Comparison of axon extension: PTFE versus PLA formed by a 3D printer

Abstract Three-dimensional (3D) printers mainly create 3D objects by stacking thin layers of material. The effect of the tools created using the fused deposition modeling (FDM) 3D printer on nerve cells remains unclear. In this study, the effects of polytetrafluoroethylene (PTFE) models and two different types of polylactic acid (PLA) models (white or natural), were created using the FDM 3D printer on axon extension were compared using the Campenot chamber. Neurons were isolated from the dorsal root ganglia and added to the central compartment of the Campenot chambers after isolation, processing, and culturing. On day 7, after the initiation of the culture, the difference of the axon extensions to the side compartments of each group was confirmed. We also compared the pH and the amount of leakage when each of these chambers was used. The PLA was associated with a shorter axon extension than the PTFE (white p = 0.0078, natural p = 0.00391). No difference in the pH was observed (p = 0.347), but there was a significant difference on multiple group comparison (p = 0.0231) in the amount of leakage of the medium. PTFE was found to be a more suitable material for culturing attachments.


Introduction
In recent years, three-dimensional (3D) printer technology has attracted considerable attention. 3D printing is a technology that creates real objects based on a 3D model designed using computer-assisted design (CAD) software. In most cases, thin layers of material are piled up by hardening with ultraviolet (UV) radiation or melting with heat to create a substance based on the 3D model.
Generally, by standardizing tools, it is possible to perform experiments with high reproducibility; if the data of the 3D model can be shared, 3D printers enable us to experiment with high reproducibility.
The mainstream 3D printing method is fused deposition modeling (FDM) [1]. This involves stacking materials, such as polymer, molten at high temperatures as a thin layer, and the common materials used are polylactic acid (PLA) and acrylonitrile-butadiene-styrene (ABS).
Attempts were being made to develop devices for in vitro research in the field of cell culture, such as using a cell culture platform created by a 3D printer to culture fibroblasts [2]. There is also a strong interest in the safety of materials formed by 3D printing, and several studies have been conducted. Some studies conclude that it is harmful [3][4][5][6][7][8], but that some pretreatment may make it safe [5][6][7]. Some PLA devices for culturing neurons have been proposed [9]; however, their effects have not been sufficiently investigated.
Polytetrafluoroethylene (PTFE) is famous for its trade name, which is Teflon. It is useful due to its resistance to chemical substances, insulating properties, and the lowest coefficient of friction among solids. PTFE has high biosafety [10]. PLA is one of the most commonly used materials for biomedical applications [11]. It is derived from plants such as corn starch and is said to be biodegradable [12,13] and biosafe [14]. It is attracting attention as a carbon-neutral material with attractive properties; it is renewable, biocompatible, processable, and energy-saving [15] and used as medical implants and drug carriers [16,18]. It is approved by the US Food and Drug Administration for human use for sutures, bone implants, and screws, as well as formulations for sustained drug delivery and vaccine antigens (proteins, peptides, and DNA) [19]. On the other hand, when it is used as a biomaterial, its strong hydrophobic property can lead to an inflammatory response [20]. In vitro, while it was reported that Chinese hamster ovary cells showed good proliferation and adhesion to PLA [21], it has been reported that PLA suppresses fibroblast adhesion, elongation, and alkaline phosphatase activity in the early stage of culturing [22].
In this study, isolated neuron dividers developed by Dr. Campenot [23] (the so-called Campenot chambers) were created using PLA and an FDM 3D printer, and the effect of the divider on nerve cells was controlled by Campenot chambers made of PTFE, which is the commonly used material for Campenot chambers [23][24][25].
Then, after that experiment, the pH of the medium due to contact with each material with a pH monitor (LAQUAtwin pH-22B; HORIBA Advanced Techno, Kyoto, Japan) and the amount of leakage of each type of chamber were measured using an absorptiometer (SmartSpec3000; Bio-Rad Laboratories Inc, California, USA).

3D printings
In this study, all PLA Campenot chambers were fabricated using the FDM 3D printer (Creator pro; Flashforge, Zhejiang, China).
In an FDM 3D printer, a filament that is easily molten in the heat was passed through a nozzle that was heated to a constant temperature to create tools. The filament is fed at a constant speed using an extruder. This caused the heat-molten resin to constantly exit the tip of the nozzle, which drew a layer with thickness programmed on the heated bed (Figure 1a). The heated bed prevented the shrinkage of PLA and made it easier to attach the modeled object by warming the adhesive surface.
By stacking the layers, the desired shape was obtained ( Figure 1b). Therefore, there must be a layer of scaffolding beneath the layer being drawn, which restricts the creation of tools with complex shapes. Therefore, when modeling an item with specific shape using CAD software, it is necessary to evaluate its formation. If not possible, it may be necessary to make plans, such as modeling support structures.
In this study, TinkerCAD (Autodesk, California, USA) was used to create the model and the stereolithography (STL) data. The STL data were translated using FlashPrint (Flashforge, Zhejiang, China) version 4.2.0 and saved with a filename extension of "x3" for 3D printing.
Two types of PLA filaments were used: white (Ender filament, Creality 3D, Shenzhen, China) and natural (PolyLite, Polymaker, Tokyo, Japan). Both filaments have a diameter of 1.75 mm. The conditions for the 3D printing are as follows: the temperature of the nozzle is 210°C and that of the heated bed is 60°C; the thickness of the first layer is 0.40 mm and that of the other layers is 0.20 mm; infill density of 100%; printing speed of 50 mm/s; the fan was run only during bottom printing; the nozzle diameter is 0.4 mm; and no auxiliary structures such as rafts, frames, or support materials were used.
In this experiment, PTFE and white PLA were first compared as white is the most basic color of PLA filaments, and then PTFE and natural PLA were compared in the next experiment; natural PLA has high purity [26].
The formed PLA chambers were washed with pure water three times, and before use, white PLA was autoclaved (121°C, 20 min) and natural PLA was UV sterilized (ultraviolet C [UVC], 15 W, 30 cm, 30 min).

Campenot chamber
A 6-well plate coated with poly-D-lysine/laminin was used for each culture. First, each well was scratched in parallel straight lines using a pin rake (Tyler Research, Alberta, Canada). These lines prevent the axon from extending in an unintended direction ( Figure 2) [23].
The circular compartments of the PTFE and PLA Campenot chambers were greased (High Vacuum Grease, Dow Corning Corporation, Michigan, USA) and placed so that the straight partition of the Campenot chambers crossed the straight lines drawn by the pin rake. Three wells were made of conventional PTFE (Teflon), and the remaining were PLA formed by the FDM 3D printer. The shape of the PLA chambers was modeled, as shown in Figure 3, to match the mass and shape of the contact surface to the medium. All the Campenot chambers, including the PTFE ones, weighed 2.7 g.
When cultured, axons pass under the left and right partitions from this central compartment to reach the left and right compartments, resulting in a high concentration of nerve growth factor (NGF).
In this study, for better adhesion of the chambers and reducing the contamination of cell bodies to the side compartments, 75 g of aluminum rods were placed as weights in each chamber for 2 h before the cells were added.

Animals
Male C57BL/6 mice aged 6-16 weeks were used in this study. The animals were maintained under a 12-h light/ dark cycle (daytime from 7 am to 7 pm) at a constant temperature. All the mice were euthanized with sevoflurane. Six chambers were allocated to one mouse, and three mice were used for each PLA group. Therefore, each group had nine PTFE chambers and nine PLA chambers.
Ethical approval: The research related to animal use has been complied with all the relevant national regulations and institutional policies for the care and use of animals and has been approved by the Animal Experimentation Ethics Committee of the Kyoto Prefectural University of Medicine.

Medium
Dulbecco's phosphate-buffered saline (DPBS) was supplemented with 1% penicillin-streptomycin. From now on, this will be referred to as "modified DPBS."  Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% bovine serum and 1% penicillinstreptomycin was used during cell isolation. From now on, this will be referred to as "modified DMEM." Neurobasal (NB) plus medium supplemented with 2% B27 supplement, 1% penicillin-streptomycin, and 0.5% lactose were used for the culture. From now on, this will be referred to as "modified NB plus medium." NGF, glial cell line-derived neurotrophic factor (GDNF), and arabinocytidine (Ara-C) were added to modified NB plus medium depending on the situation.
NB plus medium and DMEM contain phenol red.

Preparation of primary neuronal culture
Principal neurons were derived from the dorsal root ganglia (DRGs), which were dissociated by a mechanical procedure from C57BL/6 J mice. After euthanizing the mice, fresh DRGs were dissected and quickly transferred into ice-cold modified DPBS in a 10 mL falcon tube. After 3 min of centrifugation at 500×g and removing the supernatant, 3 mL of collagenase and dispase solution (collagenase 1 mg/mL, dispase 2.4 U/mL, dissolved in DPBS) was added and incubated at 37°C for 70 min. After 3 min of centrifugation at 500×g and washing with 5 mL modified DMEM, 1.9 mL of modified DMEM supplemented with 0.1 mL DNase (dissolved in DPBS) was added to the precipitate. The cells were isolated by pipetting 70 times, using a cell strainer, cell suspension, and 4 mL modified DMEM overlaid on 6 mL of 10% bovine albumin. After 10 min of centrifugation at 1,000×g and washing with 5 mL modified DMEM once, the precipitate was suspended in 900 μL modified NB plus medium (NGF 50 ng/mL, GDNF 2 ng/mL, Ara-C 10 μg/mL); 150 μL of the suspension was added to the center compartments of the six Campenot chambers, as shown in

Medium changes in Campenot chamber
On day 2, the medium of the center chamber was changed to modified NB plus medium (NGF 10 ng/mL, GDNF 2 ng/mL, and Ara-C 10 μg/mL); 250 μL of modified NB plus medium (NGF 50 ng/mL, GDNF 2 ng/mL, and Ara-C 10 μg/mL) was added to the compartments on both sides. On day 5, the medium of the center chamber was changed to modified NB plus medium (non-NGF, non-GDNF, and Ara-C 10 μg/mL), and the media on both sides were changed to the same modified NB plus medium (NGF 50ng/mL, GDNF 2 ng/mL, Ara-C 10 μg/mL). On day 7, each dish was immunostained and observed ( Figure 4).

Immunostaining
Nerve cells were immunostained for 7 days after the initiation of the cell culture. The following procedures were performed for each compartment in each chamber: First, the medium was aspirated and washed with 500 μL of DPBS, and 4% paraformaldehyde was added. After 5 min, the cells were washed with 500 μL of DPBS, and 0.1% Triton X-100 was added. After another 5 min, they were washed with 500 μL of DPBS, and Blocking One (Nacalai Tesque Inc.) was added. After incubation at 37°C for 30 min and aspirating Blocking One, an anti-beta To promote axon extension, the concentration of NGF in the center chamber was reduced by every medium change. GDNF in the center chamber was removed on day 5. The concentrations of the additives in both side compartments were kept at NGF 50 ng/mL and GDNF 2 ng/mL. Until day 2, both side compartments were kept empty to prevent neurons from flowing there.
Beta-tubulin was observed at a wavelength of 482 nm, and the cell body was observed at a wavelength of 390 nm using the EVOS FLoid Imaging System. These two images were merged to confirm axons and cell bodies in the central compartment (Figures 5a, 6b and c). When observing the extension of axons to the left and right compartments, an ordinal microscope view was added to check the absence of the cell body (Figures 5d and 6e-g).

Observation
The axons extending to the left and right compartments may have a mesh-like structure, as shown in Figure 6a, or extended, as shown in Figure 6b. As shown in Figure 6c, the degrees of extension to the left and right compartments were evaluated by adding all the lengths of the axons that did not have cell bodies in the left and right compartments. For each chamber, the sum of the axon lengths crossing to both side compartments was the subject of statistical analysis. The length was measured using ImageJ software.

Measuring pH and leakage
Each greased chamber (PTFE, white PLA, natural PLA) was placed on a scratched 35 mm Petri dish and a 75 g aluminum rod was placed on it for 2 h to make it stick. Eight chambers were prepared for each group. For each group of four chambers, 150 µL of modified NB plus medium (NGF 50 ng/mL, GDNF 2 ng/mL, and Ara-C 10 μg/mL) was added to the central compartment and 250 µL to the left and right compartments (pH group), and for the remaining 4,150 µL of pure water was added to the central compartment and 250 µL of modified NB plus medium (NGF 50 ng/mL, GDNF 2 ng/mL, and Ara-C 10 μg/mL) was added to the left and right compartments (leakage group). As a control, 1000 µL of modified NB plus medium (NGF 50 ng/mL, GDNF 2 ng/mL, and Ara-C 10 μg/mL) was placed on a 35 mm Petri dish (control group). Four dishes were prepared for the control group. All Petri dishes were incubated at 37°C.
After 72 h, the pH of the central compartment of each chamber was measured for the pH group and control group. For the leakage group, the absorbance of each compartment was measured using an absorptiometer and the absorbance at a wavelength of 560 nm was recorded [27]. Based on the measured values, the amount of leakage was estimated by dividing the average of the absorbance of the left and right compartments by the absorbance of the central compartment.

Statistical analysis
The extensions of the axons of the PTFE group and PLA groups were compared using the Wilcoxon signed-rank sum test. The Kruskal-Vallis test was used to compare pH and absorbance, and the Mann-Whitney U test with Holm's multiple comparisons, was used to compare the differences between the two groups. Two-sided tests were performed with a significance level of 95%. All the statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University) ver1.40, which is a graphical user interface for R. Figure 6: Measurement of axon extensions to both side compartments. The shape of the axon differs from axon to axon. Some axons are mesh-like (a) and others are long and straight (b). To evaluate the axon extension, the sum of the crossed-over axon lengths measured using ImageJ was used (c). On day 7, betatubulin was immunostained to observe the axon, and the cell body was stained by DAPI. Beta-tubulin was observed with GFP. The upper row is lined with images of the center compartment, which are merged with GFP and DAPI. The lower row is lined with images of the side compartment around the partition between the center and side compartments, which are merged with GFP, DAPI, and ordinal microscope view. The columns represent PTFE, white PLA, and natural PLA in that order.
In this study, Campenot chambers were used to confirm the effect of chambers made of PTFE and PLA using an FDM 3D printer based on the degree of nerve cell extension. Two types of PLA, white PLA made by Creality 3D and natural PLA made by Polymaker, were used.

Comparison of axon extension
Axon extensions to the left and right compartments were observed for the PTFE, white PLA, and natural PLA groups to a greater or lesser extent (Figure 7). A significant extension was observed in the PTFE group compared to the white PLA group (n = 9 in each group, p = 0.00781). The rank-sum test was used this time; however, the average value of axon extension per chamber was 7,306 μm for the PTFE group and 408 μm for the PLA group (Figure 8a), and four of the nine PLA chambers showed no extensions of axons to the left or right compartments.
The results of the Wilcoxon signed-rank-sum test were the same for the PTFE and natural PLA groups (n = 9 in each group, p = 0.00391), and the average values of axon extension per chamber were 21,184 and 5,733 μm, respectively (Figure 8b). One of the nine PLA chambers showed no extension of the axons to the left or right compartments.
For both groups, axon extensions to the left and right compartments were observed in all PTFE chambers, although there was variability.

Comparison of pH
Modified NB plus medium was added to each type of chamber and the pH values were compared. The pH in each group was compared as shown in Figure 9. The comparison of the four groups including the control group was p = 0.115, and the comparison of the three groups excluding the control group was p = 0.347, showed no significant difference.

Comparison of leakage
Pure water was added to the central compartment and modified NB plus medium containing phenol red was added to the side compartment, and the absorbance

Discussion
In general, white is the basic color of PLA filaments and the color of PLA is transparent; therefore, natural-colored filaments have higher purity and low crystallinity when they are formed by an FDM 3D printer [26]. We attempted to use both types of PLA. Because the extension of axons differs from cell to cell, the Wilcoxon signed-rank sum test was used to compare the results of the neurons obtained from the same mouse.
The conventional PTFE chambers showed greater axon elongation than the PLA chambers created by the FDM 3D printer. The longer elongation is expected to be due to an increase in the number of axons that elongate after responding well to NGF [28].
Recently, culturing with PLA nanosheets has often been conducted [29], and the high hydrophobic property of PLA requires pretreatment for better cell adhesion [30][31][32]; however, PLA is thought to be biosafe for neurons. In this study, PLA was not pretreated because it was not a scaffold, and the hydrophobic property of PLA may have affected cell extension even though it was not used as a scaffold.
PLA is known for its accelerated degradation by hydrothermal reactions and its degradation by prolonged UV exposure, resulting in organic acids. Therefore, there is a possibility that the sterilization process in this experiment may have been harmful to the cells. PLA is degraded to produce the monomer lactic acid, which contains several organic acids and other compounds [33]. In order to confirm the effect of this, we investigated the pH. However, the difference was not obvious, and no effect of sterilization on neuronal growth was suggested.
3D printers create structures by laminating molten filaments, and there may be slight irregularities on the surface of the PLA chambers created by the 3D printer. PLA and PTFE have different physical properties, such as friction coefficient and surface roughness [34][35][36][37]. The coefficient of friction of PTFE is very small [38] and its surface is smooth [39]. On the other hand, the surface of PLA formed by the 3D printer has a rough structure at the microscopic level [40]. PLA is known to have slight distortions due to shrinkage during the formation process. In this study, more leakage was observed in the PLA group compared to the PTFE group. This result suggests that the rough surface or shrinkage of the PLA formed by the 3D printer has a weaker ability to partition the solution compared to PTFE, which may have prevented the concentration difference in the Campenot chamber from working effectively. If that is the cause, then frequent medium changes may give better results. However, the toxicity of PLA itself cannot be denied from this study. Further research is needed.

Conclusions
When comparing PTFE and 3D-printed PLA Campenot chambers, it was found that better neurite outgrowth was obtained with PTFE. The reason for this may be due to the smoothness of the PTFE surface.
However, this study was only validated with a specific instrument shape, and only limited conclusions can be drawn. More research is needed.