Rheological and Biological Impact of Printable PCL‐Fibers as Reinforcing Fillers in Cell‐Laden Spider‐Silk Bio‐Inks

The development of bio‐inks capable of being 3D‐printed into cell‐containing bio‐fabricates with sufficient shape fidelity is highly demanding. Structural integrity and favorable mechanical properties can be achieved by applying high polymer concentrations in hydrogels. Unfortunately, this often comes at the expense of cell performance since cells may become entrapped in the dense matrix. This drawback can be addressed by incorporating fibers as reinforcing fillers that strengthen the overall bio‐ink structure and provide a second hierarchical micro‐structure to which cells can adhere and align, resulting in enhanced cell activity. In this work, the potential impact of collagen‐coated short polycaprolactone‐fibers on cells after being printed in a hydrogel is systematically studied. The matrix is composed of eADF4(C16), a recombinant spider silk protein that is cytocompatible but non‐adhesive for cells. Consequently, the impact of fibers could be exclusively examined, excluding secondary effects induced by the matrix. Applying this model system, a significant impact of such fillers on rheology and cell behavior is observed. Strikingly, it could be shown that fibers reduce cell viability upon printing but subsequently promote cell performance in the printed construct, emphasizing the need to distinguish between in‐print and post‐print impact of fillers in bio‐inks.


Introduction
In the context of tissue engineering and regenerative medicine, biofabrication can be defined as "the automated generation of DOI: 10.1002/smtd.202201717biologically functional products with structural organization from living cells, bioactive molecules, biomaterials, cell aggregates such as micro-tissues, or hybrid cell-material constructs, through bio-printing, or bio-assembly and subsequent tissue maturation processes." [1]io-inks are fundamental to highresolution bio-printing of dimensionally stable scaffolds with sufficient mechanical properties.At the same time, they must ensure cell survival during printing and enable tissue maturation in the long term, leading to the formation of a biofunctional construct.This combination has proven to be a bottleneck in the past, as increasing the polymer concentration in bio-inks to improve the printing result can lead to loss of cytocompatibility within the so-called biofabrication window.Shear stress during printing can also damage cells, and subsequent gelation can result in immobilization of cells in the dense polymer network.This prevents cells from spreading, migrating, and colonizing the hydrogel matrix with newly synthesized extracellular matrix, rendering the bio-manufactured product unusable. [2]Since the identification of this problem, numerous strategies have been explored to transform biofabrication from bare cell printing to the production of functional tissue constructs.The main focus was on the chemical functionalization of hydrogel-forming polymers to improve crosslinking properties, as well as adaptations and new developments in the field of printing processes.However, it is emerging that this multidimensional challenge can only be addressed effectively through multilateral approaches-there is talk of an indispensable convergence of biofabrication technologies that must bring together complementary strategies in a synergetic manner. [3] highly promising approach in this context is applying fiber reinforcement to mechanically support bio-inks while substituting the need for high polymer concentrations.The advantages of composite hydrogels as biomimetic constructs with tunable mechanical, physical, chemical, and biological properties have already been demonstrated for tissue engineering and regenerative medicine.These properties are attributed to their potential to revolutionize the field. [4]Fiber-reinforced hydrogels have been shown to display improved control over cell distribution, vitality, and activity as compared to purely hydrogel or purely fiberbased constructs. [5]However, most of the work relates to hydrogels with embedded nonwovens and fiber mats, which are not suitable for conventional extrusion bioprinting due to their continuous character. [6]Consequently, fibers are predominantly integrated either as an intermediate layer [7] or as a scaffold. [8]So far, it has only rarely been possible to establish fiber-reinforced hydrogels for 3D bioprinting. [9]For example, gold nanofibers were suspended in collagen-based bio-inks and aligned using an electric field during bioprinting.Furthermore, in the field of cartilage replacement, the printability of alginate bio-inks could be improved by integrating nanocellulose, [10] and polylactide (PLA) fibers were used to increase the elastic modulus of printed alginate hydrogels with improved chondrocyte viability. [11]Moreover, natural silk fibers (fibroin) were mechanically ground and used for various bio-inks to improve printability. [12]The mechanical properties of printable decellularized cartilage matrices could be improved by adding fiber pieces made of gelatin and poly(lactic-co-glycolic acid) (PLGA). [13]Noteworthy is the work of Omidinia-Anarkoli et al., [14] who used magnetically aligned fiber fragments as an anisotropic structural component to induce oriented growth and electrical activity of neuronal cells.This work underscores the hypothesis that fibers represent a highly relevant structural component for cellular behavior.
We have recently established a versatile complementary technology targeting the development and systematic research of fiber-reinforced bio-inks.The goal of this approach is to develop new composite bio-ink formulations that allow for broadening the biofabrication window toward better printing results in terms of shape-fidelity and accuracy, while simultaneously promoting cell performance (Figure 1A).To achieve this, we initially formulated a catalogue of relevant aspects for fiber design, taking into account the printing process as well as the subsequent postprinting period during which the construct should allow for tissue maturation (Figure 1).
One class of fibers that displays promising properties in accordance with these design criteria are electrospun dumbbellshaped polycaprolactone (PCL) fibers, with a diameter of 2.5 ± 0.6 μm and a length of 55 ± 29 μm. [15]We recently showed that the round-shaped ends of these collagen-coated fibers significantly enhance cell adhesion compared to collagen-coated con- tinuous fibers.In this study, we used a human glioblastoma cell line, U87, to study cell-fiber contacts, which have been previously attributed to morphologically triggered integrin-mediated cellfiber interactions at the collagen-coated fiber interface.Moreover, the fibers have been applied to tailor the rheology of alginate and Pluronic inks. [16]Based on these findings, the current study focuses on the applicability of such fiber systems to be printed together with cells.To assess the impact of fibers throughout gelation, during and after printing, we fibrillized a well-studied recombinant spider silk protein, engineered Araneus diadematus fibroin 4 (eADF4(C16)), with 16 repeating C-motifs resulting in a molar mass of 47 kDa, and applied it as a hydrogel matrix. [17]trikingly, without any RGD modification, this material does not promote cell adhesion and proliferation. [18]Consequently, the use of eADF4(C16) hydrogels allowed for determination of effects predominantly induced by collagen-coated PCL fibers and not by the matrix, rendering it ideal to explicitly study the impact of fibers on cell performance in a 3D setting.

Electrospinning
The fiber fragments were produced and coated according to a previously published method. [15]Briefly, 20% w/v polycaprolactone (45 kDa, Sigma Aldrich, MO, USA) was dissolved in 99% pure 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, abcr GmbH, Germany).The spinning solutions were electrospun at a rate of 1 mL h −1 with a voltage difference of 6 kV and a distance of 9 cm between the needle and collector.The rotating drum collector (Ø 85 mm) was set to a constant speed of 100 rpm and covered with conductive polyamide 6.6 mesh (mesh size 80 μm, Reichelt Chemietechnik GmbH+Co, Germany).After 60 min of spinning time, the resulting PCL nonwoven mats were immersed in ethanol and subjected to ultrasonication (UP200Ht, flat probe diameter 40 mm, Hielscher Ultrasonics GmbH, Germany) for fragment separation.During sonication, the PCL sheets were suspended in ethanol and cooled to below 10 °C to prevent melting of PCL.The PCL fiber fragments were filtered twice using coarse filters (mesh size 105 μm, Reichelt Chemietechnik GmbH+Co., Germany) and fine filters (mesh size 45 μm, Reichelt Chemietechnik GmbH + Co., Germany) and washed thoroughly.Afterward, the fragments were immersed in 10% 1,6hexanediamine (Sigma Aldrich, MO, USA) in 2-propanol and mixed in an overhead stirrer at room temperature for 1 h.After washing with Milli-Q water several times, the fragments were mixed with 1% glutaraldehyde (Sigma Aldrich, MO, USA) for 1 h at room temperature, washed thoroughly again, and suspended in 2.5 mg mL −1 collagen solution (ibidi GmbH, Germany) at pH 3.4.The mixture was gently mixed at 2-4 °C for 12 h.Samples were filtered and rinsed once with 0.1% acetic acid and subsequently with Milli-Q water until a neutral pH was reached.The suspension was filtered and stored as a paste with 20% w/v solid fraction.

Spider Silk Solubilization and Hydrogel Preparation
18a,19] In brief, lyophilized spider silk proteins were dissolved in 6 m guanidinium thiocyanate (Roth, Germany) at a protein concentration of 5 mg mL −1 for 1 h at room temperature.After sterile filtration through a 0.2 μm filter (Sartorius, Germany), the solution was dialyzed against 10 mm Tris/HCl pH 7.5 (Roth, Germany) using a molar mass cutoff membrane of 6-8 kDa (Spectra/Por, Fisher Scientific GmbH, Germany).The protein concentration was measured by microvolume spectrophotometry (NanoDrop2000, Thermo Fischer Scientific GmbH, Germany).To prepare the hydrogels, further dialysis steps were performed against 25% w/v polyethylene glycol (MW: 20 000 g mol −1 , Roth, Germany) using the same molar mass cutoff membrane to increase the protein concentration until final concentrations of 40 to 60 mg mg −1 were achieved.The solutions at the final concentration were stored on ice until the final mixing with cells and fiber fragments.

Turbidity Measurements
The self-assembly of eADF4(C16) spider silk into nanofibrils was analyzed in the presence of plasma-treated and non-treated PCL fragments and compared to PCL-free silk solutions.The respective PCL fragments were dissolved in the eADF4(C16) spider silk solution (1 mg mL −1 ) at a concentration of 0.5 mg mL −1 in TRIS/HCl pH 7.5 as a control.As in previous work, spider silk self-assembly was initiated by adding potassium phosphate (Roth, Germany) at a concentration of 150 mm.Samples without eADF4(C16) and PCL fragments served as controls.The fibrillization kinetics of eADF4(C16), indicated by the turbidity change of the solution, were recorded in triplicates at a wavelength of 340 nm at 20 °C using UV-micro-cuvettes (Sarstedt, Germany) and a temperature-controlled UV-spectrometer Cary 50 (Varian, Germany), as previously described. [20]

Fluorescence Analysis of Tetrametylrhodamin (TMR)-Labeled eADF4(C16)
The recombinant spider silk protein eADF4(C16) was functionalized with Tetrametylrhodamin (TMR) via an Nhydroxysuccinimide (NHS) group as described elsewhere. [21]yophilized TMR-eADF4(C16) was utilized in a similar fashion as unlabeled eADF4(C16).Due to its high fluorescence intensity, TMR-eADF4(C16) was blended with unlabeled eADF4(C16) at a ratio of 1:10 and subjected to an identical procedure as described in the section "Spider silk solubilization and hydrogel preparation," excluding the dialysis step against 25%w/v PEG.Collagen I-coated PCL-fibers were incubated with a 1 mg mL −1 TMR-eADF4(C16) solution either overnight (≈16 h) or for 2 h and subsequently washed three times by repeated centrifugation (5 min at 300×g) and resuspended in 1× PBS prior to use.The fibers were centrifuged again, and excess washing buffer was removed prior to mild air drying.The treated bones were then mixed into a 3%w/v eADF4(C16) hydrogel at 2wt% and incubated for gelation at 37 °C prior to imaging using a Leica DMi8 Fluorescence microscope with a TXR filter cube (540-580 nm).

Transmission Electron Microscopy (TEM)
TEM was performed to analyze the assembled eADF4(C16) nanofibrils, as previously described. [20,22]In brief, 5 μL of the appropriate samples were transferred onto pioloform-coated 100mesh copper TEM grids (Plano GmbH, Germany) and incubated for 5 min at ambient conditions.The samples were then washed twice with water, and the spider silk nanofibrils were counterstained with a 2% uranyl acetate solution for 2 min at RT, followed by another wash with water.The samples were allowed to dry in the dark for at least 12 h at RT before TEM imaging.TEM imaging of the dry samples was conducted using a JEM-2100 transmission electron microscope (JEOL, Japan), which operates at 80 kV and is equipped with a 4000 × 4000 charge-coupled device camera (UltraScan 4000, Gatan, Pleasanton, CA).

Scanning Electron Microscopy (SEM) Analysis of Composite eADF4(C16) Hydrogels
The eADF4(C16) hydrogel samples with and without fiber reinforcement were extruded from cartridges through a 21 G nozzle in the extrusion-based bioprinter setup (Inkredible, Cellink, AB, Sweden) as described previously.An amount of 200 mg was collected per extruded sample and transferred to a 1.5 mL Eppendorf tube with ventilation openings in the lid.The samples were frozen at −80 °C for 1 h and subsequently lyophilized in a Free Zone 2.5-L Benchtop freeze-drier (Labconco Corporation, MO, USA) for 72 h.The dried samples were carefully cut with a scalpel, and the resulting slices were fixed on SEM stubs with carbon adhesive discs.The prepared samples were stored in a desiccator to prevent rehydration until being sputtered with 1.3 nm of platinum and analyzed with SEM (Apreo VolumeScope, FEI Thermo Fischer Scientific Inc., MA, USA) using the "Standard" use case at magnifications of 1000 and 5000.

Cryo-Scanning Electron Microscopy (Cryo-SEM)
The samples were taken from cell culture and kept at 37 °C.These were rapidly frozen in slushed nitrogen at −210 °C after placing them between aluminum plates (d = 3 mm) with a 2 mm notch for sample fixation.All following transfer steps were performed at −140 °C with an EM VCT100 cryo-shuttle (Leica Microsystems).To generate a freshly fractured hydrogel surface, one of the aluminum plates was knocked off and freeze etched for 15 min at −85 °C under high vacuum (<1 × 103 mbar) in a Sputter Coater machine (ACE 400, Leica Microsystems).Afterward, samples were sputtered with 3 nm platinum and transferred to the SEM chamber (Crossbeam 340, Zeiss).Images of the hydrogel surface morphology were taken at −160 °C using an acceleration voltage of 8 kV.

Rheological Characterization
The rheological measurements were conducted using a Discovery HR-2 rheometer (TA Instruments, New Castle, DE, USA) with a 25 mm plate-plate geometry and a 200 μm gap.Temperature was controlled at 25 °C using a Peltier element in the lower plate, and a solvent trap with PBS was used to prevent evaporation effects.The samples were preconditioned for 60 s at a constant shear rate of 10 1 s −1 to eliminate any loading-induced tensions and to set the sample temperature.Frequency sweeps were performed at 1% deformation between 0.1 and 100 rad s −1 to determine storage modulus, loss modulus, and complex viscosity.Amplitude sweeps to determine storage and loss modulus were carried out at 10 rad s −1 with a deformation range of 0.01% to 500% within 100 s.Viscosity was measured using flow ramps with a logarithmical increase of stress from 1 Pa to 1 kPa within 500 s.

Bio-Ink Preparation
The PCL fiber fragments were suspended in sterile MQ water and subjected to ultrasonication in a chilled water bath, employing 1-s bursts with constant agitation.Lyophilized eADF4(C16) was then dissolved in 6 m GuaSCN for at least 30 min at room temperature and transferred into a dialysis tube with a 6-8 kDa cut-off.The protein solution was then dialyzed against 10 mm Tris/HCl pH 7.5 with 4 consecutive bath changes, after which the tube was kept in the dialysis bath for 16 h.To concentrate the spider silk solution, it was transferred to a 25 w/v% PEG solution, and UVvis spectroscopy was used to measure the concentration, which was then diluted to a stock concentration of 50 mg mL −1 .Prior to mixing all components, the spider silk solution was pre-gelled.During the pre-gelation process, first nucleation seeds formed, followed by fibrillization and subsequent gel formation by physical entanglement of these fibrils.The spider silk solution was rotated at 37 °C, agitating it every 20 min for 2-4 h until air bubbles produced by agitation would rise.The solution was kept on ice until used in cell culture.Cells were harvested, and 535 000 cells with a total medium content of 15%w/v hydrogel with different PCL-fiber (0%, 2%, or 3%) and spider silk concentrations (2% or 3%) were used in a total amount of 700 μL per condition.
The hydrogel components, fiber fragments, and cells were mixed according to the instructions in Table 1.All the components were carefully transferred under sterile conditions to 2 mL Eppendorf tubes and gently mixed by pipetting.The final solutions were loaded into 3 cc cartridges with matching pistons (Drifton, Denmark), and the cartridges were sealed by luer-lock syringe plugs (Braun, Germany).These cartridges were placed into 50 mL falcon tubes and attached to an overhead rotating shaker (Thermo Scientific Tube Revolver Rotator 88881002) to keep them at 37 °C for 16 h at 3 rpm.This was done to prevent sedimentation of fibers or cells until the complete gelation of the hydrogel had occurred.

Printing
The cartridges containing fully gelled hydrogels were fitted with straight steel needles, providing a 0.51 mm inner diameter and 30 mm full length (21G, Drifton, Sweden).Bio-ink printing was performed using an extrusion bioprinter (Inkredible, Cellink, Sweden) with a customized G-Code of a 4-layer structure resembling the shape of the number 8. The printing speed was set to 10 mm s −1 , and the printing pressure was adjusted dynamically in the range of 1.0-1.6 bar according to the bio-ink's printing behavior.Three repetitions of the same structure and bio-ink were printed on a polystyrene surface.After printing, the constructs were left to stabilize for 10 min and subsequently transferred to 24-well plates, immersed in cell culture medium, and incubated at 37 °C.

Cell Viability
The viability of the 2D and 3D samples was assessed on day 1 and 4 after printing U87 cells.The spider-silk constructs were washed in PBS and incubated with 2 × 10 −6 m Calcein-AM (green, indicating living cells; Thermo Fisher Scientific, Waltham, MA, USA) and 2 × 10 −6 m ethidium homodimer I (red, indicating dead cells; Sigma-Aldrich, St. Louis, MO, USA) in PBS at 21 °C for 30 min.For each time point, 3 images were captured from at least three independent samples to determine the number of live and dead cells.

Confocal Laser Scanning Microscopy and Image Acquisition
The fluorescence and cell viability images were captured using an inverted Olympus IX81 microscope that was equipped with an Olympus FV1000 confocal laser scanning system, a FVD10 SPD spectral detector, and diode lasers with wavelengths of 405, 495, and 550 nm, all from Olympus in Tokyo, Japan.All images displayed were captured using an Olympus UPLSAPO 10× objective (air, numerical aperture 0.4) and an UPLFLN 40× objective (oil, numerical aperture 1.3).Cell viability z-stacks were captured throughout each sample with a step size of about 2-3.52 μm.
The stacks/images were processed using ImageJ/Fiji 1.31 (Im-ageJ citierung fhelt).The Imaris software from Oxford Instruments in Abingdon, UK was used for 3D reconstruction, video generation, and analysis of the z-stack images to quantitatively assess the number of live and dead cells.

Analysis of Life Images Evaluating Shape Descriptors: Circularity, Roundness, Solidity, and Sphericity
ImageJ/Fiji 1.31 was used to analyze circularity, roundness, and solidity. [23]Image analysis was performed with maximal intensity z-projections of 10× images (life assay; U87) and 40× images (TdTomato-farnesyl expressing U87 reporter cells).At least 9-15 cells were analyzed for each condition.Sphericity was analyzed using ImageJ/Fiji 1.53t.A minimum cell volume of 100 μm 3 was assumed to remove background noise.The plug-in "3D Suite" was used, to rebuild the 3-dimensional shapes of each cell. [24]A threshold value between 128 and 255 was suitable for all images.Each area of a detected cell was enveloped with an ellipsoid to analyze the 3D cell size and sphericity.

Statistical Analysis
GraphPad Prism 8.3.0 (Graphpad Software, San Diego, CA, USA) was used to calculate mean values, standard deviation (SD), standard error of the mean, and values for statistical significance.Statistical significance was estimated *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.A test to identify outliers was performed using ROUT´s method with a Q of 0.5%.

Impact of Collagen-Coated PCL-Fiber Fragments on the Gelation of eADF4(C16)-Hydrogels
Previous studies have demonstrated accelerated nanofibril formation due to protein self-assembly and thus, gelation of eADF4(C16) when cells were incorporated.Considering that the optical properties of eADF4(C16) changed during self-assembly into nanofibrils and consequently upon gelation from transparent to opaque (Figure 2A), turbidity tests were conducted to analyze the impact of collagen-coated PCL fibers on its fibrillization behavior.PCL-seeded eADF4(C16) solutions reached half maximum after 22.5 h, whereas for pure, non-seeded eADF4(C16), half maximum occurred after 38.5 h, indicating a significantly accelerated fibrillization process (Figure 2B).In accordance with other literature, [25] the Avrami model turned out to be a sufficiently accurate fit for these self-assembly kinetics: With A ∞ = absorption after self-assembly/fibrillization (in normalized case: 1), A t = time-dependent absorption, A 0 = absorption at t = 0 min, k = rate constant (depends on nucleation and growth rate), n = Avrami's exponent (depends on geometry of growth and nucleation type).Transmission electron microscopy (TEM) imaging (see Figure 2C) showed the formation of eADF4(C16) nanofibrils with diameters of 9.0 ± 1.0 nm (n = 20), indicating fractal fibrillar growth upon protein self-assembly and consequent gelation. [26]Liu and Sawant adapted Avrami's equation for the case of time-dependent fractal growth, replacing n with the fractal dimension D f . [27]Interestingly, besides the significantly increased rate constant k, a decrease in D f from 2.85 to 2.34 was observed (Figure 2B).
Gong et al. observed comparable effects when studying the fractal behavior of regenerated silk fibroin (RSF) upon gelation. [28]They showed that increasing RSF concentration from 0.5% to 1% resulted in a change in D f from 2.40 to 2.79.They attributed this effect to the assembly of less open structures at higher RSF concentrations since a higher D f is generally associated with denser structures.In other work, it was reported that incorporation of nanocomposites results in a reduction of the Avrami exponent, which was assigned to a heterogeneous nucleation mechanism. [29]Previously, it has been shown that eADF4(C16) can form nanofibrils on eADF4(C16)particles, which act as nucleation seeds. [30]In this work, we applied the same procedure using Tetramethylrhodamine (TMR)labeled eADF4(C16) to prove silk protein assembly on the surface of collagen-coated PCL fibers (Figure S1, Supporting Information).Based on these observations, we hypothesize that collagencoated PCL fibers might act as a nucleating surface for fibrillization, resulting in changes in the internal hydrogel structure.
This hypothesis is further supported by scanning electron microscopy (SEM) images showing structural changes of freezedried eADF4(C16) hydrogels as a result of PCL fiber addition (Figure 3A1-4).Further analysis applying cryo-SEM reveals reduced wall thickness and a more filigree structure with incorporated PCL fibers (Figure 3B1-4).This examination was addi-tionally quantified by applying ImageJ software to analyze wall thickness and pore size of cryo-SEM images (Figure 3C1,C2).The data support the hypothesis of PCL fibers acting as nucleation seeds, since the fibers are all covered with silk from which branched walls extend, leading to reduced pore sizes compared to pure eADF4(C16) hydrogels (Figure 3B5).Moreover, irrespective of whether fibers were incorporated or not, most samples displayed regions with branched nanofibrils (see Figure 3B6), comparable to what was observed via TEM imaging (Figure 2B).It is most likely that these structures are present all over the hydrogels but get lost upon sample preparation for cryo-SEM.Therefore, it can be concluded that eADF4(C16)-hydrogels are composed of a microporous network of interconnected sheet-like structures that are infiltrated with a nanofibrillar matrix.The addition of PCL fibers displays a significant impact on the microporous structure, resulting in reduced wall thickness and smaller pores.

Rheological Impact of Collagen-Coated PCL-Fibers on eADF4(C16) Hydrogels
One important aspect of applying filler materials to bio-inks is tailoring rheological properties.In this work, shear rheological measurements were conducted to assess the potential of tailoring bio-ink rheology via incorporation of PCL fibers into the matrix (Figure 4A).Strikingly, in a strain sweep experiment comparing pure eADF4(C16) hydrogels with PCL-fiber-reinforced ones, the linear viscoelastic region (LVE, according to ISO 6721-10) was reduced from a critical amplitude of 8.0% (w.o.PCL) down to 1.8% (2% PCL) and 1.6% (3% PCL) (Figure 4B, green line).This correlated with the amplitude at which the flow point occurred, decreasing from 32.5% (w.o.PCL) to 14.1% (2% PCL) and 13.9% (3% PCL) (Figure 4B, gray line).Furthermore, PCL fibers had a  significant impact on the LVE storage modulus, which increased from 1.3 kPa (w.o.PCL) to 3.7 kPa (2% PCL) and 4.8 kPa (3% PCL) (Figure 4B, blue line).16a] The shear-rate dependent amount of complex viscosity was calculated from the oscillation experiments, and the power-law model (Equation S1, Supporting Information) was used to fit the shear-thinning region.Additionally, flow experiments were conducted, and the power law was used to fit the shear-thinning region of the data (Figures S2 and S3, Supporting Information).In Figure 5. A) Oscillation and flow measurements were conducted, and the resulting shear-thinning behavior was quantified using a power-law model.The dynamic consistency index K and the complex consistency index K* were then correlated.B) Furthermore, the power-law parameters based on flow experiments were used to calculate the theoretical extrusion profile and shear stress in the nozzle.To provide a better sense of scale, the sketched cells and fibers are drawn to their relative sizes compared to the radial dimensions of the nozzle.general, it was observed that the addition of 2% PCL-fibers resulted in a reduction of the consistency index to 40 ± 2% for 2% w/v eADF4(C16) and 72 ± 6% for 3% w/v eADF4(C16).Increasing the PCL-fiber concentration to 3% resulted in consistency indices comparable to the respective pure hydrogels (104 ± 2% for 2% w/v eADF4(C16) and 110 ± 16% for 3% w/v eADF4(C16)).These trends were observed in both oscillation and flow experiments (Table S1, Supporting Information).It is not surprising that the Cox-Merz-relationship was not applicable for such a complex material system, but the determined parameters differed significantly.Figure 6A shows the relationship of flowbased (dynamic) and oscillation-based (complex) consistency indices K and K* that were determined for different matrix and filler concentrations.Interestingly, the factor between the data pairs was quite distinct (19.2 ± 1.4), indicating a systematic phenomenon.When applying a capillary model (Equation (S2), Supporting Information) together with the printing parameters (needle length, needle diameter, printing pressure) and the consistency and shear thinning indices (K, n), the predicted printing velocities were by far too low in the case of oscillatory-based data (Figure S4, Supporting Information).In the case of flowbased data, the predicted printing velocities were in the right range, although higher when compared to the applied printing speed of 10 mm s −1 (Figure 5B, above).Paxton et al. determined the consistency and shear thinning indices (K, n) of Poloxamerand alginate-based bio-inks by conducting flow experiments with power-law regression. [31]Using the same capillary model to predict the velocity profile, Paxton et al. found that the experimental extrusion velocities were below the theoretically determined ones in all cases.Interestingly, power-law parameters from other publications that applied composite hydrogels as bio-inks also resulted in overestimation of hydrogel extrusion velocity. [21]This is highly consistent with the data presented here.Since these predictions are solely based on shear flow, future work will focus on unraveling the role of elongational flow at the tapering between the cartridge and the nozzle.Notably, when comparing the cal-culated shear stress profiles of pure eADF4(C16) hydrogels and PCL-filled eADF4(C16) matrices, the difference is below 25% at its maximum at the wall, even at a high filler concentration of 3% PCL (Figure 5B, below).This further emphasizes that the impact of PCL fillers is quite high in the solid state of the hydrogels (as shown in Figure 4) but relatively low in the shear-thinned liquid state.

Impact of Collagen-Coated PCL-Fibers on Cell Viability in eADF4(C16) Hydrogels
We used a glioblastoma cell line, U87 (U-87 MG), [15,32] to investigate the impact of PCL-fiber-reinforcement on cell performance in eADF4(C16) bio-inks.There are three critical steps that represent the different environmental conditions to which cells are exposed upon printing.Firstly, after combining eADF4(C16) hydrogels, fibers, and U87 cells in printing cartridges, a 24-h gelation step with rotation is required.The second step is extrusion printing via shear-induced flow of the cell-laden bio-inks through a nozzle, followed by a post-printing phase during which the material gelates, and cells are incubated to allow for potential cell adhesion, proliferation, and migration (Figure 6A).During gelation and printing, U87 cells are exposed to shear stress and elongational flow, processes that can potentially lead to cell damage and reduced viability.After printing, U87 cells typically start to adhere and proliferate, but only if the surrounding construct and its matrix have suitable properties.
In this study, we printed different bio-ink formulations with varying PCL-filler content and eADF4(C16) hydrogel concentrations.We analyzed U87 cell viability one day after printing to determine the impact of fibers on cell survival during the printing process.The comparison between 2% and 3% PCL added to 3% w/v eADF4(C16) showed a significant decrease in cell viability for both PCL concentrations (Figure 6B).This is particularly interesting because the addition of 2% PCL significantly decreased the Figure 6.A) Schematic drawing (created with BioRender) illustrating the sample preparation process.B) Correlation between PCL concentration (0%, 2%, and 3%) and cell viability (%).C) Bar diagram showing the rate of proliferation for 2% and 3% w/v eADF4(C16) and 0% versus 3% fiber loading.Significance was calculated using student's t-test: *** p ≤ 0.001.D) Bar diagram of cell viability assays performed on day 1 and 4 after printing.Cells in 2% and 3% eADF4(C16) with and without PCL fibers are depicted.Significance was calculated using student's t-test: * p ≤ 0.05; **** p ≤ 0.0001.E) Representative life (green)/dead (red) images of 3% w/v eADF4 (C16) hydrogels without PCL-fibers and with 3% PCL-fibers on day one and four.
overall viscosity of shear-thinned eADF4(C16) hydrogels (Figures S2 and S3, Supporting Information).Consequently, the mechanism causing cell damage cannot be attributed to viscosity alone, but rather it is a result of cell-fiber interaction during printing.However, both 2% PCL and 3% PCL samples tended toward an increase in cell viability at day four post-printing, with 3% PCL having a greater increase than 2% PCL (Figure 6B).Therefore, the presence of PCL-fibers can have positive effects on cell viability post-printing, despite the initial detrimental effect on cell viability during printing.
Since there was an increase in viability in 3% eADF4(C16) hydrogels with 3% PCL-fibers from day one to day four postprinting, we further examined the proliferation rate (Figure 6C).As expected, while pure eADF4(C16) hydrogels displayed excellent biocompatibility in terms of viability, U87 cells did not proliferate, and cell numbers decreased over time (Figure 6D).However, the incorporation of PCL fibers significantly increased the proliferation rate for both the 2% w/v and 3% w/v eADF4(C16) formulations, indicating an improved microenvironment for the cells (Figure 6C).Upon analyzing cell viability on day one and day four after printing, we observed a high number of viable cells in both 2% w/v (day 1: 96.0 ± 2.1%; day 4: 98.7 ± 0.3%; p = n.s.) and 3% w/v eADF4(C16) (day 1: 97.1 ± 0.9%; day 4: 98.5 ± 0.6%; p = n.s.), which is consistent with previous studies. [33]The addition of 3% PCL-fibers significantly decreased cell survival, especially in 3% w/v eADF4(C16) hydrogels (Figure 6D,E).In contrast, cell viability was not significantly decreased in 2% w/v eADF4(C16) hydrogels with the same amount of PCL (3%).This can once again be attributed to the rheological properties of the bio-inks, as the consistency index of 3% eADF4(C16) with 3% PCL-fibers is about fivefold higher than that of 2% eADF4(C16) with 3% PCL-fibers (Table S1, Supporting Information).Consequently, the decrease in cell viability can be attributed to cell-fiber interactions during printing, and the extent of this effect significantly depends on the rheological properties of the matrix.

Impact of Collagen-Coated PCL-Fibers on Cell Morphology Using 2D z-Stack Projections
Healthy cells in a cell-friendly environment undergo a change in morphology from a roundish shape, after splitting or printing, to an extended, differentiated state, characterized by elongated, triangular, or star-shaped morphologies with small protrusions. [15]o analyze the impact of fiber-reinforcement in 3% eADF4(C16) hydrogels on cell morphology, we initially used low magnification (10x) and calcein-am live staining to determine cell shape development (Figure 6E).Therefore, z-stacks were projected into a 2D-image resulting in a single display of all cells in a sample.Individual cells were analyzed according to different shape descriptors: circularity, roundness, and solidity as previously conducted by Gensi et al. [34] Moreover, recent work demonstrated the suitability of such indicators to describe neuronal and astrocyte morphology in 3D hydrogels. [35]Circularity indicates the presence of cell protrusions, roundness describes the cell's state with respect to elongation, and solidity displays spreading and thus the maturational state.A perfectly round cell without any protrusions resembles a value of 1, whereas morphological changes result in a decrease in the shape descriptors (Figure 7A).When comparing cell shapes in 0%, 2%, and 3% PCL-fiber-supplemented eADF4(C16) hydrogels (3%) with cells grown in 2D, a clear and significant trend toward more spread cells was observed when PCL fibers were present (representative images: Figure 7B).Moreover, quantitative analysis underscores this observation: upon PCL-fiber addition, cell morphology significantly changes toward a more mature state comparable to cells in the 2D environment of a cover slip (Figure 7C).
To conduct a more in-depth analysis of cell morphological changes, a td tomato-farnesyl U87 cell line was used in this study.This reporter cell line labels the cell membrane precisely, allowing for high-resolution imaging of cell morphology, especially for small protrusions.Moreover, the culture period was extended to seven days, and experiments were conducted including the maximal printable PCL-fiber loading of 5% in 3% eADF4(C16).Quantitative analysis of shape descriptors showed that the addition of 2% and 3% PCL-fibers did not have a significant impact on circularity and solidity on day one (Table S2, Supporting Information).After seven days of culture, cells displayed more protrusions evidenced by the significant decrease of circularity and solidity for both 2% and 3% PCL-fiber concentrations (Figure 8A).It was observed that cells in hydrogels with 3% PCL showed the best cell shape performance, which is consistent with the results from our previous experiments (Figure 7, additional analysis of shape descriptors is available in Figure S5, Supporting Information).
Further increasing the PCL-fiber concentration to 5% resulted in comparable effects as for 2% and 3% but did not induce further improvement.This can most likely be attributed to filler-overload of eADF4(C16) hydrogels as the consistency is paste-like, indicating that there is minimal space for cells to perform.Using reconstructed confocal imaging, we were able to clearly localize cells in direct proximity to the fibers (Figure 8C1-C3).Building upon our previous findings in 2D cell culture, we hypothesize that when cells come into contact with fibers, they adhere and spread along the fiber surface. [15]Cryo-SEM confirmed this observation by showing direct cell adhesion and spreading on PCL fibers (Figure 8D).These results are highly consistent with the shape descriptor data, which demonstrates the promoting effect of fiber addition on U87 cell performance.
In terms of circularity, we demonstrated that a higher magnification (40× vs 10×) is required to capture small cell protrusions, as the measured values changed from ≈0.6 (10×, Figure 7C) to around 0.4 (40×, Figure 8A).In contrast, solidity was consistent for both methods (10×, 40×, Figure 8B), which is expected since the underlying structural changes are on a larger scale.This result highlights the importance of well-designed methodological approaches in the context of cell morphology, as demonstrated using the U87 reporter cell line td tomato-farnesyl.

Cell Morphology in 3% eADF4(C16) Hydrogels Using 3D Cell Shape Reconstructions
Based on the 3D reconstruction of high magnification (40×) confocal images, we investigated the 3D shape descriptor sphericity.Sphericity is a 3D counterpart to circularity.A cell with a corresponding value of 1 resembles a sphere, while smaller values relate to a more differentiated or spread-out cell shape with protrusions (Figure 9A).In general, we observed that PCL fibers induced significant morphological changes in cells from day one to day seven (Figure 9B).This positive impact of PCL fibers on cell performance is highly consistent with the results of circularity (Figure 8A).To verify the methodological conclusion from Section 2.4, we complemented our examinations by applying calcein-am staining with 10× imaging (Figure 9C, see Table S2, Supporting Information).The values for sphericity obtained are clearly higher, similar to what could be seen for circularity.The consistency of our findings emphasizes the need for higher resolution imaging when studying cell morphology.

Conclusion
Since eADF4(C16) hydrogels are non-adhesive but cytocompatible, they are an excellent material system for explicitly studying the impact of fibers on cellular behavior in a 3D environment.Taking into account all the presented findings, it can be concluded that collagen-coated PCL-fillers have an impact on the different stages of the biofabrication process.Upon printing, they had only a minor influence on the rheology of the shear-thinned hydrogels but significantly reduced cell viability.Interestingly, the calculated shear stresses to which the cells are exposed inside the nozzle are below 1 kPa at their maximum and thus lower than what was previously described as critical (>5 kPa). [36]Therefore, the loss of cell viability must be attributed to the direct mechanical impact of the fibers on cells.In previous work, we demonstrated the high affinity of cells for adhering to this fiber type. [15]ased on this observation, we hypothesize that cells may already adhere to multiple fibers before printing, which may cause cell damage due to the bending or rupturing of fiber clusters or via cell penetration in the capillary flow (Figure 10A).On the other hand, we observed a clear tendency that cells printed together with fibers perform better in terms of proliferation and cell morphology within a period of 4 to 7 days compared to cells that were printed in pure eADF4(C16) hydrogels.This positive postprinting effect can be attributed to direct cell-fiber interaction, as demonstrated via confocal imaging and cryo-SEM.Here, we could observe the proximity of cells and fibers, as well as cells spreading along fibers (Figure 10B).The results obtained in this 3D environment were highly consistent with the positive impact of such fibers previously reported for 2D cell culture studies. [15]urthermore, SEM and cryo-SEM imaging have revealed that the incorporation of collagen-coated PCL-fibers results in a structural change in eADF4(C16) hydrogels.PCL-fibers act as nucleating surfaces, as demonstrated by turbidity and fluorescence measurements, as well as cryo-SEM imaging.The presence of PCLfibers leads to hydrogels with smaller pores and thinner walls, accompanied by a significant reduction of the critical amplitude of the linear viscoelastic region (LVE).This combination of characteristics, i.e., thin walls and reduced LVE, is believed to reduce the required cellular forces for plastically deforming such a microenvironment.Further research will be conducted to study this phenomenon.Nevertheless, the findings of this study, based on multiple methods, clearly underscore the importance of distinguishing between printing and post-printing, as it is likely that similar initially detrimental effects on cell viability would also occur in other composite bio-ink systems.These effects may be inevitable to improve the post-printing performance with respect to cell viability and maturation.

Figure 1 .
Figure 1.Schematic overview of the perspective of composite bio-inks regarding A) the biofabrication-window and B)the respective fiber design criteria that affect the printing process and subsequent cell performance within the composite bio-ink.

Figure 2 .
Figure 2. Pictures of cartridges filled with 3% w/v eADF4(C16) hydrogels A1) without PCL and A2) with 3% collagen-coated PCL fibers before and after gelation.B) The self-assembly kinetics were examined via turbidity measurements applying a wavelength of 340 nm for absorption.The resulting curves for eADF4(C16) and PCL-seeded eADF4(C16) hydrogels were fitted applying the Avrami model (fitting parameters are given in the legend, D f = fractal dimension, k = rate constant).C) TEM-imaging revealed eADF4(C16)-fibrillization in presence of PCL-fibers.

Figure 4 .
Figure 4. A) Rheological assessment of 3% w/v eADF4(C16) hydrogels with 2% and 3% PCL-fibers incorporated.B) Further data analysis shows the reduction of the Linear Viscoelastic Region (LVE) and the flow point whereas G' increases when PCL-fibers are added to eADF4(C16) hydrogels.

Figure 10 .
Figure 10.Schematic model to describe the impact of PCL-fibers A) during and B) after printing when compared to pure eADF4(C16) hydrogels.

Table 1 .
Bio-ink formulations as used in this study.