Bioinks for bioprinting functional meniscus and articular cartilage

Abstract

3D bioprinting can potentially enable the engineering of biological constructs mimicking the complex geometry, composition, architecture and mechanical properties of different tissues and organs. Integral to the successful bioprinting of functional articular cartilage and meniscus is the identification of suitable bioinks and cell sources to support chondrogenesis or fibrochondrogenesis, respectively. Such bioinks must also possess the appropriate rheological properties to be printable and support the generation of complex geometries. This review will outline the parameters required to develop bioinks for such applications and the current recent advances in 3D bioprinting of functional meniscus and articular cartilage. The paper will conclude by discussing key scientific and technical hurdles in this field and by defining future research directions for cartilage and meniscus bioprinting.

Diarthrodial joints are characterized by a layer of hyaline cartilage that covers opposing bone surfaces. In joints such as the knee, they are structurally supported by the menisci. These cartilaginous tissues facilitate smooth articulation and load transmission across the joint surface. Both articular cartilage and menisci can become damaged due to trauma, disease or wear and tear of the joint. For example, meniscal tears are a prevalent injury, particularly in sports [1]. Current surgical interventions involve resection of the damaged area and in severe cases a meniscectomy. Both treatment options result in altered joint biomechanics with increased contact stresses [2–4] and a concomitant increase in the risk of developing osteoarthritis (OA) [5–7]. OA is generally associated with age, however, early symptoms/markers are increasingly being seen in younger patients. In 2004, the WHO positioned OA in the top ten diseases which cause work loss due to disability. Subsequent studies have demonstrated that incidences of OA had increased by 32.9% since the WHO report [8]. The current gold standard for treating advanced OA is the surgical replacement of the diseased joint with a metallic and polymer prosthesis, however, there are a number of well-documented limitations associated with this treatment [9]. The aging worldwide population, and the associated increase in the instances of joint degeneration, are motivating the need for new treatment options for damaged and diseased joints.

3D printing of biomaterials with living cells, biomolecules and extracellular matrix (ECM) components is emerging as a powerful technique to engineer new tissues and organs to replace those lost due to trauma or disease [10–14]. This technology allows for the accurate positioning of different cell types, growth factors, genes and other biological cues within 3D biomaterials to mimic the shape, structure, composition and ideally the function of specific tissues and organs. A key component to realizing this premise is selecting an appropriate biomaterial, or ‘bioink,’ which will support cell proliferation, differentiation and tissue-specific ECM deposition. A bioink can be defined as a material containing a living cell suspension that can be extruded during a printing process while maintaining cell viability [15]. Many of the hydrogels traditionally used in biofabrication and tissue engineering are being explored as bioinks for 3D bioprinting. Hydrogels are water swollen, cross-linkable materials that are used extensively in cartilage and meniscus tissue engineering [16–18]. They can support and promote chondrogenesis by facilitating matrix deposition and by providing physical and chemical cues mimicking specific cell niches. Key hydrogel design parameters for a tissue engineering applications include matrix stiffness, degradation rate, provision of cell adhesion motifs and biocompatibility [17]. These parameters directly translate to bioinks with one critical addition, printability. Printability is an umbrella term incorporating the resolution and the shape stability of a bioink. These parameters can be influenced by the bioink rheology, concentration and cross-linking mechanism.

Currently, there are four primary approaches to printing bioinks, extrusion (via pneumatic, piston or screw), inkjet (thermal or piezoelectric), laser or acoustic. The extrusion-based method involves applying a force to extrude the bioink through a needle in the form of a strand. Strand size can be controlled by the diameter of the needle and by varying the magnitude of extrusion force. Inkjet technology creates droplets of cell suspensions or bioinks using piezoelectric or thermal affects. This technology offers high spatial resolution, however, it is only compatible with low viscosity bioinks [11,19]. Laser-assisted and acoustic approaches are both nozzle-free and can provide high-resolution printing between 3 and 300 μm [20]. This review paper will not focus on describing the printing hardware or the extrusion technology; for detailed reviews on the these topics, the interested reader is referred to other recent review papers [10,11,15,21–27].

Bioprinting anatomically accurate, functional tissues and organs will involve developing bioinks that are both printable and provide encapsulated cells with an appropriate environment to perform their desired function. For load-bearing tissues such as articular cartilage and meniscus, bioinks may also be required to provide mechanical support while new tissue is being produced. Therefore, the ideal bioink for cartilage tissue engineering would be printable, mechanically functional and support robust deposition of cartilage-specific ECM components. This paper will provide a review of the different hydrogels used for both meniscus and articular cartilage tissue engineering that have been further developed as bioinks for 3D bioprinting. As will become clear, the ideal bioink for bioprinting of meniscus or articular cartilage has yet to be developed, motivating further research that will enable the bioprinting of implants capable of regenerating damaged and diseased joints.

Tissue engineering articular cartilage & the meniscus

Articular cartilage

Articular cartilage is an avascular tissue with a limited capacity for repair once damaged. It contains an intricate zonal structure and composition that is integral to the mechanical function of the tissue [28]. The tissue is often defined as consisting of superficial, middle, deep and calcified regions/zones, with each containing a specific distribution of ECM components. Recapitulating this zonal structure is one of the central challenges in the field of articular cartilage tissue engineering. Common to each zone are chondrocytes (the cell type in cartilage), water, glycosaminoglycans and type II collagen. The collagen fibers are orientated in an arcade fashion (often described as the ‘Benninghoff architecture’), running parallel to the articulating surface in the superficial zone and perpendicular in the deep zone. Current tissue engineering approaches have sought to recapitulate the structure and composition of cartilage with either primary chondrocytes or stem/progenitor cell seeded, synthetic [29], natural [30,31] and ECM-derived hydrogels and scaffolds [32,33]. A number of studies have also utilized scaffold-free approaches that support robust chondrogenesis [34–36]. Despite these successes, to date it has not been possible to tissue engineer mechanically functional cartilage grafts mimicking the zonal nature of the native tissue. Furthermore, scaling-up these engineered grafts to biologically resurface entire joint surfaces remains an ongoing challenge.

Meniscus

The meniscus is a mobile tissue that has several critical functions including the provision of stability and congruency during joint articulation, shock absorption and joint lubrication [37]. Unlike articular cartilage, there is some blood supply via the vasculature that invades the periphery of the meniscus, however, the central regions rely on nutrient diffusion from the synovial fluid. There are two menisci in the knee, the medial and the lateral. Both differ in shape and have differing levels of vasculature (10–25% in the lateral meniscus and 10–30% in the medial meniscus) [38]. Adult menisci contain highly organized collagen fibers in both radial and circumferential directions [2,39]. The cells within the meniscus have been defined as fibroblasts, chondrocytes and fibrochondrocytes [37,40–42]. These meniscus-specific cells, unlike articular chondrocytes, secrete abundant collagen type I in addition to collagen type II, III, V and glycosaminoglycans. Current tissue engineering approaches have focused on absorbable synthetic and natural scaffolds to deliver endogenous cells or recruit host cells, thereby promoting meniscus repair [43–46]. In many cases, the crescent shape of the meniscus has been recreated, however, these scaffolds and engineered constructs generally fail to recapitulate the spatially defined collagen network and mechanical properties of the native tissue.

Despite the similar functions of articular cartilage and the meniscus, they are uniquely structured to bear the different forces that they are subjected to during normal daily activities. Both tissues are formed from a condensation of skeletal stem cells, however, the final tissue is a product of the mechanical, chemical, signaling and oxygen environment they experience during pre- and postnatal development [38]. The fully developed tissues contain approximately 70% water, making highly hydrated hydrogels attractive as biomaterials to encapsulate cells for tissue engineering and biofabrication strategies.

Cell sources for articular cartilage & meniscus tissue engineering

Hydrogel encapsulation of the primary cell of each tissue (e.g., chondrocytes or meniscal fibrochondrocytes), usually at lower passages to prevent dedifferentiation, promotes tissue-specific matrix production [36,47–52]. An alternative cell source is the mesenchymal/multipotent stromal/stem cell (MSC), which can be derived from bone marrow stem cell, adipose tissue stem cell (ADSC), synovial membrane stem cell, synovial fluid stem cell and the infrapatellar fat-pad stem cell among many others [31,53–60]. MSCs are often selected as a cell source due to their multipotentiality, expandability and availability. There have been numerous studies demonstrating that MSCs undergo chondrogenesis in the presence of TGF-β3, especially in low oxygen conditions [32,53,55,61–63]. In the presence of CTGF, FGF and/or TGF-β3, MSCs have also been shown to undergo fibrochondrogenic differentiation [64–67]. A potential limitation of chondrogenically primed MSCs is their inherent tendency to undergo endochondral ossification and form bone [55,68]. For this reason, primary cells are more attractive for tissue engineering as they better maintain the desired phenotype, however, in many cases they are either unavailable or are only available in limited quantities, rendering them unsuitable for engineering large-scale tissues. An alternative approach is to use a coculture of a small number of primary cells with a larger number of stem cells [69,70]. Numerous studies have demonstrated the benefits of coculture for both cartilage and meniscus tissue engineering [71–74]. Perhaps the most promising cell source are tissue-specific stem/progenitor cells that have been shown to be resident in both articular cartilage [75] and meniscus [76]. Articular cartilage-derived progenitor cells have been shown to undergo robust chondrogenesis with limited evidence of hypertrophy and endochondral ossification, even after extensive monolayer expansion [77–79].

Bioink printability

Bioink printability is critical to the realization of a 3D-printed tissue or organ.  It is important to assess if the extruded filaments support accurate layer-by-layer deposition and to what degree of resolution. Additionally, the bioink must support cell viability and provide a suitable environment for tissue-specific ECM deposition (Figure 1A). Numerous hydrogels have been used for both cartilage and meniscus tissue engineering, however, not all are necessarily printable with current technologies. The printability is strongly dependent on the viscosity and the yield stress of the material. The ideal bioink has been defined as [26]:

Figure 1. Factors affecting printability.

(A) The biofabrication window, finding a balance between printability and biocompatibility [85]. Reproduced with permission of Springer. (B) Graph depicting the shear-thinning behavior (viscosity decreasing with shear rate) of a gellan gum–alginate bioink [86]. Reproduced with permission from John Wiley & Sons Inc. (C) Extrusion of gelatin–alginate strand illustrating printability, with undergelation resulting in a droplet morphology and a lattice with truncated corners, ideal-gelation results a smooth uniform strand with square lattice and overgelation results in an inconsistent strand thickness [82] © IOP Publishing. Reproduced with permission. All rights reserved.

• Solid gel before printing with a shear thinning but not a thixotropic behavior (time-dependent change in viscosity) to the viscosity that permits printing with the selected technology;

• Rapid gelation after printing to obtain high shape fidelity;

• Minimum or no swelling during or after extrusion.

In general, inkjet printer heads are more suited to depositing low viscosity materials, which do not clog the print head. When using an extrusion-based printer head, bioinks that exhibit non-Newtonian, shear thinning behavior print with good shape fidelity (Figure 1B). Shear thinning is the phenomenon that causes the polymer chains to align when a force is applied, thus decreasing the viscosity with increased shear strain. At the tip of the needle, when the strain is removed, the chains reassemble resulting in a stable strand. Potentially, one of the more overlooked parameters is determining printability is the yield stress of bioinks. Yield stress is defined as the point at which the material begins to flow. It has been shown that bioinks with high-yield stresses print filaments with greater shape fidelity than bioinks with the same viscosity but lower yield stress [80]. This occurs due to a steep viscosity drop after the yield stress has been overcome.

Parameters that can alter the viscosity of a bioink include biomaterial concentration [81], temperature or the number of encapsulated cells [82,83]. Several different techniques can be used to enhance the printability of bioinks without shear-thinning properties. Such approaches include combining components such as nanoparticles to tailor the rheological properties, printing sacrificial bioinks to act as a mould, printing into a bath of suitable cross-linking material or by pre-cross-linking prior to dispensing [84]. These techniques should be assessed with cells as in some cases achieving good control of the strand size may compromise cell viability [84].

Assessment of printability & cell viability

A number of simple methods have been used to gauge printability, such as comparing the printed geometry to the target geometry obtained from the original CT scans both immediately postprinting and after 24 h in phosphate-buffered saline [14] or to the area of the inputted design [87]. Alternatively, the ratio between the nozzle diameter and the extruded filament diameter can be measured, with a value of 1 considered ideal (Figure 1C) [63,82]. This approach should be conducted in both horizontal and vertical planes, assessing bioink spreading and sagging, respectively. Sagging can be evaluated by printing a filament across vertical supports/pillars of different line spacing and measuring the aspect ratio and degree of deflection [88]. A similar approach could be used to determine how two/multiple layers interact with each other. This could be used to judge whether filaments adhere to each other and if the bioink filament can support the weight of a second/third etc. layer(s). Nonadherence of sequentially deposited layers can be problem with certain bioinks, which is an important consideration when printing constructs of scale.

A key step in bioink development is assessment of cell viability postprinting. Viability is strongly dependant on the shear stresses that cells are subjected to during the printing process, with values more than 10 kPa shown to dramatically reduce cell viability [83]. The levels of stress developed during printing depend on bioink viscosity, printing pressures and needle diameters. Comparing levels of cell viability between different studies can be difficult as these printing parameters are rarely the same. For example, Chung et al. [89] investigated the printability of an alginate hydrogel and found that when combined with gelatin, a more consistent structure could be formed. The viscosity of the alginate-gelatin bioink was dependant on the temperature and cell viability could be preserved at extrusion pressures below 24 psi. In contrast, Nair et al. [90] reported that pressure has a negative effect on cell viability with pressures of 20 and 40 psi resulting in cell death of up to approximately 45%. One of the reasons given was that the increase in pressure resulted in an exponential rise in shear stress.

Another important consideration is establishing whether a homogeneous distribution of cells is maintained in the cartridge/syringe and in the printed filament. This is of particular importance for large-scale prints, which can be time consuming. An assay was developed to determine the ‘sedimentation coefficient’ whereby cells are labeled, encapsulated in the bioink and incubated for a period of time. The bioinks are then inverted and immediately imaged via confocal microscopy, with the coefficient determined based on the cell density per zone (a sedimentation coefficient of 1 equates to uniform cell density in each zones) [91]. Strategies to overcome high sedimentation coefficients include incorporating a stirrer in the cartridge, cross-linking layer by layer [92] or by including thickening agents such as xanthan gum [91].

Bioinks for cartilage & meniscus engineering

There are numerous materials which can be used to form hydrogels for tissue engineering applications. These can be categorized as natural or synthetic. Natural hydrogels include collagen, agarose, alginate, fibrin and gelatin, while synthetic biomaterials commonly used in tissue engineering include polyethylene glycol (PEG), polyethylene oxide and polyvinyl alcohol or composites thereof. The advantage of naturally derived hydrogels is that they have molecular or structural similarities to the native ECM. Synthetic biomaterials can be attractive due to their reproducibility and tailorability [93]. As stated previously, many hydrogels have been used for cartilage and meniscus tissue engineering with varying degrees of success. This does not predispose them to making good bioinks, further modification may be required. A brief overview of the most commonly used hydrogels for cartilage and meniscus engineering is described in this section, followed by a summary of their application as a bioink.

Natural bioinks

Collagen

As the main structural protein in the articular cartilage and meniscus ECM, collagen is commonly used to produce scaffolds and hydrogels for tissue engineering applications. It can be isolated from numerous biological tissue, retaining key signalling, adhesive or other biochemical cues. It can be degraded by metalloproteases present in the body and usually only elicits a minimal immune response. It can be cross-linked via temperature or pH changes; however, this can be time intensive and results in low mechanical integrity. For traditional scaffold manufacturing, improvements can be achieved by reinforcing with other materials and/or through the use of chemical cross-linkers such as dehydrothermal treatment, UV light, glutaraldehyde or 1-ethyl-3-(3-dimethylaminopropyl)carbodiimides [94–97]. As type I collagen is the dominant collagen type found in the outer region of the meniscus, it is commonly used to produce hydrogels and scaffolds for meniscus tissue engineering, with collagen meniscus implants currently used clinically [98]. Bioprinting collagen requires both control over the pH and temperature; in most cases, the barrel is maintained below 20°C [99–101] and the collection plate is maintained at 37°C (Figure 2D) [102]. However, printed collagen can take over 40 min to cross-link [103]. Increasing the concentration of collagen type I improves printability, with an optimal concentration reported as 17.5 mg/ml, higher concentrations are reported to clog the needle [104]. Lower concentrations of collagen have been printed either by inkjetting onto an agarose-collagen-coated plate [100] or into a sacrificial mould [105]. Collagen type I has also been blended with agarose to achieve 60% printing accuracy to the programmed path with an ink consisting of 1.5% agarose and 2 mg/ml of collagen [101].

Figure 2. Bioprinting various bioinks.

(A) Composite printing of multimaterial methacylated bioinks printed through a photopermeable lumen resulting in (i) a core–shell when both inks are extruded and the UV is switched on, heterogeneous structure when each ink is extruded simultaneously and a hollow structure (ii) the respective confocal image of cells labeled with different dyes with cross-section insert [119]. Reproduced with permission from John Wiley & Sons Inc. (B) Overhanging structures printed with agarose into a fluorocarbon bath [120]. (C) Alginate sulfate–nanocellulose bioink printed into the shape of an ear [121]. Reproduced with permission of Springer. (D) Sheep meniscus shape printed with collagen type I and seeded with fibrochondrocytes [102]. Reprinted with permission from ACS Biomaterials Science & Engineering. Copyright 2016 American Chemical Society. (E) Tubular 5% PEGDA Schematic printed on a rotating rod in a photopermeable capillary [119]. Reproduced with permission from John Wiley & Sons Inc.

Collagen type II, the predominant collagen type in articular cartilage, has been shown to support articular cartilage-specific matrix production [106]. It is less widely printed, however, it has been utilized alone as a bioink [107] and as a coating for 3D-printed PLGA fibers [108].

Gelatin

Gelatin is derived from denatured and partially hydrolyzed collagen, therefore it retains some of the bioactive cues such as the arginylglycylaspartic acid sequence which facilitates cellular adhesion, while matrix metalloproteinase degradation sites allow for cellular remodeling [109]. Gelatin is a thermoreversible hydrogel, gelling at temperatures below 25–35°C [19]. Methacrylated gelatin (gelMA) is commonly used as it overcomes the limitation of reversible thermal cross-linking by allowing covalent cross-linking in the presence of UV light. Properties such as stiffness can be tailored by altering the amount of methacrylation during fabrication [110]. GelMA is widely used for meniscus and cartilage applications as it supports both cell proliferation, migration and matrix production [109,111,112]. Scaffolds composed of gelatin and hyaluronan have been shown to support regeneration and integration of a section of a lapine medial meniscus [64]. GelMA has also been combined with digested meniscus or cartilage tissue in an attempt to promote tissue-specific ECM production, although no significant improvement was observed over gelMA alone [113]. Despite the use of gelMA for engineering both tissue types, it has been shown that it favors a more fibrochondrogenic phenotype, based on positive collagen type I staining [63], suggesting it is a promising bioink for meniscus bioprinting applications.

GelMA is widely used as a bioink. Printing at low temperatures provides temporary support and structure, while the subsequent application of UV light ensures long-term shape fidelity. At temperatures between 24.5 and 30°C, it is possible to print filaments whose diameter is dependent on the pressure, needle gauge, percentage material and plotting speed. Diameters of 150–200 μm are achievable with 10% gelMA, 30G needle, 24.5°C and 300–700 mm/min plotting speed [114]. Larger struts can be attained by decreasing the plotting speed and increasing the needle size, pressure and temperature. At 37°C, printing is challenging due to its liquid form and low viscosity, however, by blending gelMA with viscous components such as hyaluronic acid (HA), gellan gum, fibrin or PEGDA more stable constructs can be printed at these higher temperatures [115–118]. An alternative approach involves the application of UV light directly to the extruding filament during the printing process, rather than postprinting, as this allows immediate cross-linking of the gelMA fiber as it is extruded from the needle orifice (Figure 2A & E) [119]. This concept has facilitated spatial control over two different methacrylated materials to print complex shapes such as hollow tubes, core–shell structures and composites (Figure 2A) [119]. The printing of gelatin is widely reviewed in the literature, for a more detailed description of its printability the interested reader is referred to these dedicated review papers [19,109].

Agarose

Agarose is a polysaccharide polymer derived from seaweed and forms a thermoreversible hydrogel. In solution at high temperatures, agarose resembles a coil-like structure and during gelation (cooling) these form bundles and finally double helices. The structure can be controlled by the material concentration, with lower concentrations yielding hydrogels with larger pores and lower mechanical properties. Agarose has been shown to support the re-establishment of a chondrocyte phenotype [49] and chondrogenesis of stem/progenitor cells [31,122,123] both in vitro and in vivo [124]. Unmodified agarose does not support cell adhesion and therefore maintains the round cell shape of articular chondrocytes, however, it is less suited to supporting the fibrochondrogenic cells of the meniscus [125]. Bioprinting agarose can be challenging due to its thermal sensitivity; it is solid at 37°C and will block the extrusion needle. When maintained at higher temperatures (38–42°C), the viscosity is lower but cell viability can be compromised [126]. Printing at temperatures slightly above 37°C has been shown to maintain cell viability and construct fidelity, however, generating complex geometries or large constructs with definition is challenging to achieve when printed without a support material [63,126,127]. Examples of this approach include printing into a perfluorocarbon, which enabled columnar and overhanging structures [101,120,128]. This novel method prevents the sagging of agarose bioink under its own weight without affecting cell viability (Figure 2B) [120].

Alginate

Alginate is a natural water-soluble material primarily derived from brown seaweed. It is a linear polysaccharide copolymer of 1,4-linked β-D mannuronic acid (M) and α-ι-guluronic acid (G) monomers [93,129]. It is a biocompatible material with low toxicity levels and is easily available and fabricated. Ionically cross-linked alginate hydrogels can be formed with calcium ions such as calcium chloride, calcium phosphate or calcium carbonate [130]. The stiffness can, therefore, be tailored based on the degree of cross-linking and the material concentration. One disadvantage is that it is nondegradable as mammals lack the appropriate enzyme. Dissolution of the gel is, however, possible, occurring when the ions are released to the surrounding environment [130]. Alginate has been extensively used in the field of tissue engineering, and has been shown to support chondrogenesis of stem cells and chondrocytes both in vitro and in vivo [31,63,131–133]. Alginate has also been used extensively for meniscus tissue engineering [42,51,63,134–136]. In most tissue-engineered applications, the concentration of alginate used is between 1 and 4%. As a bioink, these concentrations are typically not viscous enough to be printable. One method to overcome this is to precross-link with a low concentration of calcium chloride prior to printing. For example, a printable ink can be achieved using 3.5% alginate partially cross-linked with 60 mM of calcium chloride at a 7:3 ratio. Following printing, the construct was stabilized by being immersed in a 50-mM calcium chloride bath [63]. Another strategy is to print directly onto a substrate or into a bath of calcium chloride, thereby enabling immediate cross-linking [137,138]. Both of these cross-linking mechanisms have been utilized for the printing of vascular structures; first, calcium ion precross-linking was undertaken followed by extrusion into a calcium chloride bath and finally a tertiary barium ion cross-linking for long-term stability of the alginate bioink in culture [84]. A third method for alginate bioprinting uses a coaxial needle that surrounds the bioink with calcium chloride, thereby coextrusion causes immediate cross-linking to form a stable fiber [139]. Others have simply combined alginate with materials such as gelatin [82,89,140], gellan [86], agar [141], nanofibrillated cellulose [142], nanocellulose (Figure 2C) [121] or printed the bioink alongside polycaprolactone (PCL) [13,143–145].

Fibrin

Fibrin is a protein involved in the clotting of blood that is used clinically as a sealing and adhesive agent. It is formed by the action of the protease thrombin on fibrinogen that causes it to polymerize. These components can be isolated autologously from a patient's blood, thus lowering the chance of an adverse reaction. The gel promotes cell proliferation and matrix synthesis and has been used as a growth factor delivery system for chondrogenic induction of MSCs [146]. Fibrin hydrogels have also been used clinically in the repair of meniscus tears [147]. A potential limitation of fibrin is that it can degrade quickly, in some cases before cells have an opportunity to secrete matrix [148]. This can be overcome by increasing the fibrinogen/thrombin ratios or by blending it with other hydrogels. There have been a number of approaches to printing a fibrin bioink. One is to combine cells into a thrombin solution and inkjet droplets into a fibrinogen bath [149]. This strategy has been used to engineer microvascular-like structures. An alginate/gelatin/fibrinogen blend has been shown to be printable while being mechanically robust, maintaining chemical cues and supporting cell viability and growth [150–152]. Additionally fibrinogen is a critical component of a bioink that was used to print human scale tissues including the external ear, which is a cartilaginous tissue [153]. Fibrin has also been used a bioink for laser-based printing [154].

Hyaluronic acid

HA is a nonsulfated glycosaminoglycan that is native to connective tissues where it is involved in lubrication, cell differentiation and growth [155]. As a hydrogel, it is both biodegradable and biocompatible. Cross-linking mechanisms include UV, dehydrothermal treatment and carbodiimide (EDC) [155,156]. HA supports both chondrogenesis of MSCs [157] and proliferation and matrix production of meniscus-derived cells [50,158]. However, this effect is concentration dependent, with HA levels of 1% or under yielding an increase in cell proliferation, sulfated glycosaminoglycan (sGAG) synthesis and an upregulation in sox-9 gene expression [159,160]. HA hydrogels have a high viscosity even with low concentrations, which has led to its use as an additive to others’ bioinks with low viscosities [115]. When printed on its own, HA has been methacrylated and subsequently UV cross-linked [87,119,161,162] or cross-linked with tetrahedral PEG tetracrylates to produce vessel-like structures [163].

Synthetic bioinks

Polyethylene glycol

PEG is a hydrophilic polymer used widely for drug delivery and biomedical applications [164,165]. It can be easily tailored and modified to suit specific applications, such as the incorporation of arginylglycylaspartic acid peptides for cell adhesion and modification to enable proteolytic degradation [166]. PEG can be methacrylated, forming a photocross-linkable hydrogel that is more suited to bioprinting applications. It is suitable for inkjet and extrusion-based technologies [92,119,167]. For example, chondrocyte-laden PEGMA has been inkjetted directly into an ex vivo cartilage defect where homogeneous cell distribution was achieved using layer-by-layer UV cross-linking [92]. An alternative to the postcross-linking approach to the printing of homogeneous filaments of methacrylated PEG is to extrude it through a photopermeable capillary in the presence of UV light [119]. Tubular structures can be fabricated by extruding onto a rotating rod, with the diameter of the rod defining the size of the lumen and the velocity of the rod and the ink tailoring the filament size (Figure 2E) [119]. To improve the printability of PEG-based bioinks, a high percentage of alginate (12.5%) can be added that has been shown to facilitate the generation of complex structures such as heart valves [14].

Advanced bioinks

While bioprinting of ‘traditional bioinks’ has yielded some successes, a new generation of bioinks are emerging that enable the engineering of more complex structures. These smart or advanced bioinks print with high shape fidelity, have shear-thinning characteristics, are mechanically robust, biomimetic and have the ability to modulate cellular functions [85]. These new and exciting bioinks include nanocomposites, multimaterials, interpenetrating networks (IPNs) and ECM-derived based or functionalized inks.

Nanocomposite bioinks include the addition of nanoparticles to provide desired physical, mechanical and chemical characteristics, most commonly shear-thinning properties and mechanical reinforcement [168–170]. Multiple-material strategies involve the use of fabrication systems to print multiple bioinks at once (Figure 3B) [171]. This enables tunable continuous multimaterials to be extruded. This technology could prove useful for printing the zonal structure of articular cartilage. IPNs, which are composite hydrogels that are physically entangled with each other, can be used to produce bioinks with increased elasticity and stiffness. For example, in a PEG-alginate-nanoclay composite, the combination of ionic and covalent cross-linking led to a highly stretchable material that could be strained up to 400% without fracture (Figure 3A) [172]. Collagen type 1, containing cells, was infused into the bioink, creating a cell-friendly environment with high levels of viability postprinting.

Figure 3. Advanced bioinks.

(A) (i) Elastic, highly stretchable of PEG-alginate-nanoclay printed in a bilayer mesh is stretched to three-times of its initial length followed by relaxation that demonstrates almost complete recovery of its original shape. (ii) A printed pyramid of the same material undergoes a compressive strain of 95% and returns to its original shape [172]. Reproduced with permission from John Wiley & Sons Inc. (B) (i) Setup of continuous multimaterial deposition in which seven different bioinks can be extruded simultaneously or alone by controlling the valve opening, (ii) printed helix and lumen structure using three different materials and a seven-material fiber [179]. Reproduced with permission from John Wiley & Sons Inc. (C) (i) Extrusion of decellurized ECM bioink alone and in conjunction with a PCL frame, (ii) live–dead staining depicting the cell viability after deposition, (iii) a scanning electron microscope image illustrating the ECM between the PCL fibers. Reprinted by permission from Macmillan Publishers Ltd, Nature Communications [177] © 2014.

ECM: Extracellular matrix; IPN: Interpenetrating network; PCL: Polycaprolactone; PEG: Polyethylene glycol.

ECM-derived scaffolds have been widely used in the tissue engineering field [173]. The tissue is decellularized and through several processes the components can be reconstituted to form a hydrogel or scaffold. One of the advantages of ECM-derived scaffolds is they maintain some of the biological, chemical and signally cues to direct cell fate. In general, they are not rejected by the body and cell proliferation, migration and differentiation is observed on these scaffolds [174]. Bone marrow stem cells seeded onto hydrogels containing ECM derived from the outer and inner meniscus have been shown to display a differential response to such cues, differentiating toward a fibroblastic and fibrochondrogenic phenotype, respectively [175]. Incorporation of ECM into bioinks has been done in a number of ways; either ECM particles (cartilage pieces, or ECM components HA, CS and collagen) are added into a bioinks such as alginate or gelMA or the ECM is itself directly printed (Figure 3C) [176–178]. The benefits of the former approach were nicely demonstrated in a seminal study by Pati et al. [177], who demonstrated that decellurised and solubilized cartilage, heart and adipose ECM could be used to develop printable bioinks that provide tissue-specific cues to direct the differentiation of encapsulated MSCs.

Bioprinting functional meniscus & articular cartilage

Many studies in the literature are making progress in the bioprinting of homogenous cartilaginous tissues, however, the complexity of their structures makes it significantly more challenging to print true articular cartilage or meniscus. Tables 1 and 2 outline the bioinks that have been targeted toward specifically articular cartilage or meniscus tissue engineering, respectively.

Table 1. Bioinks specifically targeted toward articular cartilage repair.

Bioink	Reinforced	Cross-linking mechanism	Printability analysis	Shape	Biofabrication technique	Cell type	Viability
PEGDMA [92]	No	UV	No	Filled an in situ chondral defect	Inkjet	Human chondrocytes	Yes–no quantification
PEG–gelMA [180]	No	UV	No	Cylinder	Inkjet	Human BMSC	85%
Alginate alginate–Cs alginate-CsHA [139]	No	CaCl2 + UV	Yes (shear rate and viscosity)	Geometrical shapes	Extrusion	BMSC	>85%
GelMA alginate [181]	Yes (PCL)	Temperature CaCl2	No	Cylinder	Melt electrowriting	Human chondrocytes	80%
Collagen I [182]	Yes (PCL)	Temperature	No	Humeral head	Extrusion	–	–
Gelatin–fibrinogen–HA–glycerol [153]	Yes (PCL)	Temperature + thrombin	No	Cube, ear, bone defect, muscle	Extrusion	Human AFSC, rabbit CC, mouse myoblast	91–97%
Alginate [183]	Yes (PCL)	CaCl2	No	Lattice cube	Extrusion	Human chondrocytes	Yes–no quantification
Alginate HA–collagen I [143]	Yes (PCL)	Temperature	No	Cylinder	Extrusion	MSC	No
GelMA Alginate Agarose PEG [63]	Yes (PCL)	UV CaCl2 Temperature	Yes (line spreading ratio)	Lattice cylinder	Extrusion	Porcine BMSC	Yes 80%
GelMA–gellan gum [80]	No	UV	Rheology and effect of bioink concentration	Lattice	Extrusion	Equine CC	–
Collagen type II [107]	No	pH and temperature	No	Cylinder	Extrusion	Rabbit chondrocytes	Yes–no quantification
Silk–fibroin Silk–gelatin [184]	No	Tyrosinase	Rheology	Square lattice	Extrusion	Human chondrocytes	Yes–no quantification
GelMa–PEGDA–graphene oxide [117]	No	UV	No	Lattice cylinder	Laser	Human MSC	No
Methacrylated PEG–HA PEG-Cs [161]	No	UV	Rheology	Lattice	Extrusion	Equine CC	85%
Fibrinogen–collagen type I [185]	Yes (PCL)	Thrombin and temperature	No	Square	Inkjet	Rabbit CC	No
Gellan gum–alginate–cartilage particles Gellan gum–alginate–HA [86]	No	Strontium chloride	Rheology	Nose Ear Dogbone Meniscus	Extrusion	Bovine chondrocytes	80%
Decellurized cartilage [177]	Yes (PCL)	Temperature and pH	Rheology	Square lattice	Extrusion	Human ADSC	>95%

ADSC: Adipose tissue stem cell; AFSC: Amniotic fluid-derived stem cell; BMSC: Bone marrow stem cell; CC: Chondrocyte; Cs: Chondroitin sulfate; CsHA: Chondroitin sulfate hyaluronic acid; GelMA: Methacrylated gelatin; HA: Hyaluronic acid; MSC: Mesenchymal stem cell; PCL: Polycaprolactone; PEG: Polyethylene glycol; PEGDA: Polyethyleneglycol diacrylate; PEGDMA: Polyethylene glycol dimethacrylate.

Table 2. Bioinks specifically targeted toward meniscus repair.

Bioink	Reinforced	Cross-linking mechanism	Printability analysis	Shape	Biofabrication technique	Cell type	Viability
Alginate [186]	No	CaSO4	Determined the error from design to printed	Meniscus from scan	Extrusion	Bovine chondrocytes	94%
Alginate–acrylamide–acrylated urethane [187]	Yes (acrylated urethane)	UV	Viscosity analysis	Dog bone and basic meniscus shape	Extrusion	Acellular	–
PCL [66]	Yes (PCL)	–	No	Meniscus from scan	Extrusion	–	–
Collagen type I [104]	No	Temperature	Determined the error from design to printed	Meniscus from scan	Extrusion	Bovine fibrochondrocytes	Yes–no quantification
GelMA Alginate Agarose PEG [63]	Yes (PCL)	UV CaCl2 Temperature	Yes (line spreading ratio)	Lattice cylinder	Extrusion	Porcine BMSC	Yes 80%
Nanocellulose–alginate [168]	No	CaCl2	Rheology	Meniscus	Inkjet	Human nasal chondrocytes	∼70%
Gellan gum–alginate–cartilage particles Gellan gum–alginate–HA [86]	No	Strontium chloride	Rheology	Meniscus Nose Ear Dogbone	Extrusion	Bovine chondrocytes	80%
Alginate–PLA nanofibers [188]	Yes (PLA)	CaCl2	No	Lines Medial meniscus	Extrusion	Human ADSC	Yes–no quantification

ADSC: Adipose tissue stem cell; BMSC: Bone marrow stem cell; GelMA: Methacrylated gelatin; HA: Hyaluronic acid; PCL: Polycaprolactone; PEG: Polyethylene glycol; PLA: Polylactic acid.

As stated previously, both meniscus and articular cartilage are responsible for load bearing within the knee joint. To engineer a successful biological implant, the construct must be mechanically robust enough to maintain integrity postimplantation and provide some level of mechanical function within the joint. There are three potential mechanisms to achieve this:

• Print the selected bioink in the shape of the tissue and chondrogenically prime it in vitro with the appropriate growth factors and/or mechanical cues until the cells have generated a functional matrix;

• Develop an advanced bioink such as an IPN, which can mimic the tissues of native mechanical properties;

• Print the bioink alongside a stiffer biodegradable polymer frame and then either directly implant or culture for a shorter time-period before implanting in vivo. This strategy is particularly attractive as it permits the printing of softer and less dense hydrogels, which typically better support chondrogenesis, alongside materials that provide both temporary mechanical support and potentially structural cues to printed cells to produce organized tissues.

Reinforcement of bioinks

Typically, low concentration hydrogels are used for cartilage tissue engineering, which on their own generally lack sufficient mechanical integrity for implantation into high-load-bearing environments. The Young's modulus of cartilage is in the range of 0.2–2 MPa [28,189,190] and the meniscus is 0.1–0.3 MPa [191], whereas the modulus of hydrogels that have been shown to support chondrogenesis are in the range of 0.2–40 kPa [63,96,111,123]. Increasing the material concentration will increase the hydrogel stiffness, however, the permeability will decrease and the matrix density will increase, which can negatively impact nutrient transport and waste removal. Furthermore, the hydrogel itself can become a barrier to ECM development [192,193]. Additionally, it has been demonstrated that stiffer substrates direct stem cell fate to the undesired hypertrophic or osteogenic lineage [194].

PCL is a polymer that is used clinically for drug delivery. It is a hydrophobic, biocompatible polymer with degradation kinetics spanning 3–4 years [195]. PCL is widely used in 3D printing; its low melting temperature (59–64°C) enables it to be extruded alongside bioinks without negatively affecting cell viability [63,145]. The mechanical properties of a 3D-printed PCL scaffold can be controlled by the material molecular weight and the filament pattern, spacing and diameter [196]. Hypothetically, this enables the development of a scaffold mimicking the bulk properties of native meniscus or articular cartilage. An excellent example of this involved the use of a gelMA composite reinforced by melt-electrospun PCL [181]. The melt-electrospinning technique enabled the production of small (<100 μm) diameter fibers with scaffold porosities of 93–98%. The authors observed synergistic increases in the mechanical properties of the composite material compared with the PCL scaffold alone, whereby the deformation of the fibers under axial loads were supported by the gelMA. By manipulating the fiber diameter and pore size, they were able to generate constructs with stiffness and elasticity approaching native cartilage [181]. To date, 3D printing of PCL has found use in the development of scaffolds for osteochondral defect repair. For example, MSC-seeded 3D-printed PCL scaffolds have been used to treat osteochondral defects in a rabbit femoral condyle, with no evidence of implant failure and improved joint repair compared with empty defects [197]. A similar study was conducted but with the PCL as the support for a bilayered collagen (bone region) and HA (cartilage region)-based bioinks (Figure 4C). 8 weeks after implantation in the trochlear groove, the cartilage region demonstrated significantly improved repair compared with empty, PCL only and a bilayered alginate, based on histological scores [143]. While such studies are promising, it should be noted that these focal osteochondral defects are much simpler to regenerate compared with treating a diseased joint where tissue damage is more diffuse.

Figure 4. Polycaprolactone printing.

Printing of anatomically accurate (A) meniscus embedded with PLGA beads that spatially released either TGF or CTGF to the inner and outer regions, respectively [66]. Reprinted with permission from AAAS and (B) a humeral head fabricated to facilitate cellular homing, which was implanted into a rabbit model [182]. Reprinted from The Lancet © 2010, with permission from Elsevier. (C) A reinforced osteochondral plug with spatially distributed bioinks of HA and atelocollagen for cartilage and bone regeneration [143] © IOP Publishing. Reproduced with permission. All rights reserved. (D) composite reinforced alginate bioink for endochondral tissue engineering printed in the shape of vertebrae, which supported mineral deposition and vasculature [13]. Reproduced with permission from John Wiley & Sons Inc.

HA: Hyaluronic acid; PLGA: Poly(lactic-co-glycolic acid).

3D-printed PCL has been an integral part of several new therapies for regenerating whole joint surfaces and entire menisci. In both cases, the geometries of the target tissues were obtained through CT scans. In the case of meniscus regeneration [66], a 3D-printed PCL scaffold loaded with growth factor-releasing microspheres was grafted in place of a resected meniscus in an ovine model (Figure 4A). The scaffold had a dynamic modulus slightly exceeding native values. No dislocation was reported, and the growth factor-releasing scaffold was found to facilitate spatially defined meniscus tissue deposition and to restore mechanical functionality after 3 months [66]. In the case of synovial joint regeneration [182], a PCL-hydroxyapatite composite was printed to mimic the geometry of rabbit shoulder, with TGF-β3 infused into the construct to recruit host cells and to promote cartilage repair (Figure 4B). After 4 months, animals had normal gait and histological sections demonstrated bone growth and the beginnings of cartilage repair. These studies demonstrate the potential of bioprinting strategies for whole joint regeneration, and provide a template upon which more complex strategies can be developed to treat diseased joints.

Another emerging strategy from the field of tissue engineering that is also finding an application in 3D bioprinting, is the engineering of developmentally inspired templates for tissue and/or organ regeneration [198,199]. An example of this was conducted for a bone tissue engineering application, in which MSC-laden alginate and PCL were printed simultaneously to produce mechanically reinforced hypertrophic cartilage templates in the shape of a vertebrae (Figure 4D) [13]. When implanted subcutaneously, these bioprinted tissues matured into vascularized bone organs.

Conclusion & future perspective

The field of 3D bioprinting has advanced significantly, and in theory enables engineering constructs that mimic the complex composition, architecture and mechanical properties of soft tissues such as articular cartilage and the meniscus. Hydrogels have been widely used in tissue engineering, but with limited translational success. This can be attributed, at least in part, to a failure to recapitulate the structure and mechanical properties of the native tissues. 3D printing enables novel strategies to engineer inhomogenous and anisotropic tissue structures, although challenges still remain in developing reinforced bioinks that recapitulate native mechanical properties while providing an environment conductive to cartilage or meniscus development. To this end, future research directions and outstanding challenges include:

• The controlled release (both temporal and spatial) of genes, growth factors and other regulatory biomolecules to promote stem cell recruitment and/or direct tissue-specific cellular differentiation in a spatially defined manner. This might involve incorporating delivery strategies such as drug/growth factor encapsulation or adsorption into/onto degradable hydrogels, polymers or microspheres [200–203] within printed constructs. Cell-laden microspheres have been previously incorporated into a bioink and represents a promising approach for growth factor delivery [204]. Gene delivery could provide greater spatial control by ensuring localized production of growth factors or other gene products within printed tissues. There are also exciting advances in technology that is allowing temporally defined biomolecule delivery, such as ultrasound-mediated drug delivery [205]. Incorporation of such technologies into a printed scaffold would provide temporal control of the growth factor deliver, thereby adding a fourth dimension to the printing process.

• Mimicking the spatial complexity of soft tissues could also be enabled by the spatially defined printing of ECM components and/or by depositing region-specific cells. Both articular cartilage and the meniscus have anisotropic distributions of collagen and sGAG, and this can be mimicked by spatially printing collagen type I and/or II as well as different proteoglycans in the appropriate region. Studies have already demonstrated that ECM isolated from different regions of the meniscus can appropriately direct stem cell fate [175]. This raises the question as to whether it is possible to tune the printed environment using spatially deposited ECM components to direct a single population of stem cells to produce spatially defined tissue. Alternatively, or perhaps in conjunction with spatial patterning of ECM components, zonal subpopulations of chondrocytes [206–210] from articular cartilage (the superficial, middle and deep) could be spatially deposited to recapitulate the native tissue. It has been demonstrated that each subpopulation of chondrocytes behave differently when cultured in vitro. Cells from the superficial zone secrete significantly more lubricin, while the middle zone cells deposited significantly more sGAG and collagen [206–208]. This approach has also been used for meniscus tissue engineering with cells from the inner and outer regions being isolated separately and cultured. Cells isolated from the inner avascular region have been shown to produce more GAGs compared with cells from the periphery [50,211,212].

• Given the limited numbers of primary cells, other cell sources must be considered for large-scale tissue bioprinting. Adult stem/progenitor cells are an attractive alternative due to their multipotency and widespread availability. However, promoting a fibrochondrogenic or chondrogenic phenotype while inhibiting hypertrophy remains a challenge. This points to the need to incorporate hypertrophy inhibiting factors into printed constructs. The literature has described several approaches including coculture with primary cells, delivery of PTHrP and tuning the mechanical environment by applying cyclic hydrostatic pressure or dynamic compression [69,71,213–215].

• The integration of printed engineered tissues to the surrounding host tissue is another consideration. For articular cartilage, this might involve printing osteochondral tissues that are easier to implant and retain within the bony bed. This, however, introduces a further challenge of spatially controlling angiogenesis, osteogenesis and chondrogenesis. Traditional techniques to fix meniscal implants include sutures, however, it remains unclear if this can be translated for use with bioprinted implants.

• Recapitulating native mechanical properties is one of the most significant obstacles for the development and application of bioinks. The human medial meniscus has an aggregate modulus and permeability of 0.11 MPa, 2.74 × 10^-15 m⁴/Ns posteriorly to 0.15 MPa, 1.84 × 10^-15  m⁴/Ns anteriorly. The tensile properties equally are location dependant, ranging from 138 MPa in the deeper regions of the tissue to 59.8 MPa at the surface [216]. The depth-dependant properties of articular cartilage have been well documented [28,189,217–219], with the superficial zone playing a critical role in preserving the dynamic modulus [217]. With the emergence of advanced bioinks that incorporate nanomaterials or have an IPN structure [220,221], the capacity to achieve complex mechanical properties increases. The challenge for these materials is not only mimicking native tissue properties but to also be biocompatible and to support chondrogenesis in vitro and in vivo. As discussed previously, one of the limitations with IPNs is they often contain a high polymer density which is not necessary conductive to ECM deposition and accumulation. One method to overcome this is to induce pores into the system, which has been effectively used before to promote bone formation [222].

• The final step will require printing the chosen bioinks and encapsulated cells at a human scale in anatomically accurate shapes. This may beget several challenges including maintenance of cell viability and ensuring appropriate tissue development throughout such constructs. Tissue engineering on a human scale may, therefore, require bioreactor culture postprinting, which can ensure appropriate nutrient transport throughout the construct during the in vitro maturation period prior to implantation.

In summary, 3D bioprinting and the development of bioinks has advanced at a rapid pace. There has been an evolution in recent years, from printing uniform hydrogels to printing complex IPNs, biomimetic structures and hybrid scaffolds. These bioinks have the potential to mimic many of the mechanical and biological properties of cartilage and meniscus, hopefully leading to the development of a new class of regenerative implant to treat a wider range of musculoskeletal defects and diseases.

Executive summary

Tissue engineering articular cartilage & the meniscus

• Despite significant advances, a mechanically functional tissue-engineered replacement remains elusive. 3D printing potentially enables the recapitulation of the complex extracellular matrix arrangement to promote a more native tissue formation.

Bioink printability

• Bioink printability refers to the controllability of filaments and cell survival throughout the printing process. Bioinks with shear-thinning properties extrude stable filaments, strategies for those without include, bioink modification, sacrificial fibers or cross-linking in situ.

Bioinks for functional cartilage & meniscus engineering

• Bioinks can be developed from traditional tissue engineering hydrogels but also there has been a surge in the development of more complex advanced bioinks with tailorable mechanical properties or biomimetic components. In some cases, bioinks have been reinforced to provide the mechanical integrity that a chondroinductive bioink lacks.

Future perspective

• 3D printing has advanced rapidly, one of the biggest challenges remaining is to print an anatomically accurate construct at a human scale that will have mechanical functionality and will integrate seamlessly into the synovial joint.

Financial & competing interests disclosure

The authors have funding from the Science Foundation Ireland grant number 12/IA/1554. The authors have no other  relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.