Assessing non-synthetic crosslinkers in biomaterial inks based on polymers of marine origin to increase the shape fidelity in 3D extrusion printing

In the past decade, there has been significant progress in 3D printing research for tissue engineering (TE) using biomaterial inks made from natural and synthetic compounds. These constructs can aid in the regeneration process after tissue loss or injury, but achieving high shape fidelity is a challenge as it affects the construct’s physical and biological performance with cells. In parallel with the growth of 3D bioprinting approaches, some marine-origin polymers have been studied due to their biocompatibility, biodegradability, low immunogenicity, and similarities to human extracellular matrix components, making them an excellent alternative to land mammal-origin polymers with reduced disease transmission risk and ethical concerns. In this research, collagen from shark skin, chitosan from squid pens, and fucoidan from brown algae were effectively blended for the manufacturing of an adequate biomaterial ink to achieve a printable, reproducible material with a high shape fidelity and reticulated using four different approaches (phosphate-buffered saline, cell culture medium, 6% CaCl2, and 5 mM Genipin). Materials characterization was composed by filament collapse, fusion behavior, swelling behavior, and rheological and compressive tests, which demonstrated favorable shape fidelity resulting in a stable structure without deformations, and interesting shear recovery properties around the 80% mark. Additionally, live/dead assays were conducted in order to assess the cell viability of an immortalized human mesenchymal stem cell line, seeded directly on the 3D printed constructs, which showed over 90% viable cells. Overall, the Roswell Park Memorial Institute cell culture medium promoted the adequate crosslinking of this biopolymer blend to serve the TE approach, taking advantage of its capacity to hamper pH decrease coming from the acidic biomaterial ink. While the crosslinking occurs, the pH can be easily monitored by the presence of the indicator phenol red in the cell culture medium, which reduces costs and time.


Introduction
In the last decades, 3D bioprinting has emerged as a promising technology in the field of tissue engineering (TE) to produce high-fidelity constructs by combining biocompatible printing materials and living cells [1,2]. Compared with the traditional techniques for fabricating porous scaffolds, such as freeze-drying or electrospinning, 3D bioprinting approaches can offer reproducible structures with more precise pore size and geometry and, importantly, have the advantage that it can replicate the complex architecture and organization of native tissues, providing more personalized treatment for each clinical case [3][4][5]. However, a big challenge in the field of TE stems from involving materials from natural sources due to their variability, depending on the species, gender, age, environment, and type of polymer extraction [6,7]. For decades, the demands of TE have been filled by bioactive compounds from mammalian sources, such as collagen or hyaluronic acid mainly obtained from bovine or porcine tissues provided by the processing meat industry [8,9]. However, religious constraints and the risk associated with transmissible diseases such as bovine spongiform encephalopathy, transmissible spongiform encephalopathy, and foot-and-mouth disease are the causes of the current concerns, motivating research towards the use of marine resources [10,11]. In fact, a variety of marine-derived materials such as collagen, chitosan, fucoidan, hyaluronic acid, chondroitin sulfate, alginate or even enzymes have gained considerable interest and are being regarded as alternatives to the counterparts obtained from land mammal sources due to their similarities with proteins and polysaccharides present in the human extracellular matrix (ECM). Furthermore, they provide excellent and unique biochemical and physicochemical properties, i.e. good biocompatibility, low immunogenicity, and biodegradability [12,13]. Benefiting from these properties, the materials can be used as building block materials for TE constructs, including their use as biomaterial inks or bioinks in 3D bioprinting applications [14]. Additionally, it is important to refer that their origin should be managed according to biomass utilization strategies and industrial waste management, such as fishing by-products, under the circular economy concept [12].
Regarding 3D bioprinting, designing new innovative biomaterial inks with satisfying printability properties is not an easy task, since they depend on the rheological characteristics, which ideally should display shear thinning behavior and need to support the normal functionalities of cells, including biosynthetic activity, proliferation, and differentiation, promoting consequently the regeneration of damaged tissues [15,16]. Thus, during the printing process, and especially with the increase of layers, some physical deformations can occur, and the shape fidelity decreases. Therefore, performing a careful quantitative evaluation of the deposited filament is crucial. More specifically, shape fidelity is determined from visual macro-and microscopic images, filament fusion behavior, and collapse assessment of the suspended filament [1,17]. Unfortunately, due to their soft properties, typical for materials from natural origin, it is sometimes necessary to explore additional stabilization; for instance, by external approaches to increase the structure's stiffness and guarantee the printed materials' shape fidelity. In fact, a broad range of potential crosslinkers has been described in the literature, from both natural and synthetic sources, such as calcium chloride (CaCl 2 ) solution [18], genipin [19], ethyl (dimethylamino propyl) carbodiimide/ N-hydroxysuccinimide (EDC/NHS) [20], or even 1,6-hexane diisocyanate (HDI) [21], among others. However, there are some concerns regarding potential cytotoxicity when chemical crosslinking agents are used [22,23].
The motivation to investigate several nonsynthetic crosslinking solutions (i.e. phosphatebuffered saline, cell culture medium, CaCl 2 , and genipin), capable of increasing the material stiffness after the printing process without compromising the original shape fidelity, emerged to avoid the use of synthetic crosslinking agents for the stabilization of printed structures. The innovation of this work is essentially related to the combination of collagen (from shark skin), chitosan (from squid pens), and fucoidan (from brown algae) to use as biomaterial ink for a 3D bioprinting approach, assessing their printability, reproducibility, and shape fidelity. For this, several quantitative methods for material characterization, such as shape fidelity assessment, filament collapse, filament fusion, rheological performance, recovery capacity, and compressive behavior, were performed on the proposed marine biomaterial ink and the bioprinted marine scaffolds under the predefined printing process parameters (i.e. nozzle size, printing speed, and air pressure). Afterwards, immortalized human mesenchymal stem cells (MSCs) expressing human telomerase (hTERT) were cultured on the printed marine scaffolds to investigate if the materials provide an adequate micro-environment for cell viability and proliferation and also to assess the effect of each non-synthetic crosslinking strategy on cells as a comparative study, aiming for a future application in TE as a promising regenerative material envisaging biomedical precision therapy.

Materials
Collagen from blue shark (Prionace glauca) skin (sCOL) was provided by the Instituto de Investigacións Mariñas (CSIC), Vigo, Spain. The collagen was extracted according to the extraction protocol previously described by Diogo et al [21]. Fucoidan from brown algae (Fucus vesiculosus) (aFUC) was obtained from Marinova (Australia, product: Maritech® Fucoidan, FVF2011527) and used as received. Chitosan from squid pens (Dosidicus gigas) (sCHT) was extracted and purified according to the patent WO/2019/064231 [24]. In brief, the squid pens were milled using an Ultra centrifugal mill (Model ZM 200, Retsch) to obtain a powder, which was deproteinized and further deacetylated with a single alkaline process (50% (v/v) NaOH) connected to a constant airflow of nitrogen (N 2 ) at a temperature of 75 • C for 2 h.

Marine biomaterial ink preparation and scaffold fabrication
Initially, collagen and chitosan were separately solubilized in 0.5 M acetic acid [AcOH] (around pH 2.5), while fucoidan was dissolved in ultra-pure water. The polymeric solutions were prepared according to the best polymeric concentration obtained in previous work [7], i.e. 5% collagen (50 mg ml −1 ), and 10% fucoidan (100 mg ml −1 ). However, in the present work, the chitosan concentration was increased from 3% to 4% (40 mg ml −1 ) to improve the viscosity of the biomaterial ink. Then, to obtain an appropriate homogenous ink, the blending of the marine polymeric solutions was carried out using an overhead blender (Ultra-turrax T18, IKA Staufen, Germany) in low rotations to avoid bubbles at a temperature of 4 • C [25], then left to rest on ice for 30 min. After that, the biomaterial ink was transferred into a 10 ml plotting cartridge (Nordson EFD, Oberhaching, Germany) using a spatula. The biomaterial ink was centrifuged at 300 rpm for 5 min to remove residual air bubbles and was ready for printing.

Printed constructs-shape fidelity assessment
For the determination of shape fidelity, all 3D marine scaffolds (n = 6 per condition) were imaged immediately after being printed (without crosslinking) and after 24 h of crosslinking (submerged separately in four non-synthetic crosslinkers) using a Leica M205 C stereo microscope equipped with a DFC295 camera (Leica Microsystems, Wetzlar, Germany) at a magnification of 0.78× (resolution 2048 × 1536 pixels).
The structure length and width, distances of diagonal, distance between strands, and strand diameter were measured from the acquired images using ImageJ (National Institutes of Health, Bethesda, MD, USA) [26].

Biomaterial ink rheological assessment and recovery properties
The rheological tests were conducted using a Rheotest 4.1 rheometer (Medingen, Germany), plate-plate geometry (d = 36.6 mm) with 0.1 mm gap at room temperature (≈25 • C). This method was used to determine if the developed marine ink exhibited a shear-thinning behavior. Thus, rotational shear sweep tests were carried out by measuring the intrinsic viscosity as a function of the shear rate, from 0.1 to 100 s −1 , increasing by 0.1 s −1 . In addition, to study the extrusion process of the polymeric ink, viscosity recovery was tested by applying three sequential shear rates of 1 s −1 for 200 s, then 100 s −1 for 50 s, and finally, 1 s −1 for 200 s again, called a cycle. The pre-defined process was repeated twice for this analysis, and the following formulation was applied (equation (1)) to calculate the recovery percentage. The average viscosity after 1-10 s was divided by the 1-10 s before applying high shear. All data analysis was obtained using an average of at least three experiments % Recovery = Average at (1 − 10 s) after high shear Average at (1 − 10 s) before high shear × 100. (1)

Filament collapse test
The filament collapse test is based on the work by Ribeiro et al, where suspended ink filaments' deflection is assessed [1]. For this test, pillars  Theoretically, it was assumed that 20 s after printing, the forces acting on the filament had reached equilibrium, i.e. no more deformations were experienced. For this, considering that the filament remains constant, an equilibrium equation (equation (2)) can be written as: where F g represents the gravitational force and F σ is the material resistance force resulting to yield. Then, if we consider the equilibrium in the vertical direction, (equation (2)) can be rewritten, being now expressed as shown in (equation (3)): where θ is the angle of filament deflection with the printed horizontal direction. Subsequently, it is necessary to consider the forces exerted on an infinitesimal volume element δV. For this, (equation (3)) now takes the form of (equation (4)): Moreover, the angle of deflection θ should be related to the stresses acting on the volume δV, as demonstrated in (equation (5)): where m is the filament mass, L is the distance between the pillar's edge and the suspended filament's midpoint, and δA an infinitesimal cross-sectional area of the filament. Taken all this into consideration, (equation (5)) can be rewritten as (equation (6)): where ρ is the density of the polymeric ink material prepared, which is approximately 1126 kg m −3 (calculated from the well-established formula: density = mass/volume), g is the gravitational acceleration (around 9.8 m s −2 ), and σ yield represents the yield stress of the biomaterial ink, which is approximately −1.211 (calculated from the slope of the rheological assessment).

Filament fusion test
The filament fusion test consisted in printing a zigzag pattern of parallel strands with increasing spacing, which started at 0.40 mm, and then increased by 0.20 mm per row, ending at 3 mm spacing (g-code in supplementary info) with a constant plotting speed at 20 mm s −1 , and 60 kPa of air pressure. After that, a second and third layers were printed to evaluate the filament fusion by increasing the layer height. Each layer was imaged immediately after being printed using a Leica M205 C stereo microscope equipped with a DFC295 camera (Leica Microsystems) at a magnification of 0.78× (resolution 2048 × 1536 pixels). The distance between filament lines was measured from the acquired images using ImageJ (National Institutes of Health, Bethesda, MD, USA) [26]. The plotted values represent the mean of measurements over three repetitions of the test for each formulation.

Swelling test-crosslinker medium uptake quantification
The swelling properties, or water uptake abilities, of the printed marine scaffolds were studied. For this, cuboidal scaffolds were printed using a 410 µm dosing needle, with a distance of diagonal of 8 mm, a strand distance of 2.8 mm (5 strands per layer), and a layer height of 2.7 mm (10 layers). After printing, each scaffold was immediately weighted to obtain the initial weight (W 0 ) and then incubated separately in four types of non-synthetic crosslinking solutions: (1) PBS solution; (2) RPMI 1640 medium; (3) 6% CaCl 2 solution, and (4) 5 mM Genipin at 37 • C for 21 d. Then, at different previously defined time points (1, 2, 3, 6 h and 1, 3, 7, 14, and 21 d), the samples were withdrawn, the excess solution soaked up with dried filter paper, and weighed immediately (W 1 ), using an analytical balance (Denver Instrument, Germany).
In addition, images of each scaffold were acquired by the stereo microscope at each time point to visualize structural differences along the swelling experience. All the assays were performed in quintuplicate (n = 5). Lastly, the percentage of each solution absorbed by the samples was calculated using the following (equation (7)):

Characterization of mechanical properties by compressive testing
The printed scaffolds were subject to uniaxial mechanical compressive testing using Zwick-Roell Z010 equipped with a 100 N load cell (Zwick-Roell, Ulm, Germany) to evaluate the jellification ability of the different crosslinker solutions (PBS, RPMI, 6% CaCl 2 , and 5 mM Genipin) after 24 h of incubation. For this analysis, the samples were printed according to the dimensions previously described (in section 2.2), which is approximately 11.3 mm (length) × 11.3 mm (width) × 2.7 mm (thickness). The compressive tests were conducted at room temperature using a load cell of 100 N and a pre-load of 10 mN with a crosshead speed of 1 mm min −1 until a maximum deformation of 50%. The compressive modulus (E) was calculated from the standard force versus deformation (%) as the slope of the linear region at 0%-10% deformation. Six specimens per condition were tested; the results are expressed in mean ± SD.

Determination of the acid neutralization capacity (ANC) of 3D-printed scaffolds
The printed marine scaffolds were submerged in the different non-synthetic crosslinker solutions for 24 h at room temperature to evaluate the time required for each solution to neutralize the acidic pH of the marine ink, resulting from the polymer solubilization process. The pH of each solution was measured before and after the scaffolds were submerged at different time points (0 min, 30 min, 1 h 30 min, 2 h, 2 h 30 min, 3 h, and after 24 h). Additionally, each solution was changed after 1 h and 2 h, being as well measured immediately after the submersion. All the assays were performed in triplicate (n = 3) using a pH meter (WTW InoLab pH/Ion level 2, Germany).

Cell pre-expansion
Immortalized human MSCs expressing hTERT [27] were previously expanded in monolayer culture at 37 • C and 5% CO 2 in Dulbecco's modified Eagle's medium (Gibco, Germany), supplemented with 10% fetal calf serum (Corning, USA), and 1% antibiotic and antimycotic solution (i.e. 100 U ml −1 penicillin and 100 µg ml −1 streptomycin; Gibco, UK), 7.4 pH; the medium was changed twice a week. Then, cellladen 3D printed scaffolds were incubated for 7 d with the time points of 1, 3, and 7 d in a 48-well plate with a distribution cell of 0.25 million cells/scaffold to perform the live/dead assessment. For this, 40 µl of concentrated cells in cell medium were directly transferred onto the top of each scaffold, incubated at 37 • C for 3 h, for cell attachment, and then increasing amounts of media were added till cover the scaffolds.

Cell viability assessment by live/dead staining
The cell viability of the seeded immortalized human MSCs expressing hTERT was evaluated by live/dead assay stained with Calcein-AM/ethidium homodimer-1 (Invitrogen™ LIVE/DEAD™ Viability/Cytotoxicity Kit for mammalian cells, Thermo Fisher Scientific, Waltham, MA, USA) at 37 • C for 20 min, according to manufacturer's protocol, i.e. 0.6 µl of calcein and 1.2 µl of ethidium per ml staining solution. In brief, the cytoplasm of living cells was stained with Calcein-AM by marking the esterase activity, giving the stained cells green fluorescence color. At the same time, cells without intact membranes were labeled with ethidium homodimer-1, giving the nuclei a red fluorescence color. At each time point of the experiment, z-stack images were acquired using fluorescence microscopy (Keyence BZ-X800 fluorescent light microscope, Germany) at 4× magnification (1.887 µm/px) with an image resolution of 1920 × 1440 pixels. Excitation/emission wavelengths for Calcein-AM and ethidium homodimer-1 were 495/515 and 528/617 nm, respectively. The number of live and dead cells were counted using the ImageJ analyze particles plugin (Fiji, Version 1.52p), and applied to all image stacks (n ⩾ 3 per condition). The cell viability was then determined by calculating the ratio of the living cells divided by the sum of live and dead cells.

Statistics
Statistical analysis was performed by two-way ANOVA followed by Tukey's post hoc test, using GraphPad Prism 8.0.1 (GraphPad Software, Inc., La Jolla, Ca). Differences between the groups with a confidence level of 95% (p < 0.05) were considered statistically significant. The significance level between the groups was represented by symbols of * (p < 0.05), * * (p < 0.01), * * * (p < 0.001), * * * * (p < 0.0001), and by ns (no significance). All data were presented as mean ± SD. As additional information, the equations present in this work were designed using the MathType 6.9 software (Design Science).

Scaffolds shape fidelity assessment
In a first assessment, several 3D printing pre-tests were performed to find the most promising biomaterial ink based on the best concentration of blended three marine compounds (i.e. collagen-chitosanfucoidan, adapted from previous study [7]), and the proper processing parameters (i.e. nozzle size, printing speed, and air pressure) in order to obtain controlled printability and shape fidelity. For this, qualitative visual inspections were recorded using stereo microscopy to determine if the developed marine ink was adequate for extrusion printing by investigating printability and shape fidelity. The structural size, filament diameter, and strand (filament) distance were evaluated and compared with the original parameters of the CAD model, as demonstrated in figure 2. After a simple qualitative visualization of the printed material, it was clear that the proposed marine polymeric combination could provide a satisfactory blend capable of being used as biomaterial inks for 3D extrusion printing. In figures 2(a) and (b), it is possible to see a well-structured, reproducible, and stable printed material able to support its structure in height without compromising the strand (filament) distance (sfd) significantly (i.e. the structure porosity that is defined as the distance between parallel filaments), generally associated to the collapse when increasing the number of layers. To prove this fact, several measurements of length, width, dd, sfd, and filament thickness (ft) were performed on constructs after being printed (without crosslinking) and after being submerged in different crosslinking solutions (PBS, RPMI, 6% CaCl 2 , and 5 mM genipin) for 24 h, as demonstrated in figures 2(b) and (c). It was proved that adding crosslinking agents does not significantly change the original shape of the construct. On average, as additional measurement, the length and width of all printed formulations after crosslinking point to 11.37 ± 0.05 and 11.35 ± 0.04 mm, respectively. To validate the obtained dd value per condition, each length and width were divided by 2 to serve as a major and minor leg and then applied to a simple calculation of the hypotenuse, obtaining an approximate value of 8.025, which validates the shape fidelity by being close to the original settings and the dd obtained in all conditions. Regarding the strand (filament) distance (defining the microporosity) and the ft, there was a clear correlation between these two parameters, which with the increase of the ft, a decrease in the pore size was observed, which is seen as a perfectly natural behavior. In conclusion, the developed ink showed a good shape fidelity, including well-defined macroporous structures up to a height of 10 layers.

Rheological assessment and shear recovery properties
The shear viscosity and recovery behaviors were investigated on a rheometer at room temperature (∼25 • C) to investigate whether the developed marine ink had appropriate properties for biomedical applications as printing material [28,29]. The data obtained from the rheological analysis are depicted in figure 3.
According to the results shown in figure 3(a), the marine ink revealed a typical trend of shear-thinning behavior over a shear rate from 0.1 to 100 s −1 , which resembles a non-Newtonian fluid. This behavior is characterized by a continuous decrease of the viscosity (η) with the increase of the shear rate (y) [30], through which it is possible to calculate the material's yield stress (σ yield ) using the linear regression from decreasing slope of viscosity vs. shear rate, obtaining an approximate value of −1.211 (R 2 = 0,89) for the proposed ink. This definition refers to the minimum stress at which a material will undergo permanent deformation without a significant increase in the load or external force [31].
Regarding the shear-thinning behavior, when a polymeric material presents this characteristic, it typically is related to its capacity to be used as injectable material since it is fluid enough to pass, for example, through syringes/needles. However, this property is not the unique characteristic that fluid material should comprise. After any material is subjected to higher shear forces to pass the needle canal, the material should recover, in large part, to its initial viscosity [32]. Therefore, to evaluate the recovery capacity of the developed marine ink paste, it was submitted to low (LS) and high shear rate (HS) in two distinct and sequential cycles, i.e. 1 s −1 for 200 s (LS), then 100 s −1 for 50 s (HS), and finally, 1 s −1 for 200 s again (LS), repeated twice, as demonstrated in figure 3(b). After an analytical verification, when a higher shear force was applied for the first time, the ink material had a recovery capacity of around 65% (b/a), noticing an increase of this property after applying LS again along the time, to 84% (c/b).  Furthermore, after analyzing the second shear cycle, it was noticed that after high shear has been applied, the material lost part of its initial viscosity (before second HS = c) but continued to have an interesting recovery behavior, which was about 79% (d/c). Finally, the recovery property after the second HS (d) was compared to initial viscosity (before first HS = a), and it was observed that the recovery values after the second-high shear were very similar to the values obtained after the first high shear (d/a ≈ b/a).
These similar values demonstrate an attractive recovery capacity of the developed biomaterial ink, relevant when assessing the stability and printability of the ink during several printing steps, since it could support sequential high forces without significantly losing its viscosity properties.

Filament collapse test
The filament collapse analysis is a simple quantitative test in which a single filament is printed horizontally at pre-defined air pressure and speed over a pre-designed platform with several pillars with increasing distances between them, a method based on the work by Therriaultet et al [33]. The printed ink filament deflection was calculated through measurements of the deflection angle θ between the edge of each pillar and the middle point between pillars of the suspended filament immediately after deposition (t = 0 s) and after 20 s (t = 20 s). Ribeiro et al [1] state that 20 s after printing the filament is stable enough to assume that deformation is negligible from this time and the gravitational forces that act on the filament reached an equilibrium. The theoretical explanation and the filament collapse test results are demonstrated in figure 4.
A qualitative observation of the collapse test ( figure 4(b)) reveals an increase in the filament sagging with the increase of the gap length between the pillars, which reflects on the increase in the deflection angle θ. This increasing deflection angle is highly visible on the 16 mm gap, followed by the 8 mm gap, especially 20 s after printing. On the other hand, especially in gap distances of 1 and 2 mm, the filament did not collapse and formed a straight bridge between the pillars, which is reasonable to classify the collapse area factor as 0% since the initial point (pillar) is the same as the one obtained in L. In quantitative analysis, shown in figure 4(c), it is perceptible that the deflection angle increased with the increase of the gap distance at both time points. To further relate this behavior with the yield stress of the prepared marine ink, the deflection angle for each gap distance was calculated using the theoretical model exhibited in section 2 by equation (6). In general, for both times, the slope of the regression lines on theoretical values followed the same trend and similar values with the data experimentally obtained, validating that the proposed ink is a suitable blend for 3D printing.

Filament fusion test
The filament fusion test is another straightforward quantitative analysis in which a single filament in a straight-line pattern of parallel strands with increasing spacing is printed at pre-defined air pressure and speed, to assess the minimum strand distance required to avoid fused strands as well as the influence of adding more layers on fusion behavior, as demonstrated in figure 5. For this, the fused filament length (fs) was initially measured at each filament distance (fd), as exhibited by figure 5(a). Then the fs was normalized and divided by the average of ft ( figure 5(b)) until obtaining three independent layers (figure 5(c)) in order to correlate the fusion behavior with the increasing height. The measurements were performed until the third layer to avoid wrong interpretations promoted by the ink spreading caused by the surface tension of the base material if only a single layer was analyzed [1].
As mentioned above, the fusion test was first conducted by printing a single layer on a base material (well plate) to evaluate the fusion behavior with the increase of the fd and to understand the base surface's influence on the filament. After that, the data was compared with the 2-and 3-layers experiments. The first qualitative observation revealed a clear relationship between the strand closure, fd, and a well-consistent marine ink paste. As expected, the material fuses together if the strand distance is too close (in the present case, less than 1 mm). In fact, our results suggest that fusion happened at lower fd, comprised between 0.40 and 1.60 mm, and with the decreasing of fd, an increase of fs was observed. Additionally, it is important to highlight that this behavior is common when the ink is formed by soft materials, mainly based on biological compounds; during the printing procedure, the ft can increase significantly compared to the nozzle diameter. This difference, mostly associated with the spread of the material, can induce more fusion, especially in lower fd. Another reason is related to the fact that most biological materials possess weaker mechanical properties compared to (semi)synthetic materials and the missing of fast crosslinking after printing. In fact, after printing, many of them may require external crosslinking agents for the stabilization of the printed construct.
Regarding the comparative study between layers, it is noticed in figures 5(b) and (c) that the addition of layers also influenced fusion behavior since the fd, before fusion occurred, increased per each layer added, in most parts promoted by the spreading of material. This behavior could be observed macroscopically when the three individual layers are compared qualitatively, allowing us to assess the minimum pore size achievable for each layer, i.e. 1.20 mm for 1 layer, 1.40 mm for 2 layers, and 1.60 mm for 3 layers. Furthermore, despite the increase of ft compared with the nozzle diameter, with the increase in height, the printed filament maintained its structure without collapsing. If there were some collapse, an increase in the thickness of the filament would be noticed in layer 3 compared to the 1st and 2nd layers, which was not the case. In general, the results suggest that the ink, based on marine compounds, had exciting properties to be used in bioprinting since it provided well-defined patterns, was reproducible, and supported their structure in height without deformation (in height), and even on the first printing layer did not present significant spreading of material when being in direct contact with the base.

Swelling test-crosslinker medium uptake quantification
The printed marine scaffolds were submerged in each non-synthetic crosslinker solution (PBS, RPMI, 6% CaCl 2 , and 5 mM genipin) for 21 d to estimate the swelling ability (hydration property) and the structural fidelity of the printed material along the experimental time. All data collected, including photographs from several time points, are demonstrated in figure 6.
The swelling data results shown in figure 6(a) demonstrate that the printed material submerged in all crosslinker solutions had the ability to absorb large amounts of water (doubling its initial weight), especially during the first hour (time point 1), followed by a stabilization after the second time point (2 h). After this point, almost all conditions tended to maintain their weight until the experiment's end, reaching a saturation point. Moreover, it is visualized in figure 6(b) that during the incubation time, the scaffold geometry and morphology of the macro-pores did not change significantly, which indicates that the printed materials had the potential to maintain their shape fidelity after being submerged in aqueous solutions, making them suitable materials for long-term experiments. Additionally, it is noticed that the samples immersed in CaCl 2 and PBS solution promote a higher absorption ability, comprising percentage values of (122 ± 5) and (109 ± 19)%, respectively. In contrast, the materials in the genipin solution demonstrated less absorption capacity (93 ± 12)%, while those submerged in RPMI comprised (103 ± 17)%. In general, all the conditions have the ability to increase their swelling by 100% of their initial mass. In the case of genipin, it is noted that the materials start losing their weight after 24 h, being associated with a slight compaction observed in this condition without compromising their original structural form. This behavior may be related to more crosslinks promoted by the lengthy exposition to genipin molecules, in which the compaction does not allow space for the presence of water molecules in the structure, and also by the presence of ethanol in the genipin solution. It is also essential to describe that the samples crosslinked with RPMI and genipin solutions started to change their original color (brown) to a darker shade. In the case of RPMI, this may be due to the presence of the pH indicator phenol red, which causes a red discoloration. In contrast, the genipin-treated samples change to a dark blue color induced by the reaction of the amine groups present in the marine polymers in the presence of oxygen [34]. Thus, depending on the intended application, the samples may require an additional processing step to remove most of the genipin reagent or even for their decolorization, despite the associated extra costs. However, in the case of implantable materials, the unwanted color is inconsequential. Furthermore, a simple visualization reveals that natural structural degradation is not noticed during the experimental period, indicating remarkable molecular stability. To demonstrate this stability, our previous work utilized the same marine biopolymers to develop compact hydrogels without chemical crosslinking agents [35]. In this study, we subjected the developed marine scaffolds to an enzymatic cocktail consisting of collagenase, hyaluronidase, and lysozyme, at concentrations identical to those found in human blood plasma, for 21 d. Although some degradation was observed over time, it was not significant enough to compromise the polymeric systems' structural integrity or even their handling and manipulation. This indicates that our materials based on these marine biopolymers have promising behavior for a TE approach.

Characterization of mechanical properties by compressive testing
The uniaxial compressive tests were conducted to evaluate the compressive modulus and the yield stress until 50% deformation for each printed sample immersed in each one of the four non-synthetic crosslinker solutions (PBS, RPMI, 6% CaCl 2 , and 5 mM genipin) for 24 h to allow for the maximum swelling of the samples, according to the aforementioned water uptake quantification. The stress-strain curves were used to determine the Young's modulus, as described in figure 7. Additionally, after each compressive data acquisition, the scaffolds' height was measured immediately after compression and 24 and 48 h after deformation to evaluate their recovery capacity.
The compressive mechanical behavior and recovery properties are both essential factors to consider in manufacturing of scaffolds, especially when the final application is the regeneration of human tissues that can be subject to significant stress and strain, such as articular cartilage. In fact, cartilage, bone, muscles, and tendons are four tissues that are constantly subject to stretching and contracting movements exerted by body locomotion [36]. Mechanically, each tissue presents a different compressive modulus, which, for example, increases from cartilage (around 0.7-0.8 MPa) to bone tissue (from 170 to 193 MPa) [37]. Furthermore, it is well-known that glycosaminoglycans such as hyaluronic acid and chondroitin sulfates are responsible for supporting high compressive loads, whereas collagen fibrils have a considerable contribution in providing structure, high recruitment of water contents (maintaining hydration), high tensile strength, and the ability to tolerate variations in terms of shear stresses [38].
To evaluate the stiffness as a mechanical property of the developed 3D printed scaffolds, cross-linked in several non-synthetic crosslinker solutions, uniaxial compressive forces were exerted on these materials, and the respective Young's moduli were determined from the initial linear relationship between the tensile stress and the strain [39], as demonstrated in figure 7(a). The results exhibited significant differences between the crosslinkers, highlighting for scaffolds immersed in genipin solution, which shows the highest stiffness (8.4 ± 0.3 kPa), followed by RPMI solution (3.4 ± 0.3 kPa), while CaCl 2 and PBS as crosslinkers follow the same trend of lower stiffness (without statistical significance), with values of 2.1 ± 0.5 kPa and 1.6 ± 0.2 kPa, respectively. In fact, as expected, significant differences between the scaffolds crosslinked in genipin compared to the other crosslinkers were appreciated since it has been reported as an excellent natural crosslinker for proteins (as collagen), alginate, and chitosan due to its capacity to promote covalent interactions between the polymers [40][41][42]. In contrast, we hypothesize that CaCl 2 can have an influence with fucoidan through electrostatic interactions between the respective divalent calcium cations and sulfate groups, potentially acting as a weaker crosslinker in an interplay with the direct interaction between the biopolymers chains themselves. On the other hand, we believe that the PBS buffer does not significantly interact with these marine polymeric materials, just providing the medium for the biopolymers chains to reorganize upon printing. Unexpectedly the RPMI medium had a positive effect on the combination of these marine polymers. In fact, RPMI composition is characterized by high levels of inorganic salts and various amino acids, which exhibit negative, positive, or neutral charges. Collagen, for instance, consists also of multiple types of amino acids, which could increase the interactions between the positively charged groups (protonated amines) of collagens and/or chitosan and the negatively charged groups (e.g. ester sulfates and carboxylates) of fucoidan via ionic electrostatic interactions, promoting the formation of more connecting bridges within this internal network.
Unfortunately, the scaffolds still presented lower stiffness than native tissues like cartilage, which may be due to the fact that they are based on biopolymers network organization and the use of natural crosslinkers that, in general, are weaker than the bonds promoted by synthetic crosslinking agents since they can create a higher interconnected organization level which results in higher stability of the polymeric structure [14]. In contrast, there are some concerns regarding cytotoxicity when synthetic crosslinking agents are used, which may interfere with the materials' therapeutic capacity [22,23,43]. Additionally, it is also essential to refer to the importance of the nozzle size in bioprinting since it can reflect directly on the mechanical properties. In fact, after testing several nozzle sizes, Chung et al [44] concluded that increasing the nozzle size decreases the printed scaffolds' compressive modulus. However, it is necessary to find a good relationship between the material used and the nozzle size, since blockages can occur during the printing process if the diameter is too small.
Despite the mechanical differences between the developed scaffolds and the native tissues, they have demonstrated interesting physical properties as a cohesive material, allowing handling with spatulas, and having an impressive recovery ability after compression, as shown in figure 7(b). The results reveal that immediately after the compressive force was applied, all samples could recover around 82% of their initial height, and after 24 and 48 h, this value increased to 90%. It is essential to highlight that after 48 h, the conditions that demonstrated a more extraordinary recovery ability were those crosslinked in genipin solution, followed by the RPMI medium, which is probably due to the nature of the crosslinking promoted in each case, with covalent bonds created with genipin supporting a more stable scaffold structure, withstanding compression, while physical bonds are more prone to changes when submitted to deformation.

Determination of the ANC on 3D printed scaffolds
The determination of the ANC is a simple quantitative measurement of the buffered pH of the liquid surrounding the printed marine scaffolds. This analysis tries to understand if the non-synthetic crosslinker solutions (PBS, RPMI, 6% CaCl 2 , and 5 mM genipin) have the ability to neutralize the acidic pH environment of the scaffolds caused by the polymeric solubilization process used to prepare the ink, and the minimum time required for each solution to reach a stable pH value, as demonstrated in figure 8. In fact, pH variations can be considered problematic in cell culture environments since they can directly affect cell behavior like metabolism, cell growth, and membrane potential or even lead to cell death. However, a static pH is not strictly necessary since it can vary according to each cell type's preferences. For example, most mammalian cells show better growth at pH 7.4, while some transformed cell lines have been demonstrated to grow better at slightly more acidic environments (around pH 7.0-7.4). In contrast, some fibroblast cell lines prefer a more basic environment that comprises a pH between 7.4-7.7 [27].
Regarding our results, it is noticed that the initial pH varies according to each crosslinker solution. For example, the PBS and RPMI solutions present a pH of around 7.4, while the CaCl 2 and genipin solutions have pH values of 6.8 and 7.9, respectively. However, after immersing the scaffolds in the crosslinking solutions, a significant decrease in pH in all conditions was visible due to the rapid release of acid into the surrounding environment. To prove this fact, the scaffolds submerged in RPMI medium solution were recorded by photographs since the presence of the pH-indicator phenol red allows the easy detection of pH variations, which revealed a change from red color (neutral pH) to yellow (acidic pH) after 1 h submersion, as demonstrated in figure 8(a). After 1 h, the solutions were replaced with new ones to discard the acidic components. After another hour of incubation, the liquid was again replaced, since more acid was released, mainly in the scaffolds crosslinked with CaCl 2 and genipin, respectively. In contrast, the RPMI medium and the PBS have the ability to neutralize the acid present in the scaffolds faster, i.e. 2 h and two solution replacements are enough to maintain the pH stable at 7.4; this behavior could be explained by the PBS, and RPMI medium being buffering solutions while the CaCl 2 and genipin were dissolved only in water. Overall, the RPMI solution demonstrated the capacity to neutralize the acid material faster due to the pH buffers present. Furthermore, it had the advantage of containing a visible indicator, avoiding external measurements and waste of time.

Cell viability assessment by live/dead staining
A preliminary evaluation of the viability of immortalized human MSC expressing hTERT cultured for 7 d on the printed marine scaffolds was determined through a live/dead assay using fluorescence microscopy after staining with calcein-AM and ethidiumhomodimer. Briefly, the living cells metabolized the calcein-AM, expressing a green fluorescence color in the cytoplasm, while the dead cells were stained with ethidium-homodimer, exhibiting a red fluorescence color in the nuclei. The fluorescence images acquired of the samples crosslinked with different non-synthetic crosslinker solutions (PBS, RPMI, 6% CaCl 2 , and 5 mM genipin) are shown in figure 9. To achieve a comparative evaluation of the effect of the different crosslinkers on the seeded cells, the ratio of viable cells is respect to total number of cells was quantified using ImageJ software with representative fluorescence images; the data obtained is Plot of the viable cell ratio is respect to total cell numbers, expressed as mean ± standard deviation. The * * * * represents statistical significance at a significance level of p < 0.0001. demonstrated in figure 9(b). Furthermore, the hTERT-MSCs were selected for the present work since they can differentiate into osteogenic, chondrogenic, and adipogenic lineages and have the advantage of behaving similarly to primary MSC but can be expanded without limitation [45,46]. Live/dead assay revealed a high ratio of living cells in all samples. In fact, at the first time point (24 h after seeding), the cell viability comprised 96% on the scaffolds crosslinked in PBS (SD = 1.22%), RPMI (SD = 3.10%) and genipin (SD = 0.99%), while the scaffolds crosslinked with CaCl 2 solution supported a lower cell viability (89 ± 3%). This behavior could be explained by the data obtained on the determination of ANC (section 3.8), proving that the CaCl 2 solution has less ability to neutralize the pH present in the scaffolds, reaching a minimum pH of 6.7. Therefore, the lower pH obtained with these scaffolds compared to the other conditions may have affected cell adhesion, cell growth and may have also promoted cell death. However, with increasing of the experiment time, an increase in cell viability in samples crosslinked with CaCl 2 could be noticed, which can indicate that the pH buffer present in cell culture medium could neutralize the acidic character of the biomaterial ink over time, allowing a suitable environment for higher cell viability, approximately 93% on day 3 and 95% on day 7. In general, the cross-linking methods that demonstrated a better microenvironment for cell viability were the RPMI, followed by the genipin solution, comprising 97% of viable cells on days 3 and 7 (both with SD ≃ 2%). Furthermore, after 3 d of culturing, cells apparently started to stretch and some cell proliferation could be observed during the cultivation time. This indicates adaptability, good cell adhesion to the scaffolds, and offers adequate conditions for the cells. Overall, combining collagen, chitosan, and fucoidan to form the biomaterial ink could provide good structural support as printed materials and a suitable environment to sustain cell viability and proliferation, making it a promising printing material for TE applications.

Concluding remarks
In this study, marine collagen, chitosan, and fucoidan were blended to fabricate an adequate biomaterial ink, capable of being used for 3D extrusion printing, combining the strategies of valorization of marine biological resources and derived materials with TE and regenerative medicine applications. The marine ink demonstrated shear thinning behavior with a recovery property of around 79% after two high shear cycles. The outcome of a set of characterization tests validated that the developed marine ink combined with the pre-defined processing parameters (i.e. nozzle diameter, pressure, and speed) have the ability to print stable and reproducible structures with high shape fidelity without the occurrence of collapsing or deformation with increasing layer number. After printing, the structures were cross-linked using four types of non-synthetic crosslinker solutions, i.e. PBS, RPMI, 6% CaCl 2 or 5 mM genipin, selected to avoid the use of synthetic crosslinker agents, which are sometimes associated with cytotoxicity. Although the scaffolds crosslinked in genipin solution exhibited better compressive mechanical properties, in general the RPMI solution (cell culture medium) rendered scaffolds with the most valuable properties, also benefiting from the presence of phenol red allowing a pH visualization to indicate a complete neutralization, which is fundamental for cell viability. Additionally, hTERT-MSCs were cultured on printed marine scaffolds for 7 d, and the structures demonstrated a good environment for cell survival (cell viabilities of above 90% over 7 d). In future, the non-synthetic crosslinking strategy should be tuned to increase the mechanical properties of the printed scaffolds in order to better mimic the compressive modulus of the native tissues to be engineered and explore new methodologies to increase the pH of the polymeric blend before printing to enable the use of the biomaterial ink for bioprinting applications.

Data availability statement
The data cannot be made publicly available upon publication because they are not available in a format that is sufficiently accessible or reusable by other researchers. The data that support the findings of this study are available upon reasonable request from the authors.