a more flexible behavior with elongation at break values of PHBV similar to those of polypropylene or even low density polyethylene [57,58]. Generally, PHAs are characterized by a low glass transition temperature, between −50 and 0 ◦C, and a melting temperature lower than 200 ◦C [59]. However, probably the most attractive property of PHAs is their biodegradability, which can occur both in aerobic [60] and anaerobic [61] environments, without developing toxic products. The biodegradation of PHAs evolves in three main stages: (1) biodeterioration, which consists in the colonization of the surface, or the bulk of the material, by microorganisms which modify the physical properties of the polymer; (2) biodepolymerization, which is the conversion of polymers into oligomers and monomers induced by enzymes (i.e., PHA depolymerases), secreted by microorganisms, such as bacteria or fungi, which hydrolyze the ester bond of the PHAs; and (3) assimilation, where these low-molecular-weight molecules are metabolized as carbon and energy sources by microorganisms that convert carbon of PHAs into CO2, water, and biomass [62,63].

Considering the similarity in mechanical, thermal and barrier properties of PHAs with commodity polymers along with their bio-based origin and biodegradability, this leads to a great interest of PHAs as possible replacements of conventional polymers in different industrial applications [22], such as household or agricultural items manufacturing [64] and packaging [65,66]. However, the higher prices of PHA make them noncompetitive in the current market compared to the fossil-based polymers. In fact, whilst common polyolefins like polyethylene and polypropylene nowadays cost less than 1 €/kg [56], PHAs can range from 2 to 5 €/kg depending on the grade [67]. Their higher prices are mainly due to the cost of carbon sources, substrates and to the low extraction yield at industrial scale [68]. PHAs are largely hydrophobic and soluble in chlorinated hydrocarbons, such as chloroform or dichloromethane. Considering the biomedical applications, the PHA hydrophobic behavior is a suitable property to avoid that the devices undergo a rapid dissolution and a consequent loss of structural properties, once they are implanted in the aqueous body environment. However, it is well-known that wettable scaffolds are conducive to better cellular adhesion, growth and proliferation, due to the ability of maintaining a humid environment and hence promoting fluid exchange between the designed part and the surrounding [69]. In order to tune this hydrophobic behavior, the PHA matrix can be compounded with hydrophilic filler, such as montmorillonite [70], to increase the water affinity of the composites. Two other key properties for PHA medical applications are biocompatibility and biodegradability in physiological environments, which make them suitable for the production of resorbable biomedical devices, which support cellular adhesion, proliferation, and differentiation [49]. A great benefit in biomedicine is the possibility to implant a device that matches the host tissue mechanical property, and hence it decreases stress concentrations at the device–tissue interface. Therefore, the advantage of PHA compared to other polymers clinically used such as poly(lactide-*co*-glycolide) (PLGA), poly(ε-caprolactone) (PCL), poly(glycolic acid) (PGA), or poly(lactic acid) (PLA) is their wide variety of mechanical properties depending on the chemical structure of the monomers. In fact, PLA and PGA have a high Young's modulus (i.e., 3 and 6 GPa respectively) and a limited elongation at break (i.e., around 2%); differently, PCL has an inferior Young's modulus (i.e., 0.35 GPa), but a much higher elongation at break (i.e., 400%). These materials are optimal for specific biomedical applications, according to their inherent properties. Due to the possibility of tailoring the Young's modulus of PHAs, via compounding or synthetic copolymerization, the applicability of this class of biopolymer is potentially much wider and it gives the chance to choose the best grade of copolymer or monomer to mimic the final destination environment [71]. The mechanical properties of human tissue can considerably vary, for example the Young's modulus for granulation tissue is ~0.2 MPa, for fibrous tissue is ~2 MPa, for articular cartilage is 1–20 MPa, for intervertebral disc is 6–50 MPa, for tendon is 1–3 GPa and for mature bone is ~6 GPa [72,73]. Similarly, the Young's modulus for PHA family may range from ~600 MPa for some grade of copolymers such as P(3HB-4HB) to ~3 GPa for PHB. It is important to note that also the Young's modulus of a same copolymer can be tuned by the variation of the molar composition ratio, for example, the P(3HB-4HB) Young's modulus decreases at the increase of 4HB monomer content [74].

Moreover, compared to the abovementioned polymers, PHA has a better interaction with the immune system, due to the unchanged local pH value during its degradation, without toxic or inflammatory reactions [75]. As the other properties, also degradation times for PHAs depend on the chemical structure of the polymer. A previous research study for bioresorbable cardiovascular scaffolds showed that P4HB has a degradation time ranging between two and twelve months. Differently, PGA has an approximate degradation time, starting from six months; PLLA and PCL degradation take longer than two years [76].

#### **3. Overview on the Main Production Techniques for Biomedical Implants Using PHA**

Advances in the biomedical field are not limited to their final applications or the materials used, but they may also concern advancements in the processing techniques of the final implants and devices. Considering the thermoplastic behavior and the solubility in organic solvents of PHA, different approaches have been followed for transforming PHA raw material into architectures with various potential biomedical applications. The first PHA biomedical devices were simple systems with no control on the structure development, and they were obtained by traditional methods, such as (1) solvent casting, (2) salt leaching, (3) thermally induced phase separation (TIPS), (4) non-solvent-induced phase separation (NIPS), (5) emulsification, and (6) electrospinning. Here, the main features of these techniques are reported and summarized.

**Solvent casting** is probably the most common and the simplest technique for polymer film samples production. PHA are dissolved in an organic solvent (e.g., chloroform or dimethyl sulfoxide) at a typical concentration between 2 and 5 wt%; then, the solution is cast into a mold and the solvent is drawn off to obtain a polymer film with a final thickness of about 100 µm [77,78]. An actual problem of this technique is the impossibility of totally controlling the kinetics of the drying process, which could lead to some stress formation into the film structure and to a wrinkled surface.

**Salt leaching** is a straightforward technique to obtain porous scaffolds, which is a key feature for cell adhesion and proliferation. This process consists in mixing a salt powder, for example, NaCl, with a solution of PHA, and then, after solvent evaporation, leaching out the salt from the structure by soaking the membrane in water [79]. Compared to the solvent cast films, the scaffolds obtained via salt leaching are slightly thicker, varying in a range between 250 and 500 µm [80,81], and with an additional porosity ranging from a few to tens of microns, depending on the size of the salt particles. To avoid using organic solvents, alternatively to the first solvent casting step, a melt molding process is possible. In this case, PHA and salt powders are mixed and poured in a mold, which is first heated above the PHA melting temperature and then cooled down for scaffold solidification. For example, Baek et al. compounded PHBV and hydroxyapatite powder (9:1 *w/w*) with NaCl particles (100–300 µm) at a 1:17 weight ratio and then cast in a mold at 180 ◦C. The final structure is a porous network with pore sizes ranging from several microns to around 400 µm [82].

**Thermally induced phase separation (TIPS)** is a common alternative approach used in the fabrication of porous PHA scaffolds. The physical principle on which it is based is the changing of the temperature condition of a polymer solution, in order to induce a separation into two distinct phases. First, PHA is dissolved in an organic solvent and then frozen. Next, the solvent is removed by a sublimation process (e.g., freeze-drying), leaving a final porous structure. As an example, You et al. dissolved PHBH in 1,4-dioxane under vigorous agitation at 65 ◦C, to promote solubilization. The polymer solution was then frozen at −80 ◦C and lastly freeze-dried for two days. Vacuum drying was applied to completely remove any possible solvent remaining in the scaffolds. Morphology of the scaffolds showed porous structures with pore sizes of approximately 60–100 µm in diameter and 9.3 ± 1.4% in porosity. Moreover, micropores with 5–10 µm diameters were observed interconnected inside the scaffolds, which may help improve intercellular communication [83,84].

**Non-solvent induced phase separation (NIPS)** is another technique used to produce films and thin membranes of PHA. In this case, first PHA is dissolved in an organic solvent and then a phase separation is obtained when this solution enters in contact with a nonsolvent, and hence PHA precipitate forming a film. This technique can be used with direct injection in local body sites, and in these cases, it is important to use a non-toxic organic solvent (e.g., dimethyl sulfoxide (DMSO)) for dissolution of PHA, and when this solution comes into contact with aqueous body fluid (a non-solvent for PHA), a PHA membrane is formed, and the polymer solution leads to the precipitation of PHA, which consists in film formation. Dai et al. investigated different non harmful organic solvents: *N*-methyl

pyrrolidone (NMP), dimethylacetamide (DMAC), 1,4-dioxane (DIOX), dimethyl sulfoxide (DMSO), and 1,4-butanolide (BL) to be used with PHBH, at 15 wt% concentration, as injectable systems in rats at the intra-abdominal position. The results showed that PHBH films with a porous structure were formed and their surface morphologies depended on the different solvent-exchange rate in the phase separation process involving organic solvents and aqueous liquid. PHBH films prepared from NMP, DMAC, and DMSO showed larger porous structures both on the surface and in the cross-section. Those from DIOX and BL had very low porosity on the surfaces [85].

**Emulsification** is the most prevalent technique to obtain PHA microspheres or nanoparticles, which are further used as drug carriers for pharmacological agents. The derived applications are particularly appropriate for topical therapies at controlled-release rate, to safely achieve the desired therapeutic effects [86]. The oil-in-water emulsion-solvent evaporation method is the standard procedure for PHA nanoparticles fabrication. It consists of mixing an organic phase, PHA polymer dissolved in a solvent, to an aqueous solution with an emulsifier, e.g., poly(vinyl alcohol) (PVA). The organic solvent is then removed by volatilization. Finally, nanoparticles are harvested by centrifugation, washed, and dried. The final dimensions of the nanoparticles are usually between 100 and 200 nm, when ultrasonication is used as mixing step [87–89]; differently, if a homogenizer process is used, the dimensions are slightly higher and they vary into a range between 150 and 300 nm [87].

**Electrospinning** is a microfiber production method, and, nowadays, it is the most widely used technique for fabrication of fibrous microporous scaffolds, which simulate the structure of the extracellular matrix. Unlike melt-spinning or wet-spinning, electrospinning does not require a thermal or a chemical coagulation step to produce microfibers. A syringe is filled with a PHA solution and then placed in a high-voltage electric field, usually at 20 kV; thereby, the liquid starts to charge electrically. When the voltage is high enough for the electric repulsion to exceed the surface tension of the droplet at the end of the needle, a thin fluid jet erupts in the direction of the collector, which can be a flat metallic plate or a rotating mandrel. During the travel, the solvent evaporates and the jet dries; hence, electrospun microfibers with a mean diameter of about 500 ± 150 nm [90–92] are collected in the form of a microporous film, with a pore size of 1–1.5 µm [92].

Figure 6 summarizes the above-described conventional processing techniques and graphically represent the final shapes and morphologies of different PHA-based medical devices.

From the techniques presented so far, we conclude that the sustainability aspect, coming from the production of a bio-based and biodegradable polymer, is undermined by the technological approaches requesting a high amount of harmful organic solvents. Moreover, all these techniques are only suitable for the manufacturing of devices with a very limited 3D structure and, overall, with a maximum thickness of hundreds of microns, which is an evident drawback for an extensive use for biomedical applications.

With the spreading of **additive manufacturing (AM)** techniques, a new light on the modern research scene has been turned on 3D printing for biomedical applications (e.g., tissue engineering, prosthesis, or drug delivery), due to the possibility of tailoring the final design and the manufacturing of complex structures, eliminating the costs and time needed for the construction of molds [97,98]. Three-dimensional printers are commanded by a sequence of instructions, expressed in a computer numerical control programming language (e.g., g-code), to build a three-dimensional object starting from a computeraided design (CAD) model. Particularly interesting in biomedical applications is the possibility of customizing and elaborating the starting model, in accordance with the morphological structure of the body in which the device is supposed to be implanted, thus achieving optimal compatibility [99,100]. Moreover, with AM approach is possible to tune the mechanical properties of the final device in order to modify the stiffness of the implant to match that of the original tissue, and hence mitigating the problem of stress concentrations. In fact, varying the structure and the design of the 3D-printed device, it is possible to increase the porosity and thereby to decrease of one order of magnitude the Young's modulus of the implant [73,101,102].

**Figure 6.** Morphology of PHA scaffolds produced with conventional techniques. (**a**) Visual appearance of a poly(3-hydroxybutyrate-co-3-hydroxyhexanoate), PHBH, film obtained via solvent casting (scale bar = 10 mm; adapted from Reference [93]). (**b**) SEM image of a PHBH conduit cross section with uniform wall porosity obtained via salt leaching; the white arrow indicates the internal side (scale bar = 100 μm; adapted from Reference [81]). (**c**) SEM image of a porous scaffold made of a blend of PHB/PHBH obtained via thermally induced phase separation (TIPS) (scale bar = 500 μm; adapted from Reference [94]). (**d**) SEM image of a PHBV membrane cross-section obtained via non-solvent-induced phase separation (NIPS) (scale bar = 50 μm; adapted from Reference [95]). (**e**) Optical microscopy image of PHBH microspheres prepared via emulsification (scale bar = 50 μm; adapted from Reference [96]). (**f**) SEM image of a porous PHBH film obtained via electrospinning (scale bar = 20 μm; adapted from Reference [92]). **Figure 6.** Morphology of PHA scaffolds produced with conventional techniques. (**a**) Visual appearance of a poly(3 hydroxybutyrate-co-3-hydroxyhexanoate), PHBH, film obtained via solvent casting (scale bar = 10 mm; adapted from Reference [93]). (**b**) SEM image of a PHBH conduit cross section with uniform wall porosity obtained via salt leaching; the white arrow indicates the internal side (scale bar = 100 µm; adapted from Reference [81]). (**c**) SEM image of a porous scaffold made of a blend of PHB/PHBH obtained via thermally induced phase separation (TIPS) (scale bar = 500 µm; adapted from Reference [94]). (**d**) SEM image of a PHBV membrane cross-section obtained via non-solvent-induced phase separation (NIPS) (scale bar = 50 µm; adapted from Reference [95]). (**e**) Optical microscopy image of PHBH microspheres prepared via emulsification (scale bar = 50 µm; adapted from Reference [96]). (**f**) SEM image of a porous PHBH film obtained via electrospinning (scale bar = 20 µm; adapted from Reference [92]).

modern research scene has been turned on 3D printing for biomedical applications (e.g., tissue engineering, prosthesis, or drug delivery), due to the possibility of tailoring the final design and the manufacturing of complex structures, eliminating the costs and time needed for the construction of molds [97,98]. Three-dimensional printers are commanded by a sequence of instructions, expressed in a computer numerical control programming language (e.g., g-code), to build a three-dimensional object starting from a computeraided design (CAD) model. Particularly interesting in biomedical applications is the possibility of customizing and elaborating the starting model, in accordance with the mor-Many different techniques of 3D printing have been invented according to the characteristics of the material processed. For PHA 3D printing, the most applied approach is the one of extrusion-based techniques, in which the biopolymer is either melted or dissolved in a solvent and then extruded through a nozzle and deposited on a printing bed, layer-by-layer. Hereafter, the essential extrusion-based AM techniques used in the production of PHA biomedical applications are discussed and compared to the traditional ones: (1) Direct Ink Writing (DIW), (2) Fused Deposition Modeling (FDM), (3) Selective Laser Sintering (SLS), and (4) Computer Aided Wet-Spinning (CAWS).

With the spreading of **additive manufacturing (AM)** techniques, a new light on the

phological structure of the body in which the device is supposed to be implanted, thus achieving optimal compatibility [99,100]. Moreover, with AM approach is possible to tune the mechanical properties of the final device in order to modify the stiffness of the implant to match that of the original tissue, and hence mitigating the problem of stress concentrations. In fact, varying the structure and the design of the 3D-printed device, it is possible to increase the porosity and thereby to decrease of one order of magnitude the Young's modulus of the implant [73,101,102]. **Direct Ink Writing (DIW)** is an extrusion-based 3D-printing technique in which the material is loaded in the form of an ink with rheological properties that allow flowing through the nozzle, as well as supporting its own weight during assembly. In this technique, unlike FDM, the shape retention does not rely on solidification, but rather on shear thinning behavior of the inks. The material is extruded through a thin nozzle, using a computercontrolled robotic deposition system [103]. The final shape of the CAD model is first sliced into layers of height proportional to the nozzle diameter, and it is achieved layer-by-layer. In the production of PHA biomedical devices, the ink is generally obtained by dissolving the

biopolymer in a solvent; however, it is also possible to print directly the biopolymer pellets, using a high-temperature print head and thus exploiting the thermoplastic properties of the material. After printing, a final step of cooling or drying occurs, depending if the material underwent a heating process or not.

**Fused Deposition Modeling (FDM)** is the most popular AM technique, due to its straightforwardness and its design freedom. It is a layer-by-layer melt-extrusion approach that consists in heating up a continuous filament of a thermoplastic material above its glass transition temperature (Tg), and then deposing the extruded material still hot to ensure the adhesion with the underneath layer, already cooled down and hardened. The result is a fully solidified structure whose final design accuracy is guaranteed by a computer control of movements of both printing platform and 3D-printer extruder head [104]. Although FDM can be considered as the most-used 3D-printing technique in a wide range of applications, with different polymeric materials, its utilization for PHA biomedical devices is still extremely limited. Only four scientific research works were published so far, and they evaluate either the applicability as preliminary investigations [105–107] or the use of this technique for the production of an external medical aid in the form of a finger cast [108].

**Selective Laser Sintering (SLS)** is another AM technique, and it was the first one investigated for production of PHA-based biomedical devices [109]. This approach uses a high-power laser beam to locally sinter the biopolymeric powder bed. This procedure is repeated layer-by-layer, to form a 3D structure with a predesigned architecture, generated by CAD software and transferred to the 3D printer. Due to a suboptimal definition of the sintering process, pore areas of the printed scaffolds are generally reduced, compared to the initial designs. An important influence over this effect depends on the powder layer thickness (PLT) and the scan spacing (SS). Pereira et al. investigated the effect of the variation of these printing parameters over the morphological structure, and it was demonstrated that the increase of SS reduces the size deviation; for example, with a PLT of 0.18 mm and different SS (0.15, 0.20, and 0.25 mm) pores of 0.60 <sup>±</sup> 0.04 mm<sup>2</sup> , 0.64 <sup>±</sup> 0.04 mm<sup>2</sup> , and 0.68 <sup>±</sup> 0.05 mm<sup>2</sup> were obtained, respectively. Similarly, the increase of PLT also decreased the reduction of pores with respect to the digital model. Printed scaffolds with SS of 0.15 mm showed pore area values of 0.39 <sup>±</sup> 0.07 mm<sup>2</sup> , 0.60 <sup>±</sup> 0.045 mm<sup>2</sup> , and 0.73 <sup>±</sup> 0.07 mm<sup>2</sup> for PLT of 0.08, 0.18, and 0.28 mm, respectively [110].

**Computer Aided Wet-Spinning (CAWS)** can be considered as an evolution of the wet-spinning technique implemented with a computer control. Wet-spinning consists of extruding from a syringe a PHA solution that precipitates and solidifies in a coagulation bath (e.g., ethanol), due to a non-solvent induced phase separation [111]. The novelty introduced by this technique is the computational control layer-by-layer of the syringe movements, affecting the final shape of the 3D-printed object. This technique allows us to obtain structures with high definition, with a fiber diameter of about 100 ± 20 µm [112,113] and a high porosity, above 80% [112,114]. Due to the non-solvent induced phase separation, this particular technique leads to a multi-scale porous structure in which microporosity, inside the single filaments, is added to a designed macroporous structure. This double scale of porosity has a positive effect on cellular interaction and tissue regeneration [115].

Figure 7 displays SEM images of scaffolds 3D printed by different AM techniques, showing the final microstructure of the PHA-based medical devices.

showing the final microstructure of the PHA-based medical devices.

**Figure 7.** SEM images of PHA scaffolds 3D printed with different AM techniques. (**a**) PHBH scaffolds loaded with antituberculosis drugs 3D printed via Direct Ink Writing (DIW) (scale bar = 500 μm; adapted from Reference [116]). (**b**) PCL/PHBV (50/50) scaffolds 3D printed via Fused Deposition Modeling (FDM) (scale bar = 1 mm; adapted from Reference [107]). (**c**) PHBV scaffolds 3D printed via Selective Laser Sintering (SLS) (scale bar = 500 μm; adapted from Reference [117]). (**d**) Top view of PHBH scaffold 3D printed via Computer Aided Wet-Spinning (CAWS); (insert) detail of the fiber–fiber contact region (scale bar = 500 μm; adapted from Reference [114]). **Figure 7.** SEM images of PHA scaffolds 3D printed with different AM techniques. (**a**) PHBH scaffolds loaded with anti-tuberculosis drugs 3D printed via Direct Ink Writing (DIW) (scale bar = 500 µm; adapted from Reference [116]). (**b**) PCL/PHBV (50/50) scaffolds 3D printed via Fused Deposition Modeling (FDM) (scale bar = 1 mm; adapted from Reference [107]). (**c**) PHBV scaffolds 3D printed via Selective Laser Sintering (SLS) (scale bar = 500 µm; adapted from Reference [117]). (**d**) Top view of PHBH scaffold 3D printed via Computer Aided Wet-Spinning (CAWS); (insert) detail of the fiber–fiber contact region (scale bar = 500 µm; adapted from Reference [114]).

All presented techniques used with PHAs for biomedical-device production are summarized and compared in Table 1, with an evaluation of the main advantages and disadvantages of each method. All presented techniques used with PHAs for biomedical-device production are summarized and compared in Table 1, with an evaluation of the main advantages and disadvantages of each method.

obtain structures with high definition, with a fiber diameter of about 100 ± 20 μm [112,113] and a high porosity, above 80% [112,114]. Due to the non-solvent induced phase separation, this particular technique leads to a multi-scale porous structure in which microporosity, inside the single filaments, is added to a designed macroporous structure. This double scale of porosity has a positive effect on cellular interaction and tissue regeneration [115]. Figure 7 displays SEM images of scaffolds 3D printed by different AM techniques,

**Table 1.** Outline and comparison of the traditional and additive manufacturing (AM) techniques used to produce medical devices from PHA. **Table 1.** Outline and comparison of the traditional and additive manufacturing (AM) techniques used to produce medical devices from PHA.



#### **Table 1.** *Cont.*

#### **4. Different Biomedical Applications: From Conventional to Innovative Technologies**

In the following sections, for a better and clearer understanding for the reader, we decided to use an iterative structure of the paragraphs, dividing every application according to its final utilization: drug delivery, vessel stenting, bone tissue engineering, and cartilage tissue engineering. Then, for each different medical purpose, initially the traditional fabrication techniques of PHA devices are described, highlighting the most important results obtained. Beyond this, the results achieved with AM techniques are illustrated. Particular importance is given to the advancements that AM techniques introduced in the biomedical field, and to the overcoming of some big limitations, which were encountered with traditional techniques.

#### *4.1. Drug Delivery*

Drug delivery was the first biomedical application for PHAs that was investigated [124], and in 1983, Korsatko et al. published the first research work for long term medication dosage [125]. Since then, the use of PHAs as drug carriers met a good success in the biomedical field due to their cytocompatibility and their biodegradation properties in different environments. Particularly for drug carriers, the mechanism of PHA extracellular degradation is important since it is strictly related to the amount and the rate of drug released. The basic idea is to degrade the PHA polymer chains into simpler oligomers or monomers and this can occur via lipase-catalyzed chain scission reactions [126] or via PHA depolymerases enzymatic degradation [127]. Both of them substantially hydrolyze carboxyl-ester bonds in alkanol and alkanoic acid, but they differ according to the substrate preference: lipids for lipases and PHA for depolymerases. However, even lipases showed a degradation activity with PHA polymers [128].

The factors that influence the degradation rate of PHA are different and they can be substantially distinguished between environmental factors and intrinsic PHA properties. Generally, we can state that PHA degrades faster in areas with abundance of bacteria, due to an easy colonization of the biopolymer surface by these microorganisms [129]. However, we have to consider also the PHA chemical structure; for example, if we consider PHA with aromatic side chains, not all microorganisms can decompose them [130]. It was found that an increase in anaerobic conditions [131], temperature [132], and humidity [133] can increment, as well, the degradation rate of the PHA, similar to other biodegradable polymers. On the contrary, an inverse correlation was found between the degradation rate and some properties of the PHA, such as the side chain length [134], the molecular weight, and the degree of crystallinity [75]. Therefore, a useful aspect of this biopolymeric family is the possibility of foreseeing a tunable degradation of the final device, according to the particular application. In Table 2, the main correlations between affecting factors and degradation rate are summarized.


**Table 2.** Main correlations between the degradation rate and affecting factor of degradation. The ↑ symbol indicates an increase; the symbol ↓ indicates a decrease.

The traditional technique that has undoubtedly met the greatest success is the **emulsification** process, which generates nanoparticles that can be loaded with antimicrobial agents or any other drug. One of the first experiments that used the emulsification/solvent diffusion method is dated back to 2008: Yao et al. realized a drug-delivery system that was composed of PHA nanoparticles, phasin (PhaP), and protein ligands. Varying the protein ligands, these systems were tested both in vitro for macrophages hepatocellular carcinoma and in vivo for liver hepatocellular carcinoma. PHAs were suitable for this application, because, due to their hydrophobicity, they had a good affinity with hydrophobic drugs, such as PhaP bound with ligands, which are able to pull the PhaP–PHA nanoparticles to the targeted cells [89].

Xiong et al. demonstrated, for the first time in 2009, the efficiency of employing PHB and PHBH nanoparticles for intracellular controlled drug release via endocytosis by macrophages, which allow the delivery into the cells without receptor mediation. The intracellular drug release was monitored by the amount of change in cells of the retained lipid-soluble colorant, rhodamine B isothiocyanate (RBITC). Both the PHB and PHBH nanoparticles were prepared at two different average sizes of 160 and 250 nm, with a classic **emulsification** procedure, using dichloromethane as organic solvent. It is noteworthy that the drug-loading efficiency decreases with the increase of the PHA nanoparticles dimensions. This study showed that PHA is a class of biopolymer particularly convenient for this application. In fact, it was proved that PHA uptake by macrophages was not harmful for cell viability; moreover, the use of PHA nanoparticles as carriers extends the drug release time. A control sample of free RBITC, not loaded in nanoparticles, was directly added into the culture medium and absorbed by the macrophages in a week. Differently, the use of PHA nanoparticles led to an intracellular sustained drug release period of at least 20 days, meaning an almost threefold increase in drug release time [87].

More recently, Luo et al. used the **emulsification** technique to produce some PHBHbased polymer micelles loaded with docetaxel (DTX) for melanoma treatment. The PHBHbased system is particularly useful to encapsulate DTX, because it avoids using nonionic surfactants that are currently employed for marketed DTX product and that are reported to cause hemolysis, hypersensitivity reactions, or neuro-toxicity. Interestingly, this micelle formulation shows a drug loading efficiency higher than 90%, it improves DTX solubility in aqueous medium and it reduced hemolysis for better blood compatibility. In vivo tests were run by subcutaneous inoculation of a solid tumor, A375 cells, in a mouse and then applying and comparing a control test with PBS (Phosphate Buffered Saline), a marketed DTX treatment and a DTX-loaded PHBH-based micelle treatment. After a week, the results showed an expected increase of about 450% in final tumor volume for the control group, whereas with commercial DTX and experimental micelle the melanoma underwent a volume reduction of 50% and 80%, respectively. Therefore, the results, shown in Figure 8a, demonstrated not only a better blood compatibility but also a better inhibitory ability of the DTX-loaded PHBH-based micelle, compared to a commercial DTX treatment [135].

Rebia et al. produced a fully natural nanofiber composite via **electrospinning** that can mimic the native extracellular matrix (ECM), and therefore increase the compatibility with the host body. The researchers loaded a PHBH matrix with natural antibacterial reagents (*Centella*, propolis, and hinokitiol) to produce antibacterial wound dressings. The obtained structures have a thickness varying from 50 to 140 µm, and they can withstand only moderate mechanical stresses. The in vitro antibacterial activity was evaluated by using the inhibition zone method both for Gram-positive bacteria, tested with *S. aureus*, as well as for Gram-negative bacteria, tested with *E. coli*. The results with propolis and hinokitiol loading gave promising outcomes (Figure 8b) [91].

Traditional techniques are positively used to fabricate drug nanocarriers, but the biggest limitation is that the obtained devices have a very low versatility in the design structures, which are thin membranes in the case of electrospinning or nanospheres obtained by emulsification. Therefore, the introduction in this application field of AM permitted to

obtain complex architectures, extended in all three dimensions, which could operate not only as drug carriers but also as structural support, in the target site.

Duan et al. was one of the first researchers that investigated the 3D printability of PHA for biomedical application. Starting from a micropowder, obtained by double emulsion solvent evaporation method, the researchers decided to further use it, not as a simple drug carrier, but as a powder bed for **SLS** technique. First, a calcium phosphate (Ca-P)/PHBV composite powder loaded with bovine serum albumin (BSA) was prepared, and then the scaffolds (L <sup>×</sup> <sup>W</sup> <sup>×</sup> H = 8 <sup>×</sup> <sup>8</sup> <sup>×</sup> 15.5 mm<sup>3</sup> ) were designed and 3D printed [136]. *Bioengineering* **2021**, *8*, 29 17 of 31 material, another attractive property is the slower and controlled release of antituberculotic drug, up to three months, lengthening the healing period and reducing systemic side effects [116,137].

**Figure 8.** Experimental PHA drug release applications. Panel (**a**) shows in vivo investigation of mice melanoma treatment with PHBH-based polymer micelles loaded with docetaxel (DTX-loaded poly(5%PHBHx/PEG/PPG urethane) and different control groups: Phosphate Buffered Saline (PBS), a commercial docetaxel treatment (Taxotere), and unloaded PHBHbased polymer micelles (poly(5%PHBHx/PEG/PPG urethane). Visual appearance of subcutaneous tumor sizes and tumor volume measurements, within treatment time, are displayed at the top and bottom of the panel, respectively (adapted from Reference [135]). Panel (**b**) represents the inhibition zones of neat PHBH and PHBH composite electrospun nanofibers with centella (30EC) and (30MC), propolis (30EP), and hinokitiol (30EH) on Gram-positive bacteria (*S. aureus*) (A) and Gram-negative bacteria (*E. coli*) (B) (adapted from Reference [91]). Panel (**c**) shows two photographs of a cylindrical scaffold (D × H = 6 × 8 mm2) 3D printed via DIW and implanted in a rabbit's femur for post-surgical treatment of osteoarticular tuberculosis (adapted from Reference [137]). Wu et al. investigated the possibility of 3D printing a clinical device via **FDM**, which could also have an antibacterial activity. They melt-compounded a maleic anhydride grafted PHA (PHA-g-MA) with multi-walled carbon nanotubes (MWCNTs) for the pro-**Figure 8.** Experimental PHA drug release applications. Panel (**a**) shows in vivo investigation of mice melanoma treatment with PHBH-based polymer micelles loaded with docetaxel (DTX-loaded poly(5%PHBHx/PEG/PPG urethane) and different control groups: Phosphate Buffered Saline (PBS), a commercial docetaxel treatment (Taxotere), and unloaded PHBH-based polymer micelles (poly(5%PHBHx/PEG/PPG urethane). Visual appearance of subcutaneous tumor sizes and tumor volume measurements, within treatment time, are displayed at the top and bottom of the panel, respectively (adapted from Reference [135]). Panel (**b**) represents the inhibition zones of neat PHBH and PHBH composite electrospun nanofibers with centella (30EC) and (30MC), propolis (30EP), and hinokitiol (30EH) on Gram-positive bacteria (*S. aureus*) (A) and Gram-negative bacteria (*E. coli*) (B) (adapted from Reference [91]). Panel (**c**) shows two photographs of a cylindrical scaffold (D <sup>×</sup> H = 6 <sup>×</sup> 8 mm<sup>2</sup> ) 3D printed via DIW and implanted in a rabbit's femur for post-surgical treatment of osteoarticular tuberculosis (adapted from Reference [137]).

duction of a FDM filament, which can be further used to 3D-print different geometries

tested with the inhibition zone method both for Gram-positive bacteria, tested with *S.* 

Li et al. and Zu et al. suggested an interesting application for a mesoporous bioactive glass (MBG) and PHBV composite, 3D printed via **DIW** starting from a polymer ink dissolved in chloroform and dimethyl sulfoxide. The final goal of this application is meant for post-surgical treatment of osteoarticular tuberculosis, and specifically the 3D-printed scaffolds can be implanted in the surgical defect, combining the osseous regeneration effect with the release of an antituberculotic drug, such as isoniazid or rifampin. The studies investigated in vitro drug release and cellular proliferation, and in vivo surgical procedure was run, implanting the 3D-printed cylindrical scaffolds (D <sup>×</sup> H = 6 <sup>×</sup> 8 mm<sup>2</sup> ) into the femur of different rabbits, represented in Figure 8c. Besides the osteogenetic feature of this material, another attractive property is the slower and controlled release of antituberculotic drug, up to three months, lengthening the healing period and reducing systemic side effects [116,137].

Wu et al. investigated the possibility of 3D printing a clinical device via **FDM**, which could also have an antibacterial activity. They melt-compounded a maleic anhydride grafted PHA (PHA-g-MA) with multi-walled carbon nanotubes (MWCNTs) for the production of a FDM filament, which can be further used to 3D-print different geometries according to the final application. Only a preliminary study of the antimicrobial assay was tested with the inhibition zone method both for Gram-positive bacteria, tested with *S. aureus*, and for Gram-negative bacteria, tested with *E. coli*. Generally, the tested samples demonstrated a higher inhibition zone for *E. coli* rather than *S. aureus*; however, for both class of bacteria, the results showed an increase in antibacterial performance following an increase in MWCNTs content [106].

#### *4.2. Vessel Stenting*

One of the most recent developing field of PHA application is the stent vessel production, since biodegradable stents can provide mechanical support while it is needed, for example, for obstructive cardiovascular disease treatments, and then degrade, leaving behind only the healed natural vessel, without any foreign objects in the body.

For vascular application, the most important biological property of PHA to investigate is the hemocompatibility, for example with an erythrocyte contact hemolysis assay. The easiest way to do that was to prepare **solvent cast** films. Qu et al. fabricated samples of PHB, PHBV, and PHBH. Comparing all the films, the best results were obtained with PHBH films, which showed a two-fold reduced hemolytic activity and also a lower number of bound blood platelets, after a 120-min exposure to platelet-rich plasma [138]. Zhang et al. tried to improve other important properties of PHBH, in order to enhance the applicability of this PHA in vascular engineering. Particularly, they blended PHBH and poly(propylene carbonate) (PPC) to obtain a higher flexibility, evidenced by an increase in elongation at break [118].

The former studies were fundamental to characterize and to state the possible use of this class of polyester for vascular engineering applications. However, there was a big technological issue with this traditional technique, because solvent casting is not suitable for the production of final devices with complex and 3D structures, which are meant to be implanted in human blood vessels. Gao et al. suggested the use of **electrospinning** to fabricate two kinds of PHBH vascular grafts, including straight and corrugated structures with 6 mm inner diameters. These devices have been tested mechanically, to undergo radial compression and circumferential tensile stresses, as well as for suture retention strength and radial compliance. Moreover, the biocompatibility was evaluated in vitro with hemolytic and cytotoxicity tests. The results obtained in this study demonstrated good application value in the field of stent vessel engineering, even comparing the final properties of the experimental grafts with those of commercial ones [139]. Electrospinning is a well-known technique for production of microporous films, but the production of devices with an actual 3D structure is time-consuming. For example, in this study, the realization time of a vascular graft with a thickness of 200 µm took 6 h. Figure 9a shows the final macroscopic aspect of such electrospun PHBH vascular grafts.

**Figure 9.** Experimental PHA applications for vessel stenting. (**a**) Macro morphology of corrugated tubular PHBH scaffold obtained via electrospinning (adapted from Reference [139]). (**b**) Representative photograph of a stent 3D printed via CAWS for small-caliber blood vessels (measure unit = 1 mm; adapted from Reference [123]). **Figure 9.** Experimental PHA applications for vessel stenting. (**a**) Macro morphology of corrugated tubular PHBH scaffold obtained via electrospinning (adapted from Reference [139]). (**b**) Representative photograph of a stent 3D printed via CAWS for small-caliber blood vessels (measure unit = 1 mm; adapted from Reference [123]).

the construction of 3D tubular structures by winding the coagulating wet-spun biopolymer fiber around a rotating mandrel with a predefined pattern. The biopolymer solution is extruded through a needle directly above a rotating mandrel immersed in a non-solvent bath of ethanol; the movement of the needle and the mandrel rotational velocity were controlled by an experimental computer-controlled system. The presented technique

showed a great versatility in the customization of stent fabrication [141].

*4.3. Tissue Engineering*  A challenging frontier of modern medicine is the repairing of damaged tissue of the human body, and it is called regenerative medicine. The main goal of this particular application field is to promote and enhance the formation of new viable tissues by biochemical and cellular processes. A key feature is represented by the positive effects of biocompatible materials and the innovations of technologies that can enhance the fabrication of devices able to simulate the original body environment. In order to achieve this, a connection among different disciplines (biomedicine, material science, and engineering) has to be done, and for this reason, a new interdisciplinary research field was created, i.e., tissue engineering. Due to the good cytocompatibility and to the tunable mechanical properties and degradation rate, PHA demonstrated to be suitable for both hard tissues, i.e., bone and cartilage, and soft tissues [142], nerve, tendon, bone marrow, or vascular applications. Even if the volume of research is still limited, the innovation that AM introduced in this subject of study is noteworthy. Balogová et al. carried out a preliminary study for production of urethra replacement via AM. They prepared a prototype via **DIW** a PLA/PHB tubular structure with the same length and thickness of the aforementioned vascular grafts, which only took 10 min. Compared to the previous research, the use of AM allowed a 36-fold reduction in production time, which is an evident advantage for technological applications. In this first research study, the authors focused on the technological aspect of the production, and they investigated only geometrical and viscoelastic properties of 3D-printed samples, such as the shape retention over time and the deviation from designed sizes. It is possible to state that DIW had sufficient precision to produce tubular samples usable as a replacement for urethra; further mechanical and biological characterizations have to be done to further validate the in vivo implantation [140].

In tissue engineering the device customization is a great advantage; therefore, we can state that this area is the most promising and with the highest potential for biomedical 3D printing. 4.3.1. Bone Tissue Engineering One of the first in vitro research studies of biocompatibility of PHBH was conducted by Yang et al., and they demonstrated that bone marrow stromal cells can attach, proliferate, and differentiate into osteoblasts on PHBH films, obtained by **solvent casting** [143]. Wang et al. used the **salt-leaching** technique to obtain porous scaffolds, in order to demonstrate an increased attachment and proliferation of bone marrow cells, as well as an earlier osteogenesis, onto a rough surface. The optimal pore size detected is about 3 μm in diameter. In this study, PHBH scaffolds (Figure 10a) showed better performance Puppi et al. realized via **CAWS** some PHBH stents for small-caliber blood vessels, and one example is shown in Figure 9b. The developed stents sustained proliferation of human umbilical vein endothelial cells in vitro, and they showed encouraging low levels in terms of thrombogenicity when in contact with human blood. Besides the advance in medical application, this study is also technologically interesting because it widened the field of application of the CAWS technique. It introduced a novel approach that allows the construction of 3D tubular structures by winding the coagulating wet-spun biopolymer fiber around a rotating mandrel with a predefined pattern. The biopolymer solution is extruded through a needle directly above a rotating mandrel immersed in a non-solvent bath of ethanol; the movement of the needle and the mandrel rotational velocity were controlled by an experimental computer-controlled system. The presented technique showed a great versatility in the customization of stent fabrication [141].

#### for osteoblast proliferation rather than PHB and PLA scaffolds [144]. The same authors *4.3. Tissue Engineering*

investigated also the compounding of PHB and PHBH with hydroxyapatite (HAP), and they found that the mechanical properties (compressive elastic modulus and maximum stress) and the osteoblast response improved for the PHB matrix and decreased for the PHBH blend [145]. More recently, Wu et al. studied how to enhance the cell compatibility of the PHA matrix varying the surface morphology of the solvent cast film by compounding the PHBH with carbon nanotubes (CNTs), which resulted in a higher surface roughness and an electrical conductivity. The proliferation of human mesenchymal stem cells A challenging frontier of modern medicine is the repairing of damaged tissue of the human body, and it is called regenerative medicine. The main goal of this particular application field is to promote and enhance the formation of new viable tissues by biochemical and cellular processes. A key feature is represented by the positive effects of biocompatible materials and the innovations of technologies that can enhance the fabrication of devices able to simulate the original body environment. In order to achieve this, a connection among different disciplines (biomedicine, material science, and engineering) has to be done, and for this reason, a new interdisciplinary research field was created, i.e., tissue engineering. Due to the good cytocompatibility and to the tunable mechanical properties and degradation rate, PHA demonstrated to be suitable for both hard tissues, i.e., bone and cartilage, and soft tissues [142], nerve, tendon, bone marrow, or vascular applications. In tissue engineering the device customization is a great advantage; therefore, we can state that this area is the most promising and with the highest potential for biomedical 3D printing.

#### 4.3.1. Bone Tissue Engineering

One of the first in vitro research studies of biocompatibility of PHBH was conducted by Yang et al., and they demonstrated that bone marrow stromal cells can attach, proliferate, and differentiate into osteoblasts on PHBH films, obtained by **solvent casting** [143]. Wang et al. used the **salt-leaching** technique to obtain porous scaffolds, in order to demonstrate an increased attachment and proliferation of bone marrow cells, as well as an earlier osteogenesis, onto a rough surface. The optimal pore size detected is about 3 µm in diameter. In this study, PHBH scaffolds (Figure 10a) showed better performance for osteoblast proliferation rather than PHB and PLA scaffolds [144]. The same authors investigated also the compounding of PHB and PHBH with hydroxyapatite (HAP), and they found that the mechanical properties (compressive elastic modulus and maximum stress) and the osteoblast response improved for the PHB matrix and decreased for the PHBH blend [145]. More recently, Wu et al. studied how to enhance the cell compatibility of the PHA matrix varying the surface morphology of the solvent cast film by compounding the PHBH with carbon nanotubes (CNTs), which resulted in a higher surface roughness and an electrical conductivity. The proliferation of human mesenchymal stem cells (hMSCs) were demonstrated to be outstanding when nanocomposite films contained 1 wt% CNTs, compared with that on pristine PHBH [77].

Assuming that porosity is an increasing factor of cellular proliferation, Xi et al. investigated the possibility of controlling it and they identified **TIPS** as a straightforward technique that allows the regulation of the scaffold pore diameters by varying the quenching temperature and time. The researchers obtained a series of interconnected highly porous scaffolds with pore sizes ranging from 30 to 150 µm. They demonstrated that the pore diameter decreases with decreasing quenching temperature and consequently also the overall porosity of the scaffold [146]. *Bioengineering* **2021**, *8*, 29 22 of 31

**Figure 10.** Experimental PHAs applications for bone tissue engineering. (**a**) SEM images of a porous PHBV scaffold obtained by salt leaching (scale bar = 10 μm; adapted from Reference [144]). (**b**) SEM micrographs of PHBH/silk fibroin (1:1) electrospun films after 14 days of human-umbilical-cord-derived mesenchymal stem cells culture. The white arrow indicates the cells homogeneously distributed on the microporous film (scale bar = 150 μm; adapted from Reference [92]). (**c**) Visual appearance of Ca-P/PHBV scaffolds loaded with BSA and 3D printed via SLS, using different sintering parameters: (A) laser power = 12.5 W and scan spacing = 0.1 mm; (B) laser power = 15 W and scan spacing = 0.1 mm; (C) laser power = 15 W and scan spacing = 0.15 mm (adapted from Reference [136]). (**d**) Visual appearance of PHBH scaffolds 3D printed by CAWS (adapted from Reference [114]). (**e**) Calcium phosphate (Ca–P)/PHBV nanocomposite 3D printed via SLS for the fabrication of a proximal femoral condyle scaffold (scale bar = 1 cm; adapted from Reference [147]). 4.3.2. Cartilage Tissue Engineering **Figure 10.** Experimental PHAs applications for bone tissue engineering. (**a**) SEM images of a porous PHBV scaffold obtained by salt leaching (scale bar = 10 µm; adapted from Reference [144]). (**b**) SEM micrographs of PHBH/silk fibroin (1:1) electrospun films after 14 days of human-umbilical-cord-derived mesenchymal stem cells culture. The white arrow indicates the cells homogeneously distributed on the microporous film (scale bar = 150 µm; adapted from Reference [92]). (**c**) Visual appearance of Ca-P/PHBV scaffolds loaded with BSA and 3D printed via SLS, using different sintering parameters: (A) laser power = 12.5 W and scan spacing = 0.1 mm; (B) laser power = 15 W and scan spacing = 0.1 mm; (C) laser power = 15 W and scan spacing = 0.15 mm (adapted from Reference [136]). (**d**) Visual appearance of PHBH scaffolds 3D printed by CAWS (adapted from Reference [114]). (**e**) Calcium phosphate (Ca–P)/PHBV nanocomposite 3D printed via SLS for the fabrication of a proximal femoral condyle scaffold (scale bar = 1 cm; adapted from Reference [147]).

of PHB/PHBH or even with PLA [153].

Differently from bone regeneration, cartilage structure cannot be self-recreated and

The **TIPS** technique was used as another simple approach to fabricate PHB/PHBH porous scaffold upon which human adipose-derived stem cells (hASCs) were seeded to

Currently, the most common treatments involve only the use of painkillers or surgeries, such as microfracture, osteochondral transfer or autologous chondrocyte implantation. However, these treatments present no actual restoration of cartilage tissue and, in general, an unsatisfactory average long-term result [150]. In the last decade, a new approach for cartilage repair was suggested, and it consists in the use of engineered scaffolds able to support the growth of chondrocytes. However, still further research is required to develop suitable scaffolds, because the neo-generated tissue is often fibrocartilage, which is mechanically inferior and less durable than the one found in healthy articular joints [151]. Since the beginning of the investigation, a particular interest was attributed to PHA as interesting material for the recreation of a favorable environment for the growth of chondrocytes from stem cells. The first works focused on the interaction of chondrocytes with polymer matrices. Deng et al. blended PHBH and PHB and then porous scaffolds were fabricated by the **salt-leaching** method. In order to evaluate the compatibility with this material and the production of extracellular matrix, the chondrocyte cell lines were isolated from rabbit articular cartilage, seeded on the scaffolds and incubated over 28 days [80,152]. Following research explored the best ratios between different component polymers, which could positively combine mechanical properties and biological compatibility. Considering collagen II as a differentiation marker of chondrocytes maturation, blended scaffolds of PHB and PHBH (ratio 1:2) gave the best results, compared with other ratios

Ang et al. successfully fabricated **electrospun** films made of PHBH compounded with silk fibroin (SF), and these devices were able to support the human umbilical cord-derived mesenchymal stem cells proliferation and differentiation into the osteogenic lineage. The obtained electrospun films are in the form of a porous matrix with randomly distributed fibers, with an average diameter in the range of 600 and 980 nm. The mean pore diameter of the electrospun films ranged from 1 to 1.5 µm. Silk fibroin demonstrated an enhancing effect on the proliferation and osteogenic differentiation of stem cells, compared to the pristine PHBH. In Figure 10b, the spread of the cells over the electrospun membrane is shown [92].

Even if the techniques described so far have been instrumental in starting to investigate the use of PHA for bone regeneration, traditional scaffolds have some big limitations, such as very little thickness (i.e., hundreds of microns) and no real direct control over porosity, nor over the dislocation and size of the pores. These aspects have been positively overcome with the use of AM, which has widened the range of application. Three-dimensional printing allows us to build geometries with customized and controlled designs, including the internal pattern, and with a development even in height of several centimeters.

**SLS** was the first AM technique used to fabricate PHA 3D scaffolds. Pereira et al. realized a tetragonal structure squared base (13 <sup>×</sup> 13 mm<sup>2</sup> ) and 26 mm high, with a designed porosity of 1 mm<sup>2</sup> area. PHB scaffolds were 3D printed with different properties, due to the change in the values of the scan spacing (SS) and powder layer thickness (PLT). The results showed that a decrease of the values of PLT or SS involved an increase in the compressive mechanical properties of scaffolds, such as ultimate compressive strength and compressive modulus [110].

Duan et al. studied a system that provided a biomimetic environment for cell attachment, proliferation and differentiation, based on a composite of PHBV compounded with calcium phosphate (Ca-P) nanoparticles, which was proved to be an osteoconductive component. The researchers carried out a study aimed to optimize the **SLS** 3D-printing parameters, i.e., laser power, scan spacing, and layer thickness, according to the final resolution and mechanical properties of a tetragonal porous scaffold (L <sup>×</sup> <sup>W</sup> <sup>×</sup> H = 8 <sup>×</sup> <sup>8</sup> <sup>×</sup> 15.5 mm<sup>3</sup> ), of which three examples are shown in Figure 10c [122]. The final nanocomposite revealed to have not only positive mechanical properties but also good cytocompatibility, tested with a human osteoblast-like cell line [117]. To prove the possibility of using the SLS technique for real medical applications, a human proximal femoral condyle model was obtained from computer tomography scans and then 3D printed into a porous scaffold model with a pore size of 2 mm; an image of this medical prosthesis is shown in Figure 10e [147].

In 2013, **DIW** was investigated by Yang et al. for the first time, among all extrusionbased AM approaches, as a possible technique for PHA bone scaffolds production. Yang et al. fabricated composite scaffolds made of PHBH and mesoporous bioactive glass (MBG) through a combination of 3D printing and surface doping. The MBG coating was found to improve surface hydrophilicity and bioactivity, as well as provide a better environment for human mesenchymal stem cells viability, proliferation, and osteogenic differentiation [148]. Based on the promising results of in vitro biological characterization of the nanocomposite, the research was further carried out by Zhao et al., who selected MBG/PHBH composite scaffolds 3D printed via DIW for in vivo evaluation of osteogenic capability. The scaffolds stimulated bone regeneration in rat calvarial defects within eight weeks [149].

Li et al. studied a real application case of interest for a PHBH and MBG drug-loaded scaffold for osteoarticular tuberculosis. After surgery, it is necessary to fill the surgical defect with an implant, which can combine the effects of osseous regeneration and antitubercular drug (e.g., isoniazid and rifampin) local delivery to treat the area affected by the disease and to avoid internal infections. The researchers 3D-printed, via **DIW**, a cylindrical porous scaffold with a height of 8 mm, a diameter of 6 mm, and an area of each pore of 0.25 mm<sup>2</sup> . The AM technique was particularly useful in this application, to realize a customized device that could perfectly fit to the size of the hole surgically drilled into the treated bone. The structure was tested both for in vitro compatibility and in vivo implantation

in a rabbit femur defect model (Figure 8c). Microtomography evaluations and histology results indicated part degradation of the composite scaffolds and new bone growth in the cavity [116,137].

Mota et al. explored another innovative 3D-printing technique, which was used for the first time with PHA, the **CAWS**. In this study, PHBH 3D-printed scaffolds with different pore sizes and internal architectures were fabricated layer-by-layer, and the processing parameters were investigated for optimization of mechanical compressive properties and biological evaluation. The scaffolds showed a porosity of 79–88%, an extruded filament diameter of 47–76 µm, and a pore size of 123–789 µm; hence, this AM technique allowed the fabrication of scaffolds with a high resolution and a good control over scaffold external shape and internal pattern. The PHBH scaffolds demonstrated also promising results in terms of cell differentiation towards an osteoblast phenotype [114]. Puppi et al. carried on the investigation on PHBH 3D printing for bone scaffold regeneration via CAWS with a pristine PHBH matrix (Figure 10d) [115] and with a PHBH/PCL blend composition [113]. All results showed a promising applicability for in vivo studies and implantations. Recently, they published a work where they used a ternary mixture of PHBH/chloroform/ethanol to prepare the polymeric ink to be used in the 3D printer. With this method, they suggested a more sustainable CAWS process for PHBH scaffolds production, which reduces the employment of halogenated solvent by replacing with ethanol up to 40 *v*/*v*% of the chloroform employed. Besides thus, they evaluated the effect of varying the solvent/nonsolvent ratio on structural morphology, such as macro- and microporosity, on tensile properties and on in vitro preosteoblast cells proliferation [112].

#### 4.3.2. Cartilage Tissue Engineering

Differently from bone regeneration, cartilage structure cannot be self-recreated and an excessive wear of this tissue can lead to a cartilage loss and to osteoarthritis problems. Currently, the most common treatments involve only the use of painkillers or surgeries, such as microfracture, osteochondral transfer or autologous chondrocyte implantation. However, these treatments present no actual restoration of cartilage tissue and, in general, an unsatisfactory average long-term result [150]. In the last decade, a new approach for cartilage repair was suggested, and it consists in the use of engineered scaffolds able to support the growth of chondrocytes. However, still further research is required to develop suitable scaffolds, because the neo-generated tissue is often fibrocartilage, which is mechanically inferior and less durable than the one found in healthy articular joints [151]. Since the beginning of the investigation, a particular interest was attributed to PHA as interesting material for the recreation of a favorable environment for the growth of chondrocytes from stem cells. The first works focused on the interaction of chondrocytes with polymer matrices. Deng et al. blended PHBH and PHB and then porous scaffolds were fabricated by the **salt-leaching** method. In order to evaluate the compatibility with this material and the production of extracellular matrix, the chondrocyte cell lines were isolated from rabbit articular cartilage, seeded on the scaffolds and incubated over 28 days [80,152]. Following research explored the best ratios between different component polymers, which could positively combine mechanical properties and biological compatibility. Considering collagen II as a differentiation marker of chondrocytes maturation, blended scaffolds of PHB and PHBH (ratio 1:2) gave the best results, compared with other ratios of PHB/PHBH or even with PLA [153].

The **TIPS** technique was used as another simple approach to fabricate PHB/PHBH porous scaffold upon which human adipose-derived stem cells (hASCs) were seeded to produce neocartilage, subsequent to a chondrogenic differentiation in vitro process. After 14 days of in vitro culture, the differentiated cells grown on the PHB/PHBH scaffold were implanted into the subcutaneous layer nude mice and after 24 weeks, the appearance of a new cartilage-like tissue could be observed [94]. To develop a higher and more homogeneous cell proliferations over the PHBH scaffolds, You et al. experimented a biological coating of the biopolymer scaffolds with PHA granule binding protein (PhaP)

fused with RGD peptide (PhaP-RGD coating). Human bone marrow mesenchymal stem cells (hBMSCs) were inoculated in the scaffolds and the findings showed that the proposed PhaP-RGD coating led to a more homogeneous spread of cells, and to a better cell adhesion, proliferation and chondrogenic differentiation [84]. RGD peptide (PhaP-RGD coating). Human bone marrow mesenchymal stem cells (hBM-SCs) were inoculated in the scaffolds and the findings showed that the proposed PhaP-RGD coating led to a more homogeneous spread of cells, and to a better cell adhesion, proliferation and chondrogenic differentiation [84].

produce neocartilage, subsequent to a chondrogenic differentiation in vitro process. After 14 days of in vitro culture, the differentiated cells grown on the PHB/PHBH scaffold were

neous cell proliferations over the PHBH scaffolds, You et al. experimented a biological coating of the biopolymer scaffolds with PHA granule binding protein (PhaP) fused with

All mentioned research studies provided a strong and valid basis to start investigating the applicability of PHA matrices for cartilage tissue engineering, but a big limitation was represented by the geometrical constraint in the final shapes of the devices obtained by TIPS or salt leaching. Starting from a real case study, Sun et al. analyzed a possible and new route to build and replace a damaged laryngeal cartilage. The noteworthy innovation of this work was the construction of a hollow, semi-flared geometry prepared by a **combination of solvent casting, compression molding** in a polytetrafluorethylene form, **and salt-leaching** methods. The morphology of the implant was shaped according to the anatomy of an adult laryngeal cartilage, as can been seen in Figure 11a. First, chondrocytes were inoculated onto the PHBH scaffold, and after one week of in vitro culture, an in vivo implantation was performed and the results showed that cartilage formed six weeks after the surgery (Figure 11b) [154]. All mentioned research studies provided a strong and valid basis to start investigating the applicability of PHA matrices for cartilage tissue engineering, but a big limitation was represented by the geometrical constraint in the final shapes of the devices obtained by TIPS or salt leaching. Starting from a real case study, Sun et al. analyzed a possible and new route to build and replace a damaged laryngeal cartilage. The noteworthy innovation of this work was the construction of a hollow, semi-flared geometry prepared by a **combination of solvent casting, compression molding** in a polytetrafluorethylene form, **and salt-leaching** methods. The morphology of the implant was shaped according to the anatomy of an adult laryngeal cartilage, as can been seen in Figure 11a. First, chondrocytes were inoculated onto the PHBH scaffold, and after one week of in vitro culture, an in vivo implantation was performed and the results showed that cartilage formed six weeks after the surgery (Figure 11b) [154].

*Bioengineering* **2021**, *8*, 29 23 of 31

**Figure 11.** Experimental PHAs applications for cartilage tissue engineering. (**a**) Representation of a PHBH medical device prepared by solvent casting, compression molding, and particulate filtering, with a final hollow semi-flared shape, which intends to mimic the laryngeal cartilage morphology (adapted from Reference [154]). (**b**) Photograph of the laryngeal cartilage PHBH specimen with chondrocytes inoculated, 18 weeks after implantation (adapted from Reference [154]). **Figure 11.** Experimental PHAs applications for cartilage tissue engineering. (**a**) Representation of a PHBH medical device prepared by solvent casting, compression molding, and particulate filtering, with a final hollow semi-flared shape, which intends to mimic the laryngeal cartilage morphology (adapted from Reference [154]). (**b**) Photograph of the laryngeal cartilage PHBH specimen with chondrocytes inoculated, 18 weeks after implantation (adapted from Reference [154]).

The former work had the great advantage to allow the construction of a complexshaped device; nevertheless, the experimental procedure for the scaffold fabrication is long and expensive, since it involves using a plastic mold, which should be, every time, customized according to the final implant. Moreover, organic solvent and long times of evaporation need to be estimated. With AM, these limitations could be easily overcome, because starting from a different CAD model, the need for the mold would be completely eliminated. Moreover, 3D printing would allow the fabrication of personalized and complex structures, which could encourage cellular growth in preferential directions or which could have architectures that optimize the contact and the stress transmission between bone and cartilage, for example, in the case of articular cartilage. To the authors' knowledge, there is only one recent work dealing with 3D printing of PHA scaffolds for cartilage tissue engineering. De Pascale et al. assessed the properties of collagen I hydrogel 3D scaffolds, strengthened with solvent cast and 3D-printed PHA polymer. The addition The former work had the great advantage to allow the construction of a complexshaped device; nevertheless, the experimental procedure for the scaffold fabrication is long and expensive, since it involves using a plastic mold, which should be, every time, customized according to the final implant. Moreover, organic solvent and long times of evaporation need to be estimated. With AM, these limitations could be easily overcome, because starting from a different CAD model, the need for the mold would be completely eliminated. Moreover, 3D printing would allow the fabrication of personalized and complex structures, which could encourage cellular growth in preferential directions or which could have architectures that optimize the contact and the stress transmission between bone and cartilage, for example, in the case of articular cartilage. To the authors' knowledge, there is only one recent work dealing with 3D printing of PHA scaffolds for cartilage tissue engineering. De Pascale et al. assessed the properties of collagen I hydrogel 3D scaffolds, strengthened with solvent cast and 3D-printed PHA polymer. The addition of solvent cast and 3D-printed scaffolds increased the mechanical resistance of the structures when compared to the collagen matrix only. Once again, the use of AM technique was an advantage related to traditional techniques, because regarding the compressive stress that the device could undergo, 3D-printed scaffolds showed the highest stiffness compared to the collagen and solvent cast polymer samples [155].

**References** 

#### **5. Future Perspective**

The introduction of PHA and AM in the biomedical field has boosted the advancements of innovative solutions for problems that were so far totally or partially unresolved. The main reasons for this success was certainly due to the high level of customization brought by AM and by the possibility of tailoring the final mechanical properties of 3Dprinted materials, in order to mimic the tissue environment. Besides, also the tunable and interesting properties of PHAs played a central role, for example the wide processing and application versatility, the biological origin, the biocompatibility and the biodegradability. Among AM techniques, FDM owns some well-known advantages, namely its simplicity, rapidity, and ecological sustainability; in fact, it does not require the use of any organic solvent. However, FDM used with PHA for biomedical application is still limited; however, according to the abovementioned properties and advantages of PHA and FDM, we believe that its use will be increasingly investigated and the number of 3D-printed devices by FDM will grow significantly in the next years.

In the field of PHA 3D-printed medical devices, the most promising results were obtained with non-toxic and safely resorbable scaffolds containing living cells that were used for hard tissue regeneration, bone and cartilage particularly. However, no studies were carried on the production of more complex-shaped devices like prosthesis or surgical implants, because these are generally 3D printed with synthetic biopolymer, such as PCL.

The production of synthetic polymer requires the use of chemical solvents, different catalysts (e.g., metal-based, organic, or even enzymatic systems) and also reaction conditions that are particularly energy consuming [156]. If compared to a bacterial synthesis of PHAs, it is quite evident the inconvenience in terms of ecological sustainability. As an indication of possible future developments, in Figure 12 a PHBH clavicle plate 3D printed by FDM is shown, which could be used to treat a broken fracture. Especially due to the resorbability, to the biocompatibility and to the osteogenesis induction of PHAs, this class of material allows us to think of a future medicine, where all components are bio-based, perfectly compatible with human body and devices can be harmlessly reabsorbed by our organism, when they are not needed anymore. *Bioengineering* **2021**, *8*, 29 25 of 31

**Figure 12.** Graphical representation of a PHBH clavicle plate 3D printed by FDM and its final surgical application for bone regeneration (adapted from istock.com/yodiyim, accessed on 21 September 2020). **Figure 12.** Graphical representation of a PHBH clavicle plate 3D printed by FDM and its final surgical application for bone regeneration (adapted from istock.com/yodiyim, accessed on 21 September 2020).

manuscript. All authors have read and agreed to the published version of the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

1. ASTM D6866-20. *Standard Test Methods for Determining the Biobased Content of Solid, Liquid, and Gaseous Samples Using Radiocarbon* 

2. ISO 14855-1:2012. *Determination of the Ultimate Aerobic Biodegradability of Plastic Materials under Controlled Composting Conditions—*

3. EN 13432:2000. *Packaging—Requirements for Packaging Recoverable through Composting and Biodegradation—Test Scheme and Evalu-*

4. Tarrahi, R.; Fathi, Z.; Seydibeyoğlu, M.Ö.; Doustkhah, E.; Khataee, A. Polyhydroxyalkanoates (PHA): From production to nano-

5. Velu, R.; Calais, T.; Jayakumar, A.; Raspall, F. A comprehensive review on bio-nanomaterials for medical implants and feasibility studies on fabrication of such implants by additive manufacturing technique. *Materials* **2020**, *13*, 92, doi:10.3390/ma13010092. 6. Li, Y.; Feng, Z.; Hao, L.; Huang, L.; Xin, C.; Wang, Y.; Bilotti, E.; Essa, K.; Zhang, H.; Li, Z.; et al. A review on functionally graded materials and structures via additive manufacturing: From multi-scale design to versatile functional properties. *Adv. Mater.* 

7. Li, P.; Wang, X.; Su, M.; Zou, X.; Duan, L.; Zhang, H. Characteristics of plastic pollution in the environment: A review. *Bull.* 

8. George, A.; Sanjay, M.R.; Srisuk, R.; Parameswaranpillai, J.; Siengchin, S. A comprehensive review on chemical properties and applications of biopolymers and their composites. *Int. J. Biol. Macromol.* **2020**, *154*, 329–338, doi:10.1016/j.ijbiomac.2020.03.120. 9. Ray, S.; Kalia, V.C. Microbial cometabolism and polyhydroxyalkanoate Co-polymers. *Indian J. Microbiol.* **2017**, *57*, 39–47,

10. Kumar, M.; Rathour, R.; Singh, R.; Sun, Y.; Pandey, A.; Gnansounou, E.; Andrew Lin, K.Y.; Tsang, D.C.W.; Thakur, I.S. Bacterial polyhydroxyalkanoates: Opportunities, challenges, and prospects. *J. Clean. Prod.* **2020**, *263*, 121500, doi:10.1016/j.jcle-

**Funding:** This research received no external funding.

*Method by Analysis of Evolved Carbon Dioxide—Part 1: General Method*; ISO Standard: Geneva, Switzerland, 2012.

*ation Criteria for the Final Acceptance of Packaging*; EN Standard: Brussels, Belgium, 2000.

architecture. *Int. J. Biol. Macromol.* **2020**, *146*, 596–619, doi:10.1016/j.ijbiomac.2019.12.181.

*Analysis*; ASTM Standard: West Conshohocken, PA, USA, 2020.

*Technol.* **2020**, *5*, 1900981, doi:10.1002/admt.201900981.

doi:10.1007/s12088-016-0622-4.

pro.2020.121500.

*Environ. Contam. Toxicol.* **2020**, doi:10.1007/s00128-020-02820-1.

In conclusion, we can foresee a quick and important development in this research field and we think that the next frontier and challenge in biomedical application of PHA could be the 3D printing by FDM of entire prosthesis, or complex surgical implants, which can replace the materials used until now, and which will notably improve the biomedical knowledge and technological state-of-the-art.

**Author Contributions:** All authors have contributed to the conceptualization and the writing of this manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Review* **Additive Biomanufacturing with Collagen Inks**

#### **Weng Wan Chan 1,**† **, David Chen Loong Yeo 1,**† **, Vernice Tan <sup>1</sup> , Satnam Singh <sup>1</sup> , Deepak Choudhury 1,\* and May Win Naing 1,2,\***


Received: 15 May 2020; Accepted: 25 June 2020; Published: 1 July 2020

**Abstract:** Collagen is a natural polymer found abundantly in the extracellular matrix (ECM). It is easily extracted from a variety of sources and exhibits excellent biological properties such as biocompatibility and weak antigenicity. Additionally, different processes allow control of physical and chemical properties such as mechanical stiffness, viscosity and biodegradability. Moreover, various additive biomanufacturing technology has enabled layer-by-layer construction of complex structures to support biological function. Additive biomanufacturing has expanded the use of collagen biomaterial in various regenerative medicine and disease modelling application (e.g., skin, bone and cornea). Currently, regulatory hurdles in translating collagen biomaterials still remain. Additive biomanufacturing may help to overcome such hurdles commercializing collagen biomaterials and fulfill its potential for biomedicine.

**Keywords:** collagen; ECM; extracellular matrix; bioinks; biomanufacturing

#### **1. Introduction**

Collagen is by far the most prevalent extracellular matrix (ECM) molecule found in adult mammals with an estimated 30% of protein mass of multicellular organisms [1]. Although the collagen molecule has 29 subtypes (variants) [2,3], approximately 90% of collagen consists of variants types I, II, III [4]. Collagen extracellular matrix can be found throughout the body in both soft and hard connective tissues including bones, skin, tendon, cartilage, cornea, lung, liver etc. [5].

Its fundamental structural unit is a 300 nm protein consisting of 3 braided α-subunits of 1050 amino acids in length. Each strand comprises the repeating amino acid motif: Gly-Pro-X (X is any amino acid). These strands form hydrogen bonds between the NH bond of a glycine and a carbonyl (C=O) group from an adjacent strand that holds the structure together and form their characteristic triple helix structure [4,6]. Collagen is a hierarchical biomaterial that is self-assembled into fibrils (containing numerous structural units) of ~1 cm length and ~500 nm in diameter (using type 1 Collagen as the archetype). Fascinatingly, the individual tropocollagen monomers are unstable at body temperature and favour random coil conformations. However, collagen fibrillogenesis gives rise to triple helix macromolecular structures with favourable mechanical strength in 3-dimensions, with resistance to enzymatic degradation [6]. Through the introduction of energy (e.g., heat energy from the surroundings), the H-bonds maintaining the orderly collagen structure are separated, causing the individual strands of the triple-helix to separate, resulting in a disorganized, denatured state known as gelatin (please see Figure 1 for more information).

misplacement of glycine due to the mutation results in unstable helices [4]. In their native microenvironment, collagen molecules interact with other biological molecules. Negatively-charged Glycosaminoglycans (linear polysaccharides) sequester growth factors within the ECM [7]. These have been used to generate bio-active collagen scaffolds for cell growth [8]. Furthermore, Collagen interacts with Elastin fibers to provide recoil to the ECM, as well as fibronectin to mediate cell attachment and function [1]. Collagen molecules can also interact with reducing sugars in the body

with various diseases such as atherosclerosis, osteoporosis, diabetes and renal failure [9].

**Figure 1.** The structural forms of collagen and their native interactions**.** The basic collagen unit is a triple-helix microfiber that denatures into gelatine or can be assembled into collagen fibrils. Decorin proteins wrap around collagen fibrils in their native context and bind with glycosaminoglycan chains such as dermatan sulphate. Created with BioRender.com. **Figure 1.** The structural forms of collagen and their native interactions. The basic collagen unit is a triple-helix microfiber that denatures into gelatine or can be assembled into collagen fibrils. Decorin proteins wrap around collagen fibrils in their native context and bind with glycosaminoglycan chains such as dermatan sulphate. Created with BioRender.com.

Collagen biomaterials have been utilised for decades to enhance cell culture/function [10]. A number of collagen or collagen-derivative based protocols and commercial culture products have been used extensively ranging from cell culture surfaces to hydrogels [10]. These include culture well inserts [11,12] (MilliCell®, Transwell®), sponge/gels (Matrigel™, Extracel™) and microcarriers (GEM™). While matrigel is derived from Engelbreth–Holm–Swarm (EHS) tumor and found to contain collagen IV, laminin and heparin sulfate, GEM ™ microcarriers coat an alginate core with gelatin to aid cell attachment. Beyond cell culture reagents, collagen biomaterials have been used for tissue engineering applications including: bone, tendon, cardiovascular therapies and disease models [13], cornea [5], skin, skeletal muscle, artery [14] etc. One usage with great popularity is using collagen scaffolds as dermal regeneration templates for severe wounds and other trauma such as burns. To date, a number The Gly-Pro-X amino acid arrangement is critical to the collagen molecule as seen from disease-causing mutations that lead to osteogenesis imperfecta or "brittle bone" disease. A single misplacement of glycine due to the mutation results in unstable helices [4]. In their native microenvironment, collagen molecules interact with other biological molecules. Negatively-charged Glycosaminoglycans (linear polysaccharides) sequester growth factors within the ECM [7]. These have been used to generate bio-active collagen scaffolds for cell growth [8]. Furthermore, Collagen interacts with Elastin fibers to provide recoil to the ECM, as well as fibronectin to mediate cell attachment and function [1]. Collagen molecules can also interact with reducing sugars in the body which result in its glycation. Glycation molecules result in the formation of advanced glycation end products (AGEs) which gives rise to the loss of soft tissue biomechanical properties and is associated with various diseases such as atherosclerosis, osteoporosis, diabetes and renal failure [9].

of scaffolds/templates containing collagen ingredients are commercially available including: Helistat (Integra ®), Instat (Johnson & Johnson), SkinTemp (BioCor), Helitene (Integra ®), Fibracol (J&J), Biobrane (UDL Laboratories), and Chronicure (Derma Sciences) – not an exhaustive list, which is currently presented in fibre, powder, composite forms etc. [15]. Collagen biomaterials as dermal templates have seen the greatest number of commercial translations to date. Recently, novel applications in sustainable cellular agriculture using collagen biomaterials include making artificial leather and bio-artificial muscle [16]. Collagen biomaterials have been utilised for decades to enhance cell culture/function [10]. A number of collagen or collagen-derivative based protocols and commercial culture products have been used extensively ranging from cell culture surfaces to hydrogels [10]. These include culture well inserts [11,12] (MilliCell®, Transwell®), sponge/gels (Matrigel™, Extracel™) and microcarriers (GEM™). While matrigel is derived from Engelbreth–Holm–Swarm (EHS) tumor and found to contain collagen IV, laminin and heparin sulfate, GEM ™ microcarriers coat an alginate core with gelatin to aid cell attachment.

Beyond cell culture reagents, collagen biomaterials have been used for tissue engineering applications including: bone, tendon, cardiovascular therapies and disease models [13], cornea [5], skin, skeletal muscle, artery [14] etc. One usage with great popularity is using collagen scaffolds as dermal regeneration templates for severe wounds and other trauma such as burns. To date, a number of scaffolds/templates containing collagen ingredients are commercially available including: Helistat (Integra ®), Instat (Johnson & Johnson), SkinTemp (BioCor), Helitene (Integra ®), Fibracol (J&J), Biobrane (UDL Laboratories), and Chronicure (Derma Sciences)–not an exhaustive list, which is currently presented in fibre, powder, composite forms etc. [15]. Collagen biomaterials as dermal

templates have seen the greatest number of commercial translations to date. Recently, novel applications in sustainable cellular agriculture using collagen biomaterials include making artificial leather and bio-artificial muscle [16].

Despite plentiful collagen biomaterial applications developed, collagen has several limitations that curtail its widespread usage: generally poor mechanical properties (vascular tissue engineering applications), thrombogenicity, contamination, source and batch variability [13]. These limitations leave many collagen biomaterial applications in the earlier technology development stages, hindering technology translation.

The emerging field of biomaterials printing - bioprinting, provides the means to create structures from collagen biomaterials, additives and cells in a reproducible and scalable way [17,18]. Adapted from methods first used to manufacture inorganic materials [19], bioprinting is an additive manufacturing approach to produce living tissue and organ analogs for regenerative medicine, tissue engineering, pharmacokinetic and disease/developmental modelling [20]. By patterning various combinations of biomaterials and cells, a goal is to reproduce complex biological architecture to recreate the anatomy in reproducible ways [21,22]. Thus, bioprinting potentially mitigates concerns of product variability by increasing process reproducibility. Moreover, increasing production throughput with bioprinting circumvents bottlenecks in production capacity, making collagen biomaterial products more cost-effective.

This article focuses on the bioprinting of collagen biomaterials/bioinks for (mostly) therapeutic purposes. Bioinks differ from biomaterials in that cells are introduced with the materials and printed, even in situ [23]. On the other hand, biomaterial scaffolds are printed alone before cellular components are added. We discuss how collagen biomaterials are isolated from different sources, processed and analysed post-processing. Thereafter, we discuss various printing methods for collagen biomaterials ranging from manually-casted production (the simplest and lowest throughput) to stereo-/digital light printing (additive manufacturing suited for producing complex shapes). The article concludes with a discussion about translational regulatory, cost and strategy issues using bioprinted collagen biomaterials/bioinks for regeneration and therapy applications.

#### **2. Processing Parameters**

Each step in the processing of collagen for additive manufacturing alters the properties and structure of collagen. Depending on the sources of collagen, extraction steps and crosslinking methods (chemical, physical), the resultant properties will differ. The effects of these processes as well as methods for analyzing collagen biomaterials will be discussed.

#### *2.1. Sources of Collagen*

For additive biomanufacturing, fibril-forming sub-types of collagen (type I, type II, type III, type V, type XI, type XXIV and type XXVII) are preferred because they contribute to the mechanical integrity of the ECM [15,24]. Fibrillar collagen is formed from the assembly of collagen molecules because of the intermolecular bonds between the individual strands to create the signature triple-helix collagen molecule (see the introduction section). These fibrils further assemble into fibre-bundles with tensile strength in tendons and skin [3] or into orthogonal transparent layers (e.g., cornea) [25].

Fibrillar collagen can be extracted from various sources. As animal skin/tendons and cartilaginous tissues are abundant in type I and type II collagen respectively, these tissues are sources of fibrillar collagen extraction [26]. Cells cultured in vitro are used to synthesize collagen as well [27,28]. Cells such as fibroblast and chondrocytes which specialize in type I and type II collagen production respectively can be cultured and the synthesised collagen harvested from media or cell layers. Recombinant collagen production is using genetically engineered microorganisms, plants or animals such as bacteria, yeast, transgenic corn and silkworms [29,30]. Synthetic peptides mimicking collagen trimeric structure have also been investigated to produce collagen-like peptides [31,32]. Collagen from cells grown in vitro, recombinant protein production as well as peptide synthesis have very low yield and are not as

cost-effective as collagen extraction from animal tissues. Hence, most commercial collagen extraction relies on animal sources. While there are variations in collagen between different animal species and tissue sources, variation of collagen exists as well, within the same species due to the nature of collagen. As the collagen molecules in animals form mature crosslinks over time, the age, gender, activity and physical state of the animals play a significant role in forming these crosslinks [2]. The variability of collagen between batches of extraction affects fibrillation and self-assembly properties, and in turn the final collagen biomaterial product.

#### *2.2. Collagen Extraction*

Collagen extraction depends on its solubility in the chosen solvent and composition of collagen types in the tissue sources [26]. Collagen extraction can be broken down into 3 stages: Pre-treatment, extraction and purification. During the pre-treatment step, non-collagen proteins are removed to increase the yield of the collagen extraction process. Depending on the tissue source, removal of the non-collagen proteins (lipids, calcium, etc.) is achieved using alkali solutions, neutral saline solutions, alcohol solutions or a combination of solution [33]. Following pre-treatment of tissues, collagen is then extracted via acid-solubilisation or enzymatic-digestion.

In the extraction of collagen by acid-solubilisation, the pre-treated tissue is added into a dilute acidic solution, typically acetic acid, to disrupt weaker hydrogen bonds between collagen molecules [26]. This allows tissue swelling and acid-soluble collagen (ASC) from the loosened structure to dissolve in dilute acid [34]. However, dilute acid does not disrupt the triple helix structure of collagen due to the strong intermolecular forces between the polypeptide strands [35]. The extracted collagen still retains its telopeptide region and is known as telocollagen.

In the extraction of collagen by enzymatic-digestion, pre-treated tissue is added into a proteolytic enzyme solution, typically pepsin which cleaves non-helical telopeptide at the ends of the collagen microfibrils. Selective cleaving of the telopeptide region results in the destabilisation of the fibril structure and increases collagen dissolution [34]. The triple helix structure of collagen is unaffected due to the selective pepsin enzyme digestion. The extracted collagen molecule does not retain its telopeptide regions and is known as atelocollagen.

While clinical use of collagen use both telocollagen as well as atelocollagen in dermal substitute product showed no collagen induced adverse immunogenic response, the removal the telopeptide regions is suspected to play a role in the immunogenicity and antigenicity of collagen [36]. This is because the immune response in the body targets the antigenic determinant are found in mostly the telopeptides of collagen [37]. However, the antigenic determinants which arise from the helical structure and the amino acid sequence of the collagen also contribute to the immunogenicity and antigenicity of collagen [37]. Additionally, antigenic determinates for immune responses in the body depends on the species as well [36].

These extraction methods are not exclusive and can be performed together. Enzymatic-digestion can be done on acid insoluble collagen to obtain higher yields [26]. The extracted collagen is then filtered to remove impurities and purified through repeated salt precipitation, centrifugation and dissolution in acetic acid. Alternatively, the filtered extract undergoes dialysis for purification before freezing and freeze-drying.

#### 2.2.1. Various Forms–Native, Gelatin (Disordered), Collagen Peptides

Depending on extraction methods used, the molecular weight, α-chain composition, and molecular structure are affected, in turn resulting in a change to the properties of the collagen (e.g., solubility, viscosity, etc.) From the extraction process, collagen can further be processed into denatured forms. Using thermal energy, acids, enzymes or a combination of methods, the intramolecular bonds between the α-chains are broken. As a result, the native helix structure transforms into a random coiled structure known as gelatin. Gelatin is formed as a result of the hydrolytic cleavage of collagen into individual protein strands [34]. Further processing of gelatin into smaller peptide chains is achieved

through proteolytic enzymes resulting in hydrolysed collagen. Hydrolysed collagens molecular weight is significantly smaller (3–6 kDA) compared to their native structure (~300 kDa) [38]. As a result, hydrolysed collagen is much less viscous and more soluble than its native counterpart. For this review, we will only limit our discussion to collagen-based inks for bio-additive manufacturing. While collagen is favoured for its excellent biocompatibility, it exhibits poor mechanical properties [34]. This limitation can be overcome by crosslinking collagen molecules which will be discussed later (Section 2.3 Methods of Collagen Crosslinking).

#### 2.2.2. Collagen Biocomposites

To enhance/modify the biological and mechanical properties of collagen, a mixture of synthetic or natural polymers are used. Blending of collagen together with synthetic polymers gives the final product enhanced mechanical and biological properties. The use of biocompatible synthetic polymers such as poly(lactic acid) (PLA) and polycaprolactone (PCL) allow products with excellent mechanical properties [39,40]. These composites have both the beneficial biological properties of collagen and the mechanical stiffness of synthetic polymers. Blending collagen with natural polymers (such as hyaluronic acid [41], alginate [42], glycosaminoglycans (GAGs) [43], growth factors [44], etc. [15]) for biomaterials that better mimic the native ECM environment or elicit a desired cellular response. Other than natural and synthetic polymers, inorganic compounds such as hydroxyapatite [45] or Tricalcium Phosphate (TCP) [46] can be incorporated to elicit desired cellular responses as well.

#### *2.3. Methods of Collagen Crosslinking*

Additional crosslinking of collagen molecules can be used to enhance the mechanical properties of collagen to provide structural integrity for additive bio-manufacturing such as for muscle tissue (8–20 kPa), cartilage tissue (20–30 kPa) and bone tissue (2–30 GPa) [45,47]. These "artificial" crosslinking bonds can be generated using chemical agents or physical treatment. Increasing concentration of chemical and treatment times generally increase collagen crosslinking. However, when using chemical agents for crosslinker, residual unreacted chemicals and/or chemical byproducts are often left behind [15]. This needs to be managed by washing to minimize cytotoxicity.

#### 2.3.1. Chemical Crosslinking

A commonly used aldehyde for collagen crosslinking is glutaraldehyde (GA). As a dialdehyde, the crosslinker reacts with available amide groups on the collagen chains via Schiff base reactions resulting in covalent imide linkages [48]. These covalent linkages stabilise the intramolecular and intermolecular collagen structure. However, unreacted GA is cytotoxic as it crosslinks cellular proteins which disrupt cellular functions, causing cytotoxicity. GA is used in varying concentrations (0.0025–2.5% wt/v) and treatment times (20 min to 24 h) [49–55]. Increasing concentration and treatment times lead to increased collagen crosslinking.

Carbodiimides can also be used for collagen crosslinking such as 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC). EDC crosslinks the amino and carboxyl groups collagen in a 2-step process: EDC first activates the carboxyl groups of collagen, the activated group then forms an amide linkage with primary amines in collagen [56]. This results in zero-length crosslinking where covalent bond formed is directly between the amino and carboxyl groups without addition of EDC. Crosslinking stabilises the intramolecular and intermolecular collage structure, improving overall mechanical stiffness of collagen as well as the bending stiffness of collagen fibrils. Typically, the use of EDC is accompanied with N-hydroxysuccinimide (NHS), which allows a higher conversion of crosslinks due to amine-reactive intermediates stabilizing [57]. EDC or EDC together with NHS are used in varying concentrations (0.01–2.5% wt/v) and treatment times (2 h to 48 h) [46,51,54–56,58–62]. Increasing concentration and treatment times lead to increased crosslinking of collagen.

Hexamethylene di-isocyanate (HDI), an isocyanate is also used for crosslinking as HDI reacts with available amide groups on the collagen in a nucleophilic addition reaction [63]. The resultant reaction forms a urea linkage to stabilize the intramolecular and intermolecular collage structure [64]. HDI was used in varying concentrations (1.5–5%) and treatment times (5 h–overnight) [55,63,65]. Increasing concentration and treatment times lead to increased crosslinking of collagen.

Plant extracts such as tannic acid and genipin have been explored as sustainable crosslinking agents as well. Tannic acid (TA) is a polyphenol extracted from plants which stabilises the intermolecular bonds of collagen via hydrogen bonds and hydrophobic interactions between TA and collagen molecules [66]. Tannic acids of varying concentrations (0.1% to 6% wt/v) and treatment times(10 min to 120 h) [66–68]. Increasing concentration and treatment times lead to increased collagen crosslinking. Genipin is an Iridoid glycoside compound extracted from plants able to crosslink the free primary amines in protein [69]. This allows genipin crosslink primary amides in collagen, stabilizing the intramolecular and intermolecular collagen structure. Genipin is used in varying concentrations (0.00025% to 0.6%) and treatment times (1 h to 48 h) [69–72].

#### 2.3.2. Physical Crosslinking

The use of chemical crosslinkers inevitably faces issues with cytotoxicity. Physical methods such as dehydrothermal (DHT) treatment and ultraviolet (UV) irradiation are used to create covalent bonds between intermolecular collagen structures.

Dehydrothermal treatment is a thermal treatment process that subjects collagen to high temperatures (>90 ◦C) for several hours or days (12 h to 5 days) under vacuum [40,51,54,73,74]. As a result, condensation reactions occur: between the free amino and hydroxyl groups of collagen (esterification); or between the carboxyl and free amino groups (amide linkage formation) [73]. These ester and amide bonds stabilise intramolecular and intermolecular collagen bonds. Despite the low water content of the collagen in vacuum, due to high temperatures, hydrolysis of the peptide bonds occurs resulting in the collagen triple-helix structure denaturing [73]. Though the mechanical properties of collagen improve with longer treatment times and higher temperature, collagen denaturing increases as well.

UV crosslinking involves irradiating collagen (15 min to 240 min) [74]. The mechanism of crosslinking is a result of free radical formation from peptide bond scissions. UV irradiation forms aromatic radicals which in turn attack the peptide bonds in collagen. These radicals then interact and crosslink, which stabilises intramolecular and intermolecular collagen structure. The effectiveness of UV irradiation depends on the sample preparation, irradiation dose and time of exposure [75]. While UV irradiation improves mechanical properties, it also denatures collagen triple-helix structures [75].

Gamma irradiation crosslinking is similar to UV crosslinking where the collagen structure is irradiated for a period of time (250 min to 1250 min) depending on the desired irradiation dosage. Gamma irradiation "radio-lyzes" water, creating radicals. The effectiveness of crosslinking depends on irradiation dose and exposure time. Compared to UV irradiation, the higher energy of gamma irradiation is able to deeper penetrate thicker collagen structures. However, its downside is denaturing collagen's triple-helix structure. Furthermore, gamma irradiation is often used for sterilization, making it unsuitable to crosslink cell-laden bioinks [76].

While cytotoxic compounds are not formed using physical crosslinking methods, they generally lead to collagen denaturation. Furthermore, physical crosslinking methods are less effective in improving mechanical properties of collagen compared to chemical methods [70].

#### *2.4. Collagen Analytical Methods*

Understanding the structural, morphological, and chemical composition of collagen is critical since additive bio-manufacturing processes may give rise to significant changes. Understanding the structural, morphological and chemical composition allows better design and processing of the collagen raw material to meet the needs of the final product [34,77].

#### 2.4.1. Structural Analysis

Differential Scanning Calorimetry (DSC) can determine collagen thermal stability. DSC compares and measures heat flow differences between a specimen and control when heat is supplied. Using this information, the denaturation temperature of collagen can be determined due to endothermic processes observed during collagen denaturation [78]. Using DSC, the denaturation temperature of soluble fish collagen was determined to be 10 ◦C lower than soluble porcine collagen [79].

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-Page) is used to visualise molecular size distribution of collagen protein fragments. SDS-Page uses an electric field to drive charged proteins through gel. Larger fragments move slower, while smaller fragments move quicker through the gel. Following separation by size, the fragments are stained with Coomassie blue or silver to obtain protein bands. Comparing these with controls of known molecular weight, can determine protein fragment weight. By comparing the banding pattern of known type I collagen chains (α1(I): 97 kDa; α2(I): 95 kDa), SDS-PAGE was used to determine the molecular weight of type V collagen chains through the relationship between relative molecular weight and migration rate [80].

Circular Dichroism (CD) is an absorption spectroscopy method to determine the presence of secondary and tertiary collagen structures. CD measures the differences in absorption of left circularly polarised light and right circularly polarised light. Due to the nature of the peptide bonds and structures, it results in characteristic absorption spectrums. From this information, the secondary and tertiary protein structures such as α -helices (negative bands at 222 nm and 208 nm; positive band at 193 nm), β-pleated sheets (negative band at 218 nm; positive band at 195 nm), triple helical conformation (negative band at 195 nm; positive band at 220 nm) can be determined respectively [81,82].

Raman spectroscopy is a label-free and non-destructive method used to determine the bonds and protein structures present in collagen. Raman spectroscopy measures inelastic light scattering of a sample from incident light generated by a laser source. The bonds and protein structures result in distinct shifts in wavelength of scattered light and hence distinct spectrum peaks such as Amide I band (1655 cm−<sup>1</sup> ), Amide III band (1268 cm−<sup>1</sup> ), α-helix shoulder (1630 cm−<sup>1</sup> ) and β-pleated sheet peak (1675 cm−<sup>1</sup> ). From this information, the relative quantities of bonds and protein structures can be determined for collagen [83].

FTIR is a spectroscopy method to determine the bonds and protein structures present in collagen. FTIR measures absorbance or emission of infrared radiation from a sample after irradiation from an infrared source. The bonds and protein structures result in distinct infrared spectrum peaks. Typical peaks of type Collagen are: Amide A (3299 cm−<sup>1</sup> ), (N–H) stretching; Amide B (2919 cm−<sup>1</sup> ), (CH3) asymmetric stretching; amide I (1628 cm−<sup>1</sup> ), (C=O) stretching; amide II (1540 cm−<sup>1</sup> ), (N–H) bending & (C–N) stretching; amide III (1234 cm−<sup>1</sup> ), (–CONH2) stretching. From this information we can determine the presence of bonds and protein structures and their relative quantities in collagen [84]. Additionally, the ratio peak intensity of 1 between the amide III peak and 1450 cm−<sup>1</sup> is indicative of the helix structure of collagen [84].

#### 2.4.2. Morphological Analysis

Scanning Electron Microscopy (SEM) uses focused beams of electrons to image surface topography of collagen samples. SEM measures the energies of elastic and inelastic-scattered electrons incident upon the sample to recreate surface topography. Typically, SEM can examine porosity of collagen sponges as well as assembled collagen fibre structures. SEM was used to study pore morphology of collagen sponge, collagen-I fibrin gel, collagen 2D nanofibers (oriented and random) [85].

Confocal microscopy can be used for structural visualization too. Confocal microscopy sections images for each focal plane using a laser source before compilation into a 3D image volume of high resolution. There are two modes of image acquisition: fluorescence [86] and reflectance [87]. Fluorescence image acquisition uses fluorescent dyes or autofluorescent properties of collagen to generate image contrast, while reflectance image acquisition relies on differences in refractive indexes. Collagen fibril diameters and pore sizes have been studied using both modes of acquisition [86,87].

Transmission electron microscopy (TEM) is a microscope used to visualise banded collagen fibril structures. TEM generates an image by transmitting an electron beam through a thin specimen on a copper grid, the image is then magnified and projected onto a stage. The regular array of gaps and overlaps in collagen microfibrils result in differences in packing density along the assembled collagen fibre. This leads to the banded structure of the collagen fibrils (64–67 nm). Cryo-TEM was used to analyse fibrillar collagen from mineralized and non-mineralized tissue [88].

Atomic force microscopy (AFM) also visualises banded collagen fibril structures. AFM generates an image by measuring deflection of a cantilever probe across collagen fibres. This information is then rendered into a topographic images. AFM is able to detect differences in packing densities that arise from the array of gaps and overlaps in collagen microfibres [89].

#### 2.4.3. Chemical Assays

Hydroxyproline is a colorimetric assay for quantifying hydroxyproline in collagen. Due to the hydroxyproline amino acid composition being approximately constant across the different types of collagen 11.3% (type I) and 15% (type III), it can indicate the amount of collagen within a sample [90].

Sircol assay is a colormetic assay to quantify collagen, binding to the [Gly-X-Y]<sup>n</sup> helical structure in collagen. Collagen content can be obtained by comparing it to standard curves for calibration [91].

2,4,6-Trinitrobenzne sulfronic acid (TNBS) assay is a colorimetric assay used to quantify free primary amines found in collagen. The amount of free primary amino groups can be obtained by comparing it to known quantities. The amount of TNBS can determine the degree of collagen methacrylation [92].

Ninhydrin assay is a colorimetric assay to quantify free primary amino groups. The dye binds to primary amines found in collagen. It was used to determine the change of free amino groups on collagen nanofibers following pre-treatment of L-lysine [93].

Western blot is a method used to identify the type of collagen following SDS-page analysis. Using monoclonal antibodies specific to the collagen types and visualisation through immunofluorescent staining, the type of collagen can be identified. Western blot was used to confirm Collagen VI chains from cell extracts and culture media [94].

Mass spectroscopy identifies proteins from gaseous ions generated from the protein fragments. These are sorted using an electric field according to mass-to-charge ratio. The relative quantities of ions are recorded. By comparing profiles of protein fragments with a database, the proteins can be identified. Mass spectroscopy was able to identify crosslinked pyridinoline and deoxypyridinoline amino acid in hydrolysed collagen [95].

#### **3. Collagen-Based Ink Printing Applications**

The application of collagen-based ink in both non-additive and additive manufacturing requires understanding of collagen processing, as well as the various printing methods. In this section, the principles behind the printing methods and their applications are examined.

#### *3.1. Non-Additive Manufacturing*

Non-additive manufacturing methods, casting and electrospinning of collagen-based inks and their applications are discussed. Casting involves pouring a liquid material into a mold of desired shape before solidifying and removal. Typically for collagen-based biomaterials, highly porous 3D structures (sponges) are obtained via the freeze-drying process while thin-films are obtained via air drying [34]. Freeze drying is a complex process where ice crystals in the frozen mold are removed by sublimation under vacuum. Pore size and direction of the sponge can be controlled during freeze-drying [96,97].

Collagen sponges are used extensively in wound healing and tissue engineering as scaffolds for bone [98], skin and soft tissues [99]. The porous nature of collagen sponges allow cell migration as well as nutrient diffusion into the scaffold while providing a substrate for growth. The collagen sponge can be loaded with drugs, growth factors and bio-additives to enhance scaffold bioactivity [50,60,98–100]. Collagen-glycosaminoglycan scaffolds have been successfully used to regenerate skin from full thickness burns [50]. Additionally, by varying the glycosaminoglycan concentration and pore size, peripheral nerve tissue was successfully regenerated too [101]. Loading TGF-β1 into a collagen sponge allowed controlled release of growth factors, enhancing bone regeneration of a rabbit skull defect [44].

When collagen is laid out to dry, a thin-film of collagen is obtained via evaporation. As water and solvents evaporate, fibres and molecules are brought closer together due to surface tension of the solvent giving rise to a thin-film layer upon drying [34]. Thin collagen films are typically used in cornea treatment owing to their optically transparent nature and biological properties [61]. However, collagen films are not limited to ocular tissue engineering, micropatterns can also be designed onto the film as part of the casting process to influence osteoblast cell orientation [54]. By stacking the collagen film layer by layer, the resulting biomatrix encouraged neo-tissue formation in a hernia repair model [71]. The films can also be wrapped into tubes for nerve grafting applications [49]. While functioning as a barrier membrane, collagen films can also be loaded with drugs, growth factors and bio-additives to enhance bioactivity. Additionally, collagen film degeneration and its mechanical properties can be controlled by varying crosslinking to control the release of its contents via degradation [102,103]. Collagen films are also suitable as edible food packaging [104].

#### Electrospinning

Electrospinning consists of loading a desired biomaterial and a volatile solvent into a syringe. By applying a voltage to the needle tip, an electric field forms between the needle tip and the collector. Once, the electrostatic forces of repulsion are greater than the surface tension of the extruded liquid, a taylor cone is formed and the charged liquid is ejected onto the collector. The volatile solvent evaporates, resulting in fine nano/microscale fibres. These fibres are then deposited onto the metallic collector. By varying the extrusion rate, voltage of charged material, needle gauge and distance between the needle and collector the fibre diameters can be controlled [105].

Processing materials via electrospinning is appealing due to the ability to produce fibre meshes with diameters similar to the native fibrillar network present in the extracellular matrix (20 nm to 40 µm) [106]. Electrospinning can be performed using pure collagen or synthetic polymer additives such as PLLA or PCL to increase mechanical stiffness. Various electrospinning set-up can be used to produce different scaffolds for a variety of applications. A co-electrospinning system containing 2 mixtures of collagen and synthetic polymers was used to produce a scaffold with different regions to mimic muscle-tendon junction properties [107]. Using multi-layered electrospinning, an arterial structure was fabricated using a PCL, elastin and collagen layer was able to achieve significant improvement in mechanical properties and designed to mimic native arterial tissue [108]. A combination of electrospinning and electrospraying technology was used to produce 3D constructs which improved cell infiltration and controlled release of bio-additives [109].

However, the solvents used in electrospinning can significantly denature collagen. Typical fluoroalcohols used in electrospinning such as 1,1,1,3,3,3-Hexafluoro-2-propanol (HFP) cause a loss of collagen's triple helical structure [106]. Fortunately, solvents have been designed to minimize collagen denaturation when electrospun using "less harsh" solvents such as acetic acid/DMSO and PBS/ethanol [110].

#### *3.2. Additive Biomanufacturing*

In this section, four additive bio-manufacturing technologies will be discussed: extrusion bioprinting, inkjet bioprinting, laser-assisted bioprinting and stereolithographic/digital light processing bioprinting. The main advantage of additive bio-manufacturing is to produce complex shapes with internal structures at high resolution and accuracy without molds or shaping tools required by non-additive methods. Moreover, additive bio-manufacturing is amenable to printing with cell-laden inks (bio-inks) [24].

While all additive biomanufacturing processes create structures via layer-by-layer deposition of biomaterials, not all collagen-based inks can be printed using the following methods. As such, flexible printing method such as extrusion printing have a larger number of applications and variation of printing formulations, whereas more restrictive printing methods such as inkjet, laser-assisted, and stereolithography printing have fewer applications.

#### 3.2.1. Extrusion

In extrusion bioprinting, biomaterial inks are loaded into a syringe and printed as filaments onto a stage via a mechanical or pneumatic dispensing system. Precise deposition of material is controlled by a dispensing stage along the x, y, and z axis. This method of bioprinting accommodates a large range of ink viscosities (30–60 <sup>×</sup> <sup>10</sup><sup>7</sup> mPa·s) [18]. Through multiple print heads, multiple materials and formulations can be printed together. However, this printing method is limited by the print resolution (100 µm), which is determined by the nozzle diameter [111]. Furthermore, printed cells experience high shear stresses when extruded under high pressure and small nozzles, resulting in lower cell viability.

Due to the nature of extrusion bioprinting, the viscosity of the collagen-based ink plays an important role in the printing process. The tendency of collagen to self-assemble into fibrillar structures at neutral pH when incubated at 37 ◦C allows collagen to form stable structures after printing [2]. Pure collagen was formulated to be self-supporting by either increasing the concentration or neutralising pH prior to extrusion. Following the extrusion process, scaffolds self-assembled in a neutral buffer to support self-assembly. This process produced tissue spheroid scaffolds as well as printing cell-laden inks into pre-set extrusion designs [47,112].

Combining collagen with other polymers, it is possible to design self-supporting structures by incorporating polymers rather than solely relying on pure collagen. An example was the use of cell-laden collagen/gelatin/alginate ink, by taking advantage of a two-step process involving thermal crosslinking with gelatin at low temperatures followed by crosslinking alginate in calcium solution [113]. The construct was printed at low temperature for gelatin to thermally crosslink and support the structure. Thereafter, it was immersed in calcium solution for ionic crosslinking of alginate to fix its shape. Gelatin and alginate was removed via diffusion and sodium citrate respectively, leaving behind a cell-laden collagen structure. A similar approach was used in cell-laden collagen/alginate ink where coaxial extrusion of collagen-alginate inks with calcium solution allowed the printed ink to be self-supporting [114]. In another, Pluronic F-127/Collagen ink was used to modify the gelation of the printed collagen ink, allowing it to be self-supported and be removed via diffusion in media [115]. A process unique to extrusion bioprinting known as freeform reversible embedding of suspended hydrogels (FRESH), non-self-supporting collagen ink formulations can print complex collagen scaffolds which are then self-assembled and collected from the hydrogel suspension [116].

Following the extrusion printing process, additional crosslinking of collagen (mentioned in earlier sections) can tune the mechanical properties of the collagen scaffold as desired [41,46,58,59,67,68,70,117]. Additionally, the extrusion printing process was able to generate collagen-composite scaffolds loaded with bio-additives such as silk fibroin, β-TCP, HA via-freeze-drying process for bone tissue regeneration [46,59]. Extrusion bioprinting can be combined with inkjet bioprinting for a one-step process to produce cell-laden 3D skin tissue (Figure 2A) [118].

#### 3.2.2. Inkjet Printing

In inkjet bioprinting, biomaterials in a liquid state are loaded into a cartridge and deposited onto a substrate via droplets. The propulsion of droplets is achieved through pulses of pressure generated via thermal, acoustic or piezoelectric elements. The precise deposition of material is controlled by the dispensing system along the x, y-axis and print platform along the z-axis [18]. Through multiple print heads and cartridges, different material formulations can be combined. Additionally, as a nozzle-less systems, cell viability via inkjet bioprinting is higher compared to extrusion bioprinting. However, there is a material viscosity limit (10 mPa·s) for the inks printed due to the limited force generated to propel droplets onto the substrate [119]. Due to low-viscosity inks used in the system, additional processing steps are required to form 3D structures.

3.2.3. Laser-Assisted Printing

The viscosity limit of inkjet bioprinting restricts bioink formulations and bioink cell concentration. However, the self-assembly of collagen after printing allows it to be printed at low viscosity and crosslinked to produce cornea-like structures loaded with corneal stromal keratocytes (Figure 2B) [120]. Collagen ink blended with agarose in cell-laden printing gave rise to mesenchymal stem cells (MSCs) with a spread morphology, resulting in osteogenic differentiation [121]. Inkjet bioprinting was also used to generate collagen ink patterns onto which smooth muscle cells as well as neuronal cells were cultured, resulting in complex cellular patterns [122,123]. Additionally, inkjet bioprinting was applied to create in vitro cancer model microtissue arrays for drug testing and studying tumor progression [124]. Moreover, by controlling the thickness of the collagen gels printed via inkjet printing and seeding cells between the layers of the 3D construct, cell aggregates have been shown to fuse together, demonstrating potential for organ printing [125]. *Bioengineering* **2020**, *7*, x FOR PEER 11 of 23 bioprinting was also used to generate collagen ink patterns onto which smooth muscle cells as well as neuronal cells were cultured, resulting in complex cellular patterns [122,123]. Additionally, inkjet bioprinting was applied to create in vitro cancer model microtissue arrays for drug testing and studying tumor progression [124]. Moreover, by controlling the thickness of the collagen gels printed via inkjet printing and seeding cells between the layers of the 3D construct, cell aggregates have been

shown to fuse together, demonstrating potential for organ printing [125].

**Figure 2.** Bioprinting of collagen-based inks for tissue engineering**.** (**A**) (a,b) Hybrid system (extrusion-based and inkjet-based dispensing modules) used for bioprinting of collagen bioink for developing human skin models, (c) bioprinted model showed good structural features and respective dermis (Col) and epidermis (K10) biomarkers [118]; (**B**) (a,b) Drop‐on‐demand (DoD) bioprinting was used for bioprinting collagen bioink to develop functional biomimetic 3D corneal model, (c) 3D view of human CSK 7 days after bioprinting stained with live/dead staining, most of cells found viable, (d) Smooth muscle actin immunocytochemical stainings of CSK‐loaded agarose‐collagen blends 7 days after bioprinting, observed positive keratocan (Kera) and lumican (Lum) expression [120]; (**C**) (a) Laser-assisted bioprinting was explored for in-situ bioprinting of collagen-based bioinks for bone regeneration applications, (b) two different printed designs: a ring and a disk, and (c) disk printed geometry showed homogeneous regeneration throughout the defect, in contrast with the ring geometry, where regeneration is mainly observed at the periphery [126]. **Figure 2.** Bioprinting of collagen-based inks for tissue engineering. (**A**) (a,b) Hybrid system (extrusion-based and inkjet-based dispensing modules) used for bioprinting of collagen bioink for developing human skin models, (c) bioprinted model showed good structural features and respective dermis (Col) and epidermis (K10) biomarkers [118]; (**B**) (a,b) Drop-on-demand (DoD) bioprinting was used for bioprinting collagen bioink to develop functional biomimetic 3D corneal model, (c) 3D view of human CSK 7 days after bioprinting stained with live/dead staining, most of cells found viable, (d) Smooth muscle actin immunocytochemical stainings of CSK-loaded agarose-collagen blends 7 days after bioprinting, observed positive keratocan (Kera) and lumican (Lum) expression [120]; (**C**) (a) Laser-assisted bioprinting was explored for in-situ bioprinting of collagen-based bioinks for bone regeneration applications, (b) two different printed designs: a ring and a disk, and (c) disk printed geometry showed homogeneous regeneration throughout the defect, in contrast with the ring geometry, where regeneration is mainly observed at the periphery [126].

biomaterial is achieved by the movement of the donor substrate in the x, y axis and the receiving substrate in the x,y and z axis [18]. By coating the donor film with different materials and focusing the laser beam on different locations of the donor substrate for deposition, a heterogenous 3D

In laser-assisted bioprinting (LAB), a layer of biomaterial is deposited onto a substrate via laser-

#### 3.2.3. Laser-Assisted Printing

In laser-assisted bioprinting (LAB), a layer of biomaterial is deposited onto a substrate via laser-induced forward transfer. A pulsed laser beam is focused on to a donor substrate coated with a laser-energy absorbing layer and a biomaterial layer. Energy absorbed by the donor would drive the biomaterial from the donor substrate onto the receiving substrate. The precise deposition of biomaterial is achieved by the movement of the donor substrate in the x, y axis and the receiving substrate in the x,y and z axis [18]. By coating the donor film with different materials and focusing the laser beam on different locations of the donor substrate for deposition, a heterogenous 3D structure can be obtained. This method of bioprinting, like extrusion bioprinting, also allows a large range of ink viscosities (1–300 mPa·s) [127]. It has the highest print resolution (10 µm) amongst additive bio-manufacturing methods and allows for a high concentration of cell loading [128]. However, preparation of a homogenous donor substrate for each cell type and biomaterials is time-consuming and may be difficult with multiple cells and material formulations.

Collagen-based inks are a suitable donor substrate due to cell biocompatibility and their potential for self-assembly and crosslinking. Laser-assisted bioprinting has been used to recreate skin substitutes [129,130] and corneal stroma-like tissue [131]. Additionally, in vivo bone regeneration was achieved by in situ printing of mesenchymal stromal cells using LAB (Figure 2C) [126].

#### 3.2.4. Stereolithography Printing

In stereolithography/digital light process (SLA/DLP) bioprinting, the ink is crosslinked by photopolymerisation. A reservoir of photo-sensitive ink is exposed to a predefined light pattern and crosslinked layer by layer onto a platform to produce a 3D structure [18]. The use of light patterns allow for high print resolution (50 µm) and accuracy [132]. Similar to nozzle free systems such as inkjet and laser-assisted bioprinting, SLA/DLP systems do not face clogging issues during printing. SLA accommodates inks with greater viscosities (<5 Pa) [133]. However, its restriction is the requirement for photopolymerisation crosslinking since not all materials are compatible for printing. Furthermore, photo-curing agents can be cytotoxic if residual components remain after printing [132]. Unlike previous methods, SLA/DLP bioprinting is unable to incorporate multiple ink formulations.

While collagen can be crosslinked by UV irradiation, on its own, it cannot crosslink sufficiently fast for viable bioprinting. This necessitates functionalisation of collagen molecules. Typically, free amine groups in collagen are replaced with methacrylate groups which can participate in free radical polymerisation (methacrylation). Additionally, this functionalised collagen retains the ability to self-assemble into fibrillar structures upon neutralisation. Modified collagen has shown successful 3D photopatterning of hydrogels loaded with human mesenchymal stem cells [134].

To aid the reader, Table 1 has been provided to summarise applications of additive bioprinting methods for collagen biomaterials/biocomposites and bioinks (cell-laden).



**Table 1.** Applications of additive bioprinting methods for collagen-based inks.




#### **4. Regulatory Considerations and Challenges for Collagen Biomanufacturing**

Currently, additive bioprinting methods have made significant progress using collagen biomaterials to repair severe skin wounds, regenerate cornea and (cranial) bone defects etc. In addition, precise spatial patterning of collagen biomaterials/biocomposites and bioinks (cell-laden biomaterials) can recapitulate complex tissue architecture for realistic in vitro testing. Being highly customisable, additive bioprinting will likely benefit the regeneration of hard- (bone) and soft- tissue trauma to kickstart tissue regeneration. Yet, regulatory and commercial aspects present a formidable bottleneck to their successful translation for therapy.

Taking bone tissue engineering (BTE) as an example, even after 25 years of research and 100's of \$ millions of federal research (in the USA alone), clinical progress is limited. For example, 75% of spinal fusion procedures performed still use traditional grafting methods, suggesting that limited clinical benefits were derived from recent tissue engineering research [137]. Yet, certain approved therapeutics such as INFUSETM from Medtronic Plc reap >\$750 Million in annual sales [138]. Thus, disparity between clinical translation success and failure is highly significant. This has been described as 'the valley of death' where promising technologies fail to transition into commercial usage. Past analysis suggests that translational failure can attributed to 2 stages: (i) between institutes of higher learning where fundamental research is carried out and industry, because promising ideas fail to attract sufficient funding to transition into industry and (ii) industry to clinical implementation—where funding is insufficient to complete human trials [139].

It might be instructive to consider regulations that govern the approval of therapeutics. In the USA, any prospective therapy would be assigned by the FDA to 3 centers: (i) regulate drugs (small-molecules, therapeutic proteins, antibodies and immune-modulators), (ii) regulate biological products (viruses, toxins, vaccines, blood components, cells, tissues gene vectors etc) (iii) medical devices. Separate offices of combination products, and cellular, tissue and gene therapies also have purview of the regulatory process. Further information is summarised in the review article by Pashuck & Stevens [138].

Broadly-speaking, therapies can be regulated as "drugs" or "devices" - a device does not "*achieve its primary intended purpose through chemical action (chemical reaction and*/*or intermolecular forces*)" [138]. These definitions have significant cost implications as new drug or biologic candidates cost approximately \$850 million taking 5–10 years [138], whereas premarket approvals (PMAs) for new medical devices cost between \$45–150 million and are typically completed within 5 years [138]. Notably, the PMA route is used for high-risk devices that require clinical safety and efficacy demonstrations involving approximately 1% of device applications. Accounting for a greater proportion, are lower-risk 510 (K) devices that utilise premarket notification (PMN) channels (\$1–50 million to develop). These need to demonstrate equivalence or substantial equivalence to an existing marketed device [138]. Thus, acellular biomaterial scaffolds versus combination bioinks laden with cells and/or chemical agents (e.g., growth factors) are regulated very differently.

One example is the role of collagen in the product Biobrane® which reportedly acts relatively passively while supporting wound healing [140]. On the other hand, combination products may have biologics and drug ingredients which require oversight from the office of combination products and/or office of cellular, tissue and gene therapies [138]. For example, bioprinting skin constructs to repair severe wounds may require adding growth factors with chemical activity to assist wound regeneration. This potentially hinders swift and cost-effective regulatory approval [141]. The "rule of thumb" in product translation is that increasing product complexity correlates with the number and magnitude of challenges that need to be overcome before regulatory approval [140].

Furthermore, cGMP (current good manufacturing practice) is a requirement for mass production and ISO 10993 tests are required to assess biocompatibility. For cGMP, design history (allowable ranges of physical properties - material, geometry, porosity, mechanical etc) and device history (testing to demonstrate manufacturing design criteria was met) files are required, along with related auditing costs. Biocompatibility testing on large preclinical animals may cost a further \$50 million prior to commencing human clinical trials [137]. One approach to cross this proverbial "valley of death" might involve developing technology in a modular manner. For example, development could begin with a minimally-modified biomaterial using the 510 (K) pathway to initiate revenue generation, before developing combination products suited for the PMA route. The likelihood of obtaining approval for the 2nd product with more complex features could be enhanced by the original (basic) product, because of its regulatory predicate [137].

A further consideration concerns differences between the EU and USA in regulating 3D bioprinted tissue engineering products. Whereas they may be considered biologics in USA, they are regulated as combined advanced therapy medicinal products (ATMPs) in EU. In general, the authors found that existing frameworks fail to address aspects of computer-aided 3D-bioprinting for additive manufacturing of customised tissue products [142]. They concluded, early and regular dialogue with regulatory authorities may alleviate these bottlenecks in manufacturing and quality development [142].

#### **5. Concluding Remarks**

As the most ubiquitous extracellular matrix material, collagen is an obvious candidate biomaterial with great promise for regenerative medicine. Collagen is a natural polymer with high biocompatibility, biodegradability and weak antigenicity [13]. Other benefits include: its evolutionary conservation [143]—suggesting it can be derived from many sources including (but not limited to) common commercial sources: rat tail, porcine tendon, bovine skin, fish skin etc. Thus, several xenogeneic acellular matrices have already obtained clinical approval [143]. Collagen is also extracted relatively easily, increasing the ease of availability. However, issues of ethical derivation and sustainability of collagen have arisen, which makes transgenic sources an attractive proposition [29]. Collagen is also a highly versatile biomaterial, denaturing into gelatin (and other derivatives), increasing crosslinking degree through chemical and physical means—rendering control over physical properties such as: mechanical stiffness, pore size and biodegradability. Its versatility extends to formulating biocomposites with inorganic and natural polymers to provide appropriate mechanical stiffness (e.g., PCL), gelation properties (e.g., alginates) etc. to develop suitable collagen bioinks and biomaterials for therapy.

Producing collagen-derived therapeutic and testing products with additive bioprinting methods provides significant benefits over non-additive production. Additive bioprinting exquisitely controls ink deposition, facilitating spatial patterning (mimicking the heterogeneity of skin dermis) [141], reproducibility, customisation, higher throughput, cost-effectiveness etc. [19]. On the other hand, non-additive methods like manual casting may limit product complexity and reproducibility, while electrospinning is limited in throughput and product complexity. These attractive attributes of additive bioprinting may significantly lower barriers to utilising collagen-based products in regenerative therapy and disease modelling etc. With increased process reproducibility, the inter-batch variability during manufacturing is likely to decrease, resulting in smaller tolerances reflected in its device master file (cGMP requirement). Therefore, strategic considerations of regulatory and cost issues in the application of additive bioprinting will help to ensure collagen biomaterials fulfil their tremendous potential in biomedicine and bioscience.

**Author Contributions:** Conceptualization, D.C.L.Y., and W.W.C.; software, D.C.; writing—original draft preparation, D.C.L.Y., W.W.C., V.T.; writing—review and editing, D.C.L.Y., W.W.C., D.C. and S.S.; supervision, D.C.L.Y., D.C.; project administration, D.C.L.Y., D.C.; funding acquisition, M.W.N. All authors have read and agreed to the published version of the manuscript. Please turn to the CRediT taxonomy for the term explanation.

**Funding:** This research was funded by SCIENCE AND ENGIEERING RESEARCH COUNCIL (SERC), A\*STAR grant number A18A8b0059.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

1. Frantz, C.; Stewart, K.M.; Weaver, V.M. The extracellular matrix at a glance. *J. Cell Sci.* **2010**, *123*, 4195–4200. [CrossRef] [PubMed]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
