Physico-chemical Modification of Gelatine for the Improvement of 3D Printa- bility of Oxidized Alginate-gelatine Hydrogels Towards Cartilage Tissue En- gineering

This work explored 3D bioplotting to mimic the intrinsic hierarchical structure of natural articular cartilage. Alginate dialdehyde-gelatine (ADA-GEL) was used as a hydrogel ink to create hierarchically ordered scaffolds. In comparison to previously reported ADA-GEL compositions, we introduce a modified formulation featuring increased amounts of thermally modified gelatine. Gelatine was degraded by hydrolysis which resulted in tailorable printability characteristics further substantiated by rheological analysis. ADA(3.75%w/v)-GEL(7.5%w/v) with gelatine modified at 80 °C for 3 h could be printed in hierarchical complex structures reaching scaffold heights of over 1 cm. The hierarchical structure of the scaffolds was confirmed via μ-CT analysis. To examine mechanical properties as well as the suitability of the hydrogel as a proper matrix for cell seeding and encapsulation, nanoindentation was performed. Elastic moduli in the range of ~ 5 kPa were measured. Gelatine heat pre-treatment resulted in modifiable mechanical and rheological characteristics of ADA-GEL. In summary, this study demonstrates the possibility to enhance the printability of ADA-GEL hydrogels to fabricate hierarchical scaffold structures with shape stability and fidelity, without the necessity to change the initial hydrogel chemistry by the use of additives or crosslinkers, providing a valuable approach for fabrication of designed scaffolds for cartilage tissue engineering.


Introduction
Osteoarthritis is the most common joint disease of the adult human being. Mostly caused by inherent form-and function disorders, illnesses or accidents almost every 10 th adult human in Germany suffers from osteoarthritis [1]. In particular, women and men aged between 50 to 70+ years are affected. Within the course of the disease, the guiding symptoms are severe pain and the loss of function of joint units like knee-, hip-or shoulder joints due to the degeneration of hyaline cartilage tissue. Hyaline cartilage is the highly specialized connective tissue of diarthrodial joints. Its principal function is to provide a smooth, lubricated surface for articulation and to facilitate the transmission of loads with a low frictional coefficient [2]. Up to date osteoarthritis is not curable. Consequently, the goal of osteoarthritis therapy is focused on pain reduction and on long-term preservation of the hyaline cartilage tissue to secure its functionality [3]. If conservative therapies like physiotherapy and medical treatments cannot grant a normal daily routine for the patient, surgery is required. State of the art are autologous osteochondral transplants (AOT) as well as matrix-linked autologous chondrocyte transplants (MACT) to restore cartilage tissue. For both types of surgery cartilage tissue is taken out of a healthy host tissue environment. In the case of AOTs, healthy cartilage-bone fragments are placed upon the damaged area. However, the newly formed tissue by the AOT approach is mainly fibrocartilage, which is not useful for a long-term cure [4,5]. In the case of MACTs, on the other hand, chondrocytes get isolated from the sample, increased in numbers, linked to a matrix and are implanted in damaged cartilage parts. Thanks to this technique, hyaline-like tissue can be formed with similar characteristics as native cartilage tissue [1]. The use of three-dimensional (3D) scaffolds in combination with chondrocytes, however, could rapidly accelerate the healing process thanks to the manifold applications of scaffolds as space-filling agents, as delivery vehicles for bioactive molecules and as three-dimensional structures that organize cells and feature stimuli for an even better-directed formation of the desired tissues [6]. To manufacture 3D scaffolds, biofabrication techniques attract increasing interest worldwide [7]. Especially in the fields of regenerative medicine and tissue engineering, biofabrication has great potential thanks to the numerous opportunities that are offered by additive manufacturing technologies such as 3D-bioplotting. Using such technologies enables scientists to manufacture complex tissue constructs consisting of the plotted material as well as encapsulated cells which leads to better biomimetic approaches to replicate native tissue [8]. However, as biofabrication represents an advanced tissue engineering approach, the current state-of-the-art therapy in the clinic involves still the seeding of cells on pre-fabricated scaffolds. Hence, the development of suitable materials for hierarchical bioprinting to manufacture advanced scaffold matrices for MACT is a current major challenge that must be overcome. For this application, hydrogels are promising materials because they can mimic the microenvironment of natural tissues and therefore can promote cell attachment, growth, and proliferation [9]. In this context, specifically naturally-derived hydrogels provide important characteristics like biocompatibility and promising cell-material interactions that path the way for developing a native ECM (extracellular matrix) analogue structure [6]. Synthetic approaches, in contrast, may often lack integrin-binding ligands which hinder cell attachment and proliferation [10]. Since the ECM is mostly composed of polysaccharides, glycosaminoglycans, and various proteins, ADA-GEL (alginate dialdehyde-gelatine) with a polysaccharide as well as the collagen derived component gelatine (Fig. 1) provides a promising matrix for biomimetic tissue engineering applications [11]. The potential of ADA-GEL for bioprinting approaches has been demonstrated [12][13][14]. However, mechanical demands in physiological conditions remain challenging for hydrogels with thermoresponsive characteristics. Therefore, the mechanical properties of ADA-GEL systems need to be tailored for the specific application. Due to physico-chemical modifications of the structure, crosslinking, and the use of additives it is possible to obtain hydrogels with modifiable viscous, rheological, and mechanical properties [11,15]. Potential additives include micro-and nanofibers as well as micro-and nanoparticles. Nanocellulose in particular is being investigated as a promising additive, as not only mechanical properties could be improved with it, but also cell proliferation, bioadhesion, and viability [16,17]. In a dualcrosslinking approach, we have shown that the mechanical properties [18] and crosslinking of ADA-GEL can be tailored by crosslinking the alginate-based component using CaCl 2 and the gelatine component enzymatically with microbial transglutaminase [19]. This work aimed to investigate the possibility of recreating the intrinsic hierarchical structure of articular cartilage found in-vivo by a bioplotting approach in-vitro. Due to its promising characteristics regarding biocompatibility, biodegradability, and cell-material interactions, ADA-GEL is used as a hydrogel-ink [20]. Hydrogel precursor modification via temperature pre-treatment of gelatine was examined regarding its effect towards modifiable mechanical and rheological characteristics of the ADA-GEL ink (Fig. 1). Various compositions of ADA-GEL were tested regarding optimised printability characteristics. Subsequently, the modified ADA-GEL hydrogels were compared to prior investigated ADA-GEL compositions which had been successfully developed for bio-fabrication as well as scaffold fabrication [11] using printability assessments, nanoindentation, and rheological characterization. To investigate the impact of the thermal pre-treatment of gelatine on a molecular level Fourier-transform infrared spectroscopy (FTIR) as well as Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-Page) were conducted.

Material Synthesis (Oxidation of Alginate)
Covalently crosslinked ADA-GEL hydrogel was synthesised as described by Sarker et al [20].
Briefly, ADA (alginate-di-aldehyde) was prepared by the oxidation of alginate (VIVAPHARM® alginate, PH176, from brown algae, pharmaceutical excipient grade, JRS PHARMA GmbH & Co. KG, Germany) using sodium (meta)periodate (NaIO 4 , Sigma-Aldrich, Germany) as an oxidising agent in an ethanol-ultrapure water mixture (1:1). The suspension was stirred for 6 h at 22 °C (room temperature, RT) under the complete absence of light. The oxidation reaction was quenched by adding ethylene-glycol (VWR Chemicals International) and stirred for additional 30 minutes. The resultant suspension was transferred into several dialysis molecular porous membrane tubes (MWCO: 6-8 kDa, Spectrum Laboratories, USA) and dialysed against ultrapure water (UPW, Milli-Q ® , Merck Millipore, Germany). UPW was changed twice a day during five days of dialysis. After dialysis, the ADA product was frozen for at least 48 hours before it was transferred into the freeze dryer (ALPHA 1-2 LDplus, CHRIST Gefriertrocknungsanlagen, Germany) for lyophilisation.

Biomaterial Ink Formulation
To produce ADA-GEL hydrogel-inks, ADA and gelatine (Type A, derived from porcine skin, gel strength 300, Sigma-Aldrich, Germany) solutions were mixed. For the preparation of 15 wt/vol% gelatine stock solutions, the gelatine was dissolved in UPW at 37, 70, 80 and 95 °C respectively while vigorously stirring for 10 min, 3h or 6 h. All preparations are depicted in

Fourier Transfer Infrared Spectroscopy (FTIR)
An attenuated total reflectance FTIR (ATR-FTIR) spectrophotometer (IRAffinity-1S, Shimadzu, Japan) was used to evaluate the impact of the temperature pre-treatment of gelatine and the crosslinking between ADA and gelatine on the chemical structure. Dried films were used to record ATR-FTIR spectra. Films of gelatine and ADA-GEL were created by casting the corresponding hydrogels into 1.5 ml microtubes (Sarstedt, Germany) and dried for 3 days at RT.

Electrophoretic Analysis
SDS-PAGE was performed to evaluate the impact of thermal pre-treatment on the molecular weight distribution of gelatine. Separating and collecting gels were prepared one day before the experiment. The separating gel was prepared and poured into the sample holder (5 ml) before the polymerization of the solution. It was then polymerized for 1 h. Afterwards, the collecting gel was prepared and added into the sample holder (approx. 1.8 ml) with a further 45 min polymerization step. Gelatine samples were diluted to a protein concentration of 2.3 mg/ml. 15 µl of the sample solution was added to 5 µl loading buffer and mixed extensively.

Rheological Characterization
Rheological properties of hydrogel-ink formulations were measured using a controlled stress

Nanoindentation
Mechanical testing was performed to determine the effective Young's moduli (E eff ) of the hydrogels using a Piuma Nanoindenter (Optics11, Netherlands) equipped with a boro-silicate subsequently also used as an optical medium during nanoindentation.

Printability Assessment
Scaffold designs were drafted using the scaffold fabrication software "ScaffoldGenerator" (GeSim GmbH, Germany) as well as G-code and 3D-CAD/CAM-Software Fusion 360 (Autodesk, US). To assess the printability of the hydrogel-inks, single-strand structures and two-dimensional grids were printed using a micro-nozzle with an inner diameter of 410 μm (Nordson Corporation, US). The printing speed was set to 2 mm/s. Air pressures were adapted to the different hydrogel-ink formulations (Table 1) to provide a continuous material flow.
Single-strand structures, square two-dimensional grids, and square 3D scaffolds had an edge length of 15 mm. Five struts were plotted over the edge length. The number of layers in zdirection was adapted to the different hydrogel-ink formulations to assess the maximum printable height. Scaffold geometry and structure were analyzed using macroscopic and light microscopy images (Stemi 508, Carl Zeiss Microscopy GmbH, Germany). Image analysis was performed using ImageJ to assess strand thickness, evaluated for n = 10 individual strands, and printing accuracy. The printed areas of square 2D grids with a determined edge length of 15 mm (A i [mm 2 ]) were compared to the designed grid area (A = 225 mm 2 ) to calculate the percentage printing accuracy for each sample using the following equation: (1) G-code was written to design a two-dimensional structure to evaluate the possible resolutions of the hydrogel-ink formulations. The distances between strands were varied from 2 mm up to 0.5 mm (Fig. 6 A) in a square structure to assess at which distance strut merging could occur.
A uniformity factor was determined to classify printed strand uniformities and compare them to a theoretically perfect uniform strand [13,21]. For each hydrogel ink, one layer was printed with optimised extrusion pressure and a velocity of 2 mm/s. Each layer was imaged using light microscopy and images were evaluated using ImageJ. The outer edges of three printed strands per hydrogel-ink were outlined and measured. The length was then divided by the length of a theoretical, perfectly uniform strand (straight lines next to the strands) to obtain the uniformity factor U (Eq. 2): Figure 7 H shows a nonuniform (U > 1) and a uniform strand (U = 1) [13]. The pore factor (Pr) was determined to compare printed pores with ideal square pores.
2D-printed hydrogel grids were imaged using a light microscope and evaluated using ImageJ.
Three pores were measured (perimeter and pore area) for each hydrogel composition (Table   1) and the pore factor was determined (Eq. 3). A pore factor Pr<1 corresponds to an undergelled, Pr=1 corresponds to a properly-gelled, and Pr>1 corresponds to an over-gelled material [13].

Micro Computed Tomography (µ-CT) analysis
X-ray computed tomography (CT) measurements were performed using the "CTportable160. calculates the number of voxels in the pore, the centre of mass of the pore in x,y,z coordinates, the aspect ratio, and the normalized pore diameter for each detected pore.

Statistical Analysis
Data are expressed as mean ± standard deviation (SD). Statistical analysis was performed using Origin2016G software (OriginLab Corporation, USA). All experiments were performed using a minimum of n = 3 replicates. Normality tests and analysis of variance homogeneity were performed using the Shapiro-Wilk test. Statistically significant differences between means were determined at a value of p < 0.05, as determined by the Bonferroni post-hoc test using one-way analysis of variances (ANOVA) testing. Different significance levels (p-values) are indicated with asterisks and the specific p-value is provided in each figure legend.

Denaturation of Gelatine via Heat Treatment
Temperature pre-treated gelatine was examined via FTIR to gather structural information and data about intra-and intermolecular dependencies (Fig. 2). The FTIR spectra of pre-treated gelatines showed characteristic absorption bands of polypeptides and proteins including amide  All samples modified by the heating pre-treatment showed no changes in wavenumbers for amide I, amide II and amide III peak positions indicating an intact primary protein structure of gelatine. In contrast to the gelatine spectra, ADA-GEL spectra featured extra peaks which are residues of the alginate at 1027 cm -1 due to C-O-C stretching and at 1408 cm -1 due to asymmetric COOstretching [25]. Amide I (1625 cm -1 ) and II (1538 cm -1 ) bands shifted and broadened indicating the additional formation of C=N bonds absorbing at 1538 cm -1 which is a sign of Schiff's base formation and therefore covalent crosslinking of ADA and gelatine [26].
However, amide III bands could hardly be seen most likely due to overlapping signals of gelatine and ADA. Yet, the pre-treatment of gelatine precursor solutions showed no changes in the wavenumbers of the ADA-GEL spectra indicating the unaltered structure of amide I and II motifs contributing to the Schiff's base formation with the aldehyde groups of ADA. Unaltered amide I and II motifs were also obtained by FTIR analysis by Hoque et al. examining the thermal treatment of cuttlefish skin gelatine [27]. In conclusion, from the FTIR study, it was hypothesized that a temperature increase firstly manipulates the tertiary and secondary protein structure leading to an irreversible denaturation upon further heating [28]. Intramolecular hydrogen bonds are broken by the heat treatment and substituted with hydrogen bonds to water molecules which prohibit the reformation of the native tertiary protein conformation [28].
Nonetheless, FTIR results indicate an intact primary structure of gelatine independent of the dissolution temperature, allowing the formation of covalent crosslinks between oxidized alginate and gelatine via Schiff's base formation, as described before [29]. The electrophoresis patterns of heat pre-treated gelatines (protein concentration of 2.3 mg/ml) are shown in Fig. 3 A. Gelatine dissolved at 37 °C served as control. The molecular weight distribution fitted the manufacturers' specification of approximately 100 kDa [30].  (Fig. 3 B).
Intensities in high molecular weight regions (> 55 kDa) decreased with increasing pretreatment temperatures and holding times. This indicated a decrease in large molecular weight molecules and therefore a degradation of gelatine. Due to the absence of enzymes, the main factor for degradation was identified as hydrolysis due to the temperature treatment during the preparation of the gelatine solutions. Similar results were achieved by Hoque et al [27].
Cuttlefish skin gelatine was heat-treated for 30 min at 40, 50, 60, 70, 80, and 90 °C. It was observed that gelatine degradation occurred using heat treatments higher than 70 °C [27]. In addition to incubation time and temperature, Van den Bosch et al. proved that gelatine concentration and solvent (type and concentration of salt ions and pH) have an additional impact on gelatine degradation [31]. In this work, gelatines with concentrations of 15 %(w/v) were used for temperature pre-treatment in the absence of additional ions. In summary, SDS-PAGE revealed that gelatine was susceptible to degradation due to the pre-treatment at elevated temperatures. Hence, it could be assumed that a pre-treatment of gelatine would allow a possible tunability of material characteristics.

Rheological Assessment
Rheological measurements were performed to determine the time-dependent viscosity, yield stress, shear rate-dependent viscosity, temperature-dependent viscosity, and the recovery behaviour of the chosen hydrogels (Fig 4). ADA-GEL hydrogel precursors were formed via mixing at 37 °C. Since all bioprinting procedures were performed at RT, a gelation process of the gelatine occurred due to the cooling-off. Hence, a time-sweep was carried out to evaluate after which time a hydrogel-ink could be considered printable. The complex viscosities of the ADA-GEL increased over time reaching a plateau region eventually (Fig.4 A). This could be attributed to the gelling of ADA-GELs during cooling. Initially, isolated colloidal aggregates, which were present in the suspension, crosslinked to form a 3D, highly-viscous network [32].
Thus, the progress of complex viscosity over time depended on polymer concentration and degree of polymer denaturation (Fig. 4 A). Similar results were found by Ouyang et al [33]. It viscosities for the ADA-GELs (Fig. 4 F). Hence, the results indicated that printing pressures had to be adjusted during continuous printing. shear rate (Fig. 4 B). Information about shear thinning is crucial for the printability of hydrogels since it is responsible for a smooth material extrusion without clogging [36]. The degree of disentanglement depended on the shear rate. At sufficiently high shear rates the polymer chains could be completely disentangled and fully aligned. In this range, the viscosity would be independent of the shear rate featuring plateau-like regions like it was seen for each ADA-GEL sample (Fig. 4 B) [37]. The recovery behaviour of hydrogel inks was especially important for the post-printing behaviour of each printed strand. Physically, recovery allows the material to rapidly increase in viscosity after extrusion and to maintain a high shape fidelity [35]. to the highest gelatine denaturation by the temperature treatment, proving its impracticality for further printing. However, all ADA-GEL hydrogels were not able to fully recover in the following 100 s after extrusion (Fig. 4 E). In general, longer recovery times led to a decrease in the retention of cylindrical fibre formation and shape fidelity (Fig. 6 E). To achieve printability improvements, the ADA concentration could be increased to gain a denser polymer network with more possibilities for entanglements and therefore faster recoveries. Temperature sweeps were conducted to characterise the thermoresponsive properties of the ADA-GELs.
Thermoresponsivity could be employed to cause changes in viscoelastic properties required for successful extrusion and solidification into 3D constructs with high shape fidelity after printing. Thermo-responsivity was characterised by a change in a material's viscoelastic properties as a result of temperature change (Fig. 4 D). During bioprinting, temperature enables the control of the hydrogel viscosity. An increase of viscosities could be seen for all ADA-GELs upon cooling (Fig. 4 D, ramp from 40 °C to 20 °C), due to increasing entanglements and alignments of polymer chains trying to restore their natural conformation upon cooling.
Hence, initially isolated colloidal aggregates present in suspension crosslinked to form a threedimensional, viscoelastic network featuring higher viscosities [38]. Since the temperature sweeps were conducted by cooling from 40 °C to 20 °C viscosity values at 25 °C appeared relatively low in comparison to other sweeps since gelation still had to take place. Regarding biofabrication, previous studies have shown that adding high cell numbers (> four million/ml-1) significantly reduces final ADA-GEL hydrogel stiffness [39]. It has been shown for a gelatinbased hydrogel that increasing cell concentration reduces bioink viscosity [40]. As a result, those studies imply that our biomaterial ink may show reduced viscosity for high cell numbers, which may impact the printability and rheological properties of the here presented hydrogel composition. In this study, we focused on the engineering and design of an ADA-GEL composition with increased printability to produce biomaterial scaffolds that do not contain cells. In future studies, we will assess the influence of higher cell concentrations on the final hydrogel rheological properties, which will provide important implications for future biofabricaton approaches with the here presented, optimized ADA-GEL matrix. calcium primary cross-links the guluronic residues. Because of high guluronic acid contents in the alginate, the gel strength could be dramatically increased [43]. Since in this work alginate was changed to GRINDSTED ® Alginate PH 176 featuring a molecular weight of 250000 g/mol but unknown guluronic acid/ mannuronic acid ratio, it was considered that this alginate had a significantly lower guluronic acid content, which could result in lower Young's moduli. Studies have shown clear correlations between elastic moduli of hydrogel matrices and proliferation as well as differentiation of encapsulated cells [44]. Yet, Zehnder et al. [11] followed a bone tissue engineering approach in contrast to this work which made higher moduli necessary for a suitable bone cell environment. The elastic modulus of mammalian chondral tissue (matrix around chondrocytes) (~ 25 ± 5 kPa), however, is lower than the one of bone (cortical bone ~ 15 ± 5 GPa) [45,46]. In comparison to these values, elastic moduli of ADA-GEL samples crosslinked with CaCl 2 are in the range of soft tissue elastic moduli (~ 4 ± 2 kPa) [45]. This indicates a suitable environment for cell encapsulation for cartilage tissue engineering since both matrix stiffness and composition are important for the retention of tissue-specific cell functionality [47]. However, the optimal value of the elastic modulus may depend on the cell type.

Printability
Heat pre-treatment of gelatine solutions significantly affected the printability of ADA-GEL formulations. These formulations were evaluated regarding printing accuracy, strand width, resolution, printing pressure, uniformity factor (U), pore factor (Pr), gelation times, and shape stability as well as fidelity in the z-direction. The material selection of ADA-GELs showed printing accuracies of over 90 % with slightly higher values compared to the ADA-GEL-T Ref due to the higher concentrations of modified gelatine which led to a greater amount of polymer chain interactions like entanglements and crosslinks in the hydrogel (Fig. 7 A). Due to higher degrees of degradation gelatines dissolved at higher temperatures led to smoother strand appearances (Fig. 6). A smoother strand appearance of an ADA-GEL with highly denatured gelatine content could be expected (Fig. 6 D) since gels of lower viscosities showed a more fluid-like behaviour and broadened strands due to a less dense polymer network and fewer entanglements of polymer chains [48]. The thinnest strands of approximately 0.52 mm were printed with formulations of ADA-GEL-T 70°C_6h and ADA-GEL-T 80 °C_3h . By using a precision tip needle with an inner diameter of 0.41 mm, the logical consequence was that the thinnest possible strand matched this exact value. The thinnest strands observed were 29 % larger than this inner diameter. However, this width increase could be expected due to gravity and surface energy and tension [49]. Increases of strand diameters in 3D structures could be further explained by the additional weight of the higher layers. Resolutions (= required strand distances to avoid merging strands) were determined using self-designed printed resolution structures (Fig. 6 A). Resolution structures supported the results of strand widths and accuracies. ADA-GEL-T 80°C_3h featured the lowest possible distances between printed strands (≈ 0.68 mm) to avoid strand merging in the same layer with significant differences to ADA-GEL-T 80°C_6h , ADA-GEL-T 95°C_3h , ADA-GEL-T Ref . Low viscous ADA-GELs with high denatured gelatine content further led to higher gelation times and lower printing pressures (Fig. 7 D+G).
Since encapsulated cells in hydrogel-inks experience shear forces during the printing process, lower printing pressures could be favourable [50]. The comparatively low printing pressure of the reference composition might be traced back to its lesser content of not pre-treated gelatine.
No significant differences (0.05 level) were found comparing the uniformity factors of the examined ADA-GELs (Fig. 7 E). Uniformity factors (U) ranged from 1.02 to 1.08 showing only slight deviations from a perfectly uniform strand (U = 1). Soltan et al. [13] slightly higher uniformity factors were found in this work. This result could be explained with an overall higher final concentration of ADA and gelatine, as well as higher printing pressures in combination with lower printing speeds (2 mm/s compared to 25 mm/s).
ADA-GELs with gelatine content heat pre-treated at 80 and 95 °C exhibited mean uniformity values which were closest to a perfect uniform strand (Fig. 7 E). This result is in accordance with the qualitative uniformity comparison of heat pre-treated 3D printed gelatine strands of Kolesky et al [51]. Gelatine heat-treated at 95 °C was shown to exhibit more uniform strands after 3D printing than gelatine heat-treated at 70 °C. Pore factors (Pr) ranged from 1.02 to 1.23. al. [13], pore factors were closer to 1. Soltan et al. [13] reported a mainly under-gelled ADA-GEL behaviour most likely due to lower final concentrations of ADA and gelatine as well as no rheological tunability using gelatine denaturation. To assess shape stability and fidelity in higher layers, high multilayer structures were printed. Only ADA-GEL-T 80°C_3h provided the necessary shape stability to maintain its structure throughout the whole printing process of over 50 layers achieving heights of over 1 cm (Fig. 6 F). Gelatines treated at 80°C for 6 h or in general at 95 °C resulted in formulations with deficient viscosity to achieve scaffolds with a sufficient height or stability. This was an expectable result since already the 2D grids featured a fluidic behaviour expressed by thick and merging strands which prohibited shape stability and layer stacking ( Fig. 6 A-D). After the evaluation and consideration of all requirements, ADA-GEL-T 80°C_3h was found to possess the best printing characteristics. In comparison to the reference ADA-GEL composition of Zehnder et al. [11], higher printing accuracies, better resolution and higher shape stability could be achieved by composition adjustments as well as structural modifications of the precursors. Hydrogel ink development as well as printing parameter adjustments were the crucial first steps to enable hierarchical 3D printing as a biomimicry approach towards cartilage tissue engineering. The first approach for hierarchical printing was the creation of a scaffold featuring layers with varying densities from bottom to top. Furthermore, to overcome the usual layer stacking approach of 3D plotting a programme was written via G-code. The G-code enabled hydrogel extrusion during a print-head movement in z-direction. Hence, a standard scaffold structure was printed on a dense bottom layer to support vertically printed pillars (Fig. 8 C+D). The dense bottom layer resembles the superficial zone whereas the pillars mimic the deep zone of natural cartilage tissue (Fig. 8 E). µ-CT analysis revealed the micro-and macroporous structures of critical point dried 3D printed ADA-GEL 80°C_3h scaffolds throughout XY (top view) and XZ (side view) planes (Fig. 9). The comparison of µ-CT images of various XY and XZ planes of a scaffold offered information about its porosity gradient. In this work, µ-CT data of a scaffold reference with consistent strut density was compared to the three-layered hierarchical scaffold ( Fig. 9 A+B). Despite its homogeneous design, the reference cylinder featured a heterogeneous porosity gradient.
Comparing its upper XY 130 plane with the subjacent XY 180 plane it becomes apparent that lower layers were consolidated (Fig. 9 A). In numbers, the pore diameter decreases from 1.29 mm to 0.61 mm (Fig. 9 C) as well as the overall porosity from 64.73 % to 37.76 % (Fig. 9 D).
This consolidation process could occur due to the gravitational force acting on printed strands and due to the additional weight of the higher layers. As a consequence, the scaffold structure appears bent due to strut sedimentation. The µ-CT evaluation of the examined hierarchical structure demonstrated how the physicochemical modification of gelatine improved the 3D printability of the ADA-GEL ink and enabled the tunability of the internal scaffold porosity gradient. This is an important characteristic of a biomimetic cartilage tissue engineering approach [2]. XZ and XY µ-CT images displayed decreasing pore diameters (from 1.45 mm to 0.23 mm) as well as decreasing overall XY plane porosities (from 81.66 % to 30.98 %) from top to bottom (Fig. 9 C+D). This achieved tunability of pore sizes met tissue engineering demands. Pore sizes in the range of 20 µm -500 µm are crucial in tissue engineering since they are suitable for cell ingrowth, bone regeneration as well as vascularization [52][53][54].
Furthermore, pore sizes and porosities were in the same order of magnitude compared to other 3D scaffold fabrication techniques for cartilage tissue engineering. For example, Tamaddon et al. [55] showed that freeze-dried type II collagen scaffolds with pore sizes ranging from 50 µm -300 µm and an overall porosity of over 99 % featured a suitable environment for human bone marrow mesenchymal stem cells and their differentiation towards chondrocytes. However, chondrogenic differentiation was especially promoted by using chondroitin sulfate [55]. Lee et al. generated salt leached and solvent cast 3D scaffolds made of poly(L-lactide)-g-chondroitin for cartilage tissue engineering. An average pore size between 50 and 250 µm with an overall porosity of over 85 % was reported [56].

Conclusions
In this study, hierarchical extrusion-based bioprinting of ADA-GEL was demonstrated for applications in cartilage tissue engineering. 3D-bioprinting was used to manufacture hierarchical scaffolds in a biomimetic approach to recapitulate the heterogeneous structure of the native cartilage tissue. Hydrogel precursor modification via thermal pre-treatment of gelatine enabled modifiable mechanical and rheological characteristics of the ADA-GEL material. Therefore, an advance in ADA-GEL hydrogel applications could be presented due to a significantly improved printability in comparison to previous studies [11]. Gelatine solutions were exposed to 37, 70, 80 and 95 °C for periods of 3 h and 6 h before ADA-GEL solution preparation and compared to prior investigated ADA-GEL compositions [57]. Substantiated by rheological characterization via time-, amplitude-, frequency-, recovery-, and temperature sweeps, ADA-GEL-T 80°C_3h was found to feature the most favourable printability characteristics, enabling hierarchical layer deposition with high shape stability. The new ADA-GEL formulation allowed the design of hierarchical scaffolds mimicking the intrinsic hierarchical structure of natural cartilage tissue. Scaffolds with heights of over 1 cm could be printed. µCT analysis confirmed the fabrication of open porous, hierarchically structured scaffolds. Therefore, in future experiments, these scaffolds combined with chondrocytes could offer a rapidly accelerated healing process for damaged or diseased cartilage tissue thanks to directed organized cell orientation and new tissue formation. Nanoindentation proved that the hydrogel stiffness could be tuned by varying the gelatine pre-treatment temperature as well as by varying the concentration and type of crosslinking divalent ions. In summary, the results have implications for advancing cartilage TE investigations. The scaffolds fabricated in this study could be used as complex-structured, 3D-printed hierarchical templates for in-vitro cell seeding in matrix-associated cartilage implant strategies in the future.