Polymer and bioink mixed strategy for human 3D printed Hippocampus model development

Felipe Valle Fortes Rodrigues PUCRS Brain Institute Gabriele Zanirati PUCRS Brain Institute Isadora Ghilardi PUCRS Brain Institute Allan Marinho PUCRS Brain Institute Thuany Maraschin PUCRS School of Technology Nara Regina Bastos PUCRS School of Technology Carlos A. dos Santos PUCRS School of Technology Vitória Pimentel PUCRS Scienti c Initiation Program Lia Siqueira PUCRS Scienti c Initiation Program Daniel Marinowic PUCRS Brain Institute Jaderson Costa DaCosta (  jcc@pucrs.br ) PUCRS Brain Institute


Introduction
Neurodegenerative diseases are de ned by impairment of neuronal network function. Such loss can be re ected in connectivity, individual cell integrity or even in the protection of the myelin sheath (1,2).
Probably one impacting factor in the incidence and prevalence of neurodegenerative pathologies is time and, as demonstrated in the 2015 United Nations report, the number of people over 60 years old tends will increase, highlighting the risk associated with these conditions. Pathologies (3,4). The hippocampus is involved in the consolidation of memory and learning (1). In the clinic, there are a variety of diseases that affect the hippocampus, such as Alzheimer's and other neurodegenerative processes, like epilepsy and depression (5). Although it is a key structure in the injury resulting from numerous pathologies, clinical and laboratory practice still lacks models capable of simulating the arrangement of this structure, harboring di culties regarding the study and development of biological interventions to these pathologies. The arrangement of monolayer cells, present in traditional culture, hinders the construction of models that mimic the cellular arrangement and physiological connectivity of the tissue. 3D printing, together with the use of biocompatible materials and hydrogels, contribute to the creation of structures with greater cytoarchitectural likelihood (6-9). These could allow a greater understanding of the intrinsic processes of the structure and, thus, help in the development of interventions for the pathological processes mentioned. Biomaterials are materials that aim to support and stimulate regeneration or replace functional tissue without causing rejection of the host, among them are biocompatible polymers and mesenchymal stem cells (MSC), in addition to their products such as exosomes (10)(11)(12)(13). Hydrogels, in turn, are oxygen and nutrition permeable compounds, which allow cell migration and are widely used as tissue-speci c extracellular matrix mimetizers, serving as scaffold for cell development (14)). Thus, this work aims to build a 3D structural model of the hippocampus, through 3D printing using biocompatible polymers and cells in hydrogel medium.

Statistical analyses
All statistical analyses were performed using GraphPad Prism 8 software. Samples were analyzed through ANOVA with Tukey post hoc and p < 0.05 was considered statically signi cant.

Cell culture
Experiments were conducted with NIH-3T3 cell lineage, from the Cell and Tissue Technology Laboratory, Brain Institute (BraIns) at PUCRS. Cells were cultivated in the following maintenance medium: Dulbecco's Modi ed Eagle Medium (DMEM) supplemented with 10% fetal serum inactivated bovine (SFB), 100 U/mL penicillin/streptomycin and 100 µg/mL gentamicin, all purchased from the manufacturer GibcoTM (Gibco™, Life Technologies, California, USA).

Polymeric biomaterials preparation
Membranes of different lm biomaterials were prepared in order to establish the most suitable for use in a 3D printed matrix. The biomaterials used were Policaprolactona (PCL), Poly Lactic-co-Glycolic Acid (PLGA), Polylactic acid (PLA) and Polypyrrole (PPY) in membrane. For the experiments, the membranes were used in the following con guration: PCL, blend of 70% of PCL and 30% of PLGA, PLA and PPY dispersed over the blend in the proportion of 10% m/m (15). For the experiments, lm/membrane biomaterials were used. Brie y, the polymer matrices or their blends were prepared using the solvent evaporation methodology. The polymers were dispersed in chloroform and later placed in petri dishes. The solvent was evaporated for 24 h at room temperature (15). Blends containing conductive polymer were prepared from the addition of previously synthesized PPy nano bers (10% w/w).

Contact angle
To evaluate lm surfaces hydrophobicity, contact angle values were determined using a Phoenix 301 goniometer, SEO Company. The experiments were carried out with deionized water and determined soon after the contact of the drop with the material to be analyzed. Five drops of deionized water were applied and considered the medium value.

Infrared spectroscopy
The molecular structure of the materials produced was determined by Fourier transform infrared spectroscopy (FT-IR) in order to identify the functional groups present in the samples. The infrared (IR) measurement was performed through absorption spectrum, between 400-4000 cm − 1, in Perkin-Elmer spectrometer. It was used materials without exposure as control.

Cell viability
To verify the cytotoxic potential of the materials, MTT protocol was performed through the extraction method following the norms of ISO 10993-5. Brie y, for the production of conditioned medium, the biomaterials were exposed, in the proportion of 3cm 2 /mL, to the medium used for the cultivation of cells in their maintenance phase for a period of 24 hours, 72 hours and 7 days at culture conditions stated above. Cells were plated at a density of 5 x 10 4 cells per well in a 96-well plate and cultured for 24 hours in maintenance medium. Afterwards, cells were exposed to this conditioned medium, for 24 h and then incubated with the MTT solution at a concentration of 5 mg/mL for 2 hours. Absorbance was measured at 570 nm in SpectraMax equipment, molecular devices, San Jose, California.

Direct cell adhesion
To evaluate the adhesion potential, cells were cultured directly on 3cm 2 of the biomaterials, at a density of 5 x 10 4 cells for 24 hours. Afterwards, the biomaterials were xed with 4% paraformaldehyde for 15 minutes, washed and incubated with nuclear staining, 4′,6'-diamino-2-phenyl-indole (DAPI), (10 µg/mL) for 15 minutes. Ten random elds were captured using a confocal microscope LSM 5 Exciter, Zeiss and the stained nuclei present on the matrices were quanti ed using Image Pro Plus® software version 6.0 (Media Cybernetics, Rockville, MD, USA)

Indirect cell adhesion
To assess the ability of the conditioned medium to in uence cell adhesion, the biomaterials were exposed, at a rate of 3cm 2 /mL, to the same medium used for cell culture or to the conditioned medium for 24 hours, 72 hours and 7 days. Cell culture was performed in 2 moments, culture phase and adhesion phase, and then they were stained and the absorbance veri ed. Figure 1 demonstrates an schematic design of this experiment. In the culture phase, cells were cultured for 24 hours with maintenance or conditioned medium, at a density of 5 x 10 4 cells per well in 96-well plates, trypsinized and replated in the same concentrations to start adhesion phase. In the adhesion phase, the cells were exposed to the maintenance medium or to the conditioned medium for 4 hours. The cells were then washed with phosphate buffered saline, xed with 4% paraformaldehyde for 15 minutes, washed with phosphate buffered saline, and incubated with cresyl violet for 15 min. After that, the absorbance was measured at 570 nm in the SpectraMax-M2 equipment, molecular devices, San Jose, California.

Magnetic Resonance (MR)
Structural magnetic resonance images were acquired on a General Electric (GE) 3 T Scanner with a 3D sagittal gradient acquisition by sequence echo (MP-RAGE). Repeat time of approximately 2300 ms, echo time of approximately 3 ms and inversion time of approximately 900 ms, and voxel dimensions were approximately 1.20 × 1.015 × 1.015 mm.

Scanning electron microscopy
The morphology and surface of the samples were analyzed using the scanning electron microscope (SEM), PHILIPS model XL30 with a resolution of 3.5 nm and a range of magni cations of 500 times for the lms of polymeric biomaterials and up to 20.000 times for the hydrogels, with accelerating voltage of 20 kV. Samples were metallized with gold in the Sputter Coater equipment (Balzers SCD050). In the case of hydrogels with cells, samples were xated with 4% paraformaldehyde for 24 hours, dehydrated with acetone prior to metallization protocol.

Polymer 3D printing
First for the hippocampal digital project model for the hippocampus, it was used Computer Aided Design (CAD) using the software CREO Parametric 3.0 in format *.stl for print. Data and measurements were extracted from the magnetic resonance imaging. For the 3D printing of the PLA mold, Fusion deposit modeling (FDM) technologies were used in Sethi S3 3D printer with simplify3D software. Print settings used as follows:

Results
To evaluate the feasibility of an structural and compatible with cell culture 3D model of the human hippocampus, it was performed a screening of biomaterials to determine which material is the most effective to produce the mold, considering biological compatibility and production e ciency. The dimensions of the hippocampus were obtained from the volumetry of the human hippocampus from magnetic resonance imaging of the brain of an adult subject without hippocampal pathology; after digital mold design and 3D printing, we evaluated the structural reliability of the mold by magnetic resonance imaging. For the structuring of the hippocampus anatomical layers, we used 3D printing a bioink

Contact angle
To analyze the materials hydrophobicity, Fig. 2 shows the drop of deionized water on the material and it´s repulsion. The difference between the angles. PPY and PLA groups stood out with the highest contact angles (80,57 ± 0,76 and 71,73 ± 1,70, respectively), PPY being the highest and with no statistical difference between the PCL materials (63,08 ± 3,17) and 70/30 (63,81 ± 2,32). The PCL group had statistically smaller angles than the PLA and PPY groups. The 70/30 group had statistically smaller angles than the PLA and PPY groups. The PLA group had statistically lower angles than the PPY group.
Scanning Electron Microscopy of Polymeric Biomaterials Figure 3 demonstrate the topographic difference between the materials, being controls materials not immersed in the medium and after being immersed in the period of production of the conditioned media and washed with deionized water. They were analyzed in the 24-hour, 72-hour and 7-day groups. The PCL and PLGA materials were atter, while PLA and PPY proved to be rougher. Figure 4 illustrate the infrared spectrum of the samples obtained from the chemical structure and demonstrate the infrared pro le of all materials, being controls materials not immersed in the medium and after immersed in the period of production of the conditioned media 24 hours, 72 hours and 7 days. The samples were characterized by FTIR in transmittance mode in the range of 400-4000 cm-1 to identify the structural characteristics of the materials. None of the materials underwent changes in chemical bonds, during the analyzed period, as can be seen in the gures when the control spectrum is superimposed with the set after incubation in culture medium for 24, 72 hours and 7 days. The infrared spectrum of the blends with the addition of polypyrrole presents functional groups characteristic of blend due to pyrrole be 10% deposited over the blends.

Direct cell adhesion
Afterwards, we proceeded with the biological compatibility experiments. Figure 5 shows direct adhesion, cells were plated directly onto the biomaterials and cultured for 24 hours under normal conditions. Cells cultured in standard commercial culture plates were used as control. The 70/30 blend (106.12 ± 27.4) was signi cantly more effective in allowing cell adhesion, demonstrating a statistical difference in relation to PCL groups (62.88 ± 15.1) and PPY (71.21 ± 9.2 ), but the same did not occur in relation to the PLA group (80.6 ± 18.6). The other groups showed no statistical difference between them.

Cytotoxicity assay
Regarding biomaterial toxicity potential, Fig. 5 demonstrates the results from the MTT assay performed using the conditioned media. Cells were cultured for 24 hours exposed to the conditioned medium, 24, 72 hours or 7 days, and then evaluated for relative viability in relation to the control. There were no signi cant differences in any time or group, except for PCL group (0.94 ± 0.41) using conditioned medium for 7 days.

Indirect adhesion
For indirect adhesion, two strategies were used: one thinking about the immediate effect of contact and the other about reversing this effect. For that, the cells were always cultivated in conditioned medium or exposed to the conditioned medium only during the adhesion period, in which case they were cultivated under normal conditions. Polymer 3D mold printing Figures 8 and 9 shows the acquisition of measurements obtained from magnetic resonance imaging of the brain regarding the hippocampus and the mold printing. Figure 8 demonstrate the axial (A, B and C) and coronal (D) planes. Figure 8 (F and G) shows the human hippocampus mold for 3D printing. Figure 9 (A and B) demonstrates the printed mold with the inlet and outlet on the top plate, to allow the mold to be associated with a pump to circulate the culture medium, aiming at the cultivation of cells inside. In Fig. 9 (C and D) we also verify the internal measurements that were in agreement with the measurements of the human hippocampus on magnetic resonance imaging.The magnetic resonance image of the impression after printing (Fig. 9, C and D) resembles the image of the human hippocampus obtained in the sagittal section (Fig. 8, E). 3D hydrogel model Figure 10 demonstrate the mold made with StandardInk™ and PlayInk™ bioinks, provided by the 3D printer manufacturer, It was built using smaller blocks with a lower support layer. This made it possible to use four different elements in its elaboration, consistent with the histological regions of the hippocampus. Figure 11 demonstrates cell adhesion by SEM.

Discussion
This work aims to enhance current knowledge regarding how to enable the use of emerging 3D technologies in the medical eld (16) for the development of a model for research on neurodegenerative diseases; we chose the hippocampus for its relationship with memory processes (17) and its involvement in pathologies such as epilepsy (18) and Alzheimer's disease (19). Thus, the results analyze the acquisition of magnetic resonance images and the translation to the 3D printer using Fusion deposit modeling (FDM). After obtaining the computational model, we questioned which would be the most suitable material and from that, PLA, PCL, PLGA and PPY were elected as candidates. We worked with the membrane materials for the tests, using the blend of PCL and PLGA in the proportion of 70 to 30 percent respectively and the PPY was dispersed directly over the blend (15). In our contact angle results, PPY and PLA materials reached more hydrophobic contact angles but within the range of 70 to 80 brought by the literature. Zelzer and colleagues also demonstrated the survival, adhesion, and migration of hydrophobic hippocampal neurons within the spectrum of our results (20). The addition of arginine-glycine-aspartic acid (RGD motifs) amino acid sequences, which are essential for neuron cell adhesion (21), increases the hydrophobicity of the materials (22). This hypothetical mechanism could be related to PLA and PPY materials having higher contact angles and would bene t from using these materials. Another point that could be explored in future is that more hydrophobic materials may be related to the activation of astrocytes because the high contact angle can lead to modulation of the m1 and m2 pro le (23,24).
To verify the interaction of biomaterials with the surrounding environment, simulating in vitro how the presence of the biomaterial in a tissue could affect its surroundings, we performed indirect adhesion experiments. With the PPY group with conditioned medium for 24 hours, there was an increase in adhesion when it was returned to the conventional culture medium, demonstrating a possible reversal of some process that was affecting the cell. With the conditioned medium in 72 hours, the opposite occurred, a possible potentiation of the cell adhesion process in contact with the conditioned medium. With PLA, a possible potentiation of this process occurred when kept entirely in conditioned medium. Considering that we conducted an innovative experimental strategy between cultivation and the performance of the indirect adhesion experiment, aiming precisely to understand more carefully the changes in the adhesion process and possible reversal of the process, it is necessary to carry out complementary experiments to con rm this process and it´s mechanisms. From these results and considering the feasibility of PLA being printed directly on the 3D printer, we considered PLA as the most suitable material for the construction of the mold due to the result in indirect adhesion can bene t the mold, did not present cytotoxicity and, statistically, was not inferior to other biomaterials in direct adhesion. We then printed based on the MRI measurements with space to be attached to an infusion pump, Jin et al. in an in vivo model of ischemia injury, demonstrated that the addition of arti cial brain spinal uid achieved good reparative potential (25).
The work from Mark and colleges. provides a solid model of the hippocampus (19). These models can be used to aid in medical diagnosis, histology (26), as a surgical targeting tool (27), and for teaching (16).
However, to evaluate treatments and perhaps enable the practice of precision medicine, it is necessary to build scaffolds that allow cellularization, stablishing parameters that can be controlled and treated in addition to the patient's anatomy.
Our next questions were related to tissue histology and cell integration with respect to brain tissue and conducted 3D printing using hydrogels was the answer to these questions (28). Decellularized tissue hydrogels can bring great bene t due to the maintenance of extracellular matrix elements and biochemical guides that support the processes of cell growth, orientation and differentiation (29). In addition, the similarity of mechanical characteristics can also bene t populations of the nervous system. Leipzig et al. demonstrated that stiffness closer to that of brain tissue bene ts the differentiation of neurospheres into neurons and myelination from oligodendrocytes (30). Crapo and colleagues also demonstrated that in addition to differentiation, neurons cultured in decellularized brain bioinks showed more projections than in bioinks from other tissues (31).

Conclusion
Our material selection of PLA that did not present cytotoxicity and, statistically, was not inferior to other biomaterials in direct adhesion, adds feasibility to the printing process because with PLA, the production of the negative mold can be done directly from the 3d software. The construction of the mold with bioink in small parts forming a whole, inspired by magnetic resonance voxels, made it possible for the mold to absorb the complexity of hippocampus regions. In addition, the model of construction of the hippocampus in hydrogel, which allow cell migration, serving as scaffold for cell development, contributes to the cellular layering, potentially mimicking real tissue cytoarchitecture. Taken together, our work can contribute to the 3D printing of a model of the human hippocampus using compatible biomaterial and bioink that allows simulating conditions in health or disease, seeking new therapeutic strategies.  Figure 1 Demonstration of the indirect adhesion experiment. Initially the cells were cultured ("Culture phase") in conditioned medium or under normal culture conditions (represented as DMEM). After the cells are placed to carry out the adhesion process exposed to normal conditions or to conditioned medium. Source: author.

Figure 2
Demonstration of the contact angle. Analysis of the contact angle of biomaterials (angle ± SD, p<0.05); * represents the statistical difference between the PCL and PLA group; ** represents the statistical difference between the PCL and PPY.

Figure 3
Scanning electron microscopy of the biomaterials demonstrating the surface behavior over the period for the production of the conditioned medium, 24 and 72 hours and 7 days. As control it was used the materials without incubation. Analysis of direct adhesion on biomaterials (Count number± SD, p<0.05); * represents the statistical difference between the 70/30 group and PCL; ** represents the statistical difference between the 70/30 group and PPY; *** represents the statistical difference between all groups and the control group. Figure 6 MTT viability assay using conditioned media for 24h, 72h and 7 days. * represents the statistical difference between PCL and the 3T3 control.    Execution of the proposed model of construction of the hippocampus in hydrogel regarding its histological regions, using arti cial material. In white, underneath, a support structure. In yellow, the region of the dentate gyrus. In pink, the region of CA3 and CA4. In blue the region of CA 1 and CA 2. SEM of cell in the hydrogel. In A, cells were bioimprinted and cultured for 24 hours before being xed and SEM performed. B and C demonstrate cell adhesion to the hydrogel