Extrusion bioprinting device for planar and non-planar mammalian cells and bacterial scaffolds prototyping

Bioprinting has emerged as an extension of material printing using living material, including microorganisms, mammalian cells and active molecules (Staton et al., 2015). Bacterial bioprinting is focused on developing multifunctional bioinks for the creation of living and responsive materials, combining 3D printed complex shapes with dynamic functionalities resulting from the products of bacterial metabolism (Kyle, 2018). Conversely, mammalian cell bioprinting implies fabricating intricate 3D scaffolds mimicking the characteristics of the extracellular matrix. This involves creating structures with interconnected porosity and specific fibre patterning to facilitate nutrient transport and support cell survival, attachment and subsequent tissue functionalization (Shao et al., 2018; Sonatkar et al., 2022). Various manufacturing methods have been explored to achieve these porous scaffolds, including freeze-drying, mold casting, stereolithography and 3D extrusion bioprinting (Flores-Jimenez et al., 2023). Among these techniques, extrusion bio-printing is the most extensively investigated due to its ease of using biocompatible materials, referred to as bioinks, primarily based on hydrogels. Structures based on hydrogel are built additively for extrusion devices, using a layer-by-layer methodology in planar trajectories. Generally, two main approaches are considered to manipulate the living material: encapsulating bacteria or cells embedded in the bioprintable hydrogel or culturing them over printed hydrogel scaffolds. These hydrogel formulations allow cell encapsulation and control of their precise placement; however, it should present specific characteristics to make it extrudable, which often depends on factors such as extrusion pressure, temperature and nozzle geometry (Chen, 2019). To determine these parameters, a rheological characterization is needed, as presented in (Gloria et al., 2020), where the viscosity of the hydrogel, as the resistance to flow, and its storage modulus, as its viscoelastic behaviour, were determined by an injectability test. In addition, the controlled variations of these parameters can also influence cell growth and viability (Flores-Jimenez et al., 2023). In general, bioinks often presents low viscosity and are dependent on crosslinking agents to consolidate a structure, which limit the resolution of bioprinted constructs. Consequently, this hampers the potential to generate scaffolds with intricate 3D curvatures or self-standing components (Chen, 2019). To overcome these challenges, researchers have explored support baths and sacrificial materials as potential solutions (McCormack et al., 2020). These strategies aim to prevent the main structure from collapsing during deposition. Nevertheless, challenges persist, particularly concerning potential damage to the structure during support removal and limitations in incorporating crosslinking factors (Pagan et al., 2023).

The well-known 3D extrusion bioprinting is constrained by its planar material deposition path adherence. This method involves stacking material layer by layer on a flat surface, resulting in a structure that lacks smooth transitions between non-continuous regions. The resolution for curves or vertical details is contingent upon approximations dictated by the number of deposited layers. Due to the intricate nature of tissues, planar 3D bioprinting faces limitations in fabricating clinically relevant tissues or in vitro models. This drawback also impedes the realization of the full potential of techniques such as in situ bioprinting and support bath bioprinting, where omnidirectional extrusion is essential (Ding et al., 2015). Furthermore, there is a lack of control over curvature resolution, a characteristic that can influence cell fate at various scales, from organelles to tissues (Schamberger et al., 2023). A few studies have explored alternatives to the non-planar printing process, particularly following curved deposition trajectories (Liao et al., 2023,Wulle et al., 2022). In Ahlers et al. (2019), an algorithm is proposed to guide the extruder of a conventional fused deposition 3D printer along the contour of a non-planar surface. The approach involves initially printing planar layers to form a base. Subsequently, these layers are 2D projected to identify the non-planar path and a z-offset is added. A similar methodology is outlined in Alkadi et al. (2020), where non-planar infill paths are divided into segments to circumvent collisions, enabling printing on a freeform substrate. Another non-planar algorithm is presented in Rodriguez-Padilla et al. (2021), where 3D lattices are printed on a curved substrate based on the calculation of normal vectors of the substrate and their intersections with lattice points. In Houzhu (2019), a method is proposed to create non-planar paths from any solid file using image processing and a Kinect sensor. Although non-planar extrusion has been minimally explored in the realm of 3D bioprinting, notable examples exist in Lee et al. (2009). Authors deposited hydrogel fibres on a curved substrate, using a non-continuous sequence of planar layers rather than a non-planar deposition path. A more sophisticated approach is detailed in Fortunato et al. (2021), where a bioprinting platform based on a robotic arm is designed to deposit material on irregular substrates, particularly for in situ bioprinting. This platform operates by following an algorithm calculating joint angles to ensure the extruder remains perpendicular to the surface. Another noteworthy development is the non-planar slicer for bioprinting alginate-based hydrogel within a gelatin support bath, as described in Albert et al. (2022).

To enhance the guidance of cellular growth in a non-planar structure, we introduced a technological methodology to develop bioprinting instrumentation, departing from printer adaptation to follow curve trajectories for embedded living material deposition. The first main contribution of the work is the conditioning of a conventional 3D printer to an extrusion 3D bioprinter by modifying the extruder and the firmware, proposing a new design methodology and g-code generation for porous scaffolds. The second key contribution of the study consisted in the validation of the 3D bioprinter using designed bioinks with living materials (bacteria and mammalian cells, respectively), confirming their growth and viability. This manuscript is organized as follows: Section 2 details materials and methods used to design and condition a low-cost commercial bioprinting device, including extruder design for bioprinting purposes, the process for non-planar scaffold design and the formulation and characterization of hydrogel and bioinks used to assess the platform’s functionality, with one incorporating bacteria and the other mammalian cells. Section 3 presents the bioprinting assays, with the first involving planar trajectories using the bacteria bioink. Following this test, a mammalian cells bioink was used in planar and non-planar trajectory bioprinting, using a rigid non-planar structure as support. Section 4 closes the study with some final remarks and future trends issued from this study.

To create the 3D bioprinter device, a commercial desktop fused deposition modelling 3D printer platform was modified, the Wanhao Duplicator i3, shown in Figure 1(a). A similar approach has been taken by Senior et al. (2019), but dedicated to simple planar extrusion. The Wanhao Duplicator i3 had its original Melzi motherboard, which was changed by a motherboard SKR V1.3 from Bightreetech. This change represents a transition from an 8-bit architecture to a 32-bit one, which increases the processing capacity of the g-code and accurate movement of the motors (X, Y and Z coordinates). A custom version of the open-source firmware, Marlin (V2.0.5.3), was downloaded and flashed to the motherboard to enable bioprinting applications. Modifications were made to optimize the printer’s capabilities and set up dimensions for bioprinting. In addition, a new extruder for hydrogel extrusion was designed. A digital display, slots for SD cards and a USB drive were added. Dimensions of the new bioprinter allow it to be used in laminar flow hoods, a necessary condition for cell culture in a sterile environment and the current configuration ensures plate temperature control (max 40°C) by adjusting gains in the internal PID control. Motors in the bioprinter were configured for linear advance according to the protocol and recommendations in (Sebastianv650 and Sineos, 2025).

The original printer head was modified to extrude bioinks based on hydrogel formulations and control their extrusion volume; we considered the use of a single cartridge (bioink reservoir) using commercial sterile syringes (3 mL capacity, internal diameter of 8.4 mm) for this purpose with available extrusion tips from the supplier Cellink®, based on this dimension we designed the holder and the mechanism to push the cartridge’s plunger, as is represented in Figure 1(b) and (c). The new bioink extruder structure was designed using SolidWorks® software and 3D printer in ABS material for the first design. After testing, the final design was manufactured using resin material in a stereolithography 3D printer (Formlabs, Model Form 2). Extrusion is a non-linear process, but motors in the bioprinter were configured for a linear advance. A calibration process was conducted for the step motor in charge of pushing the plunger, ensuring the correct volume selection of the extruded material correlated with the number of motor steps. Water was used as a testing material, filling our cartridge (3 mL sterile syringes) with 1 mL at every repetition. We determined the correct linear distance (baseline) to match the volume extrusion of 1 mL, corresponding to the number of steps to extrude this control volume accurately.

In this work, we proposed using Rhinoceros 3D software (Version 6) with its integrated visual programming tool Grasshopper to obtain the G-code of a scaffold. The scaffold was modelled in a planar a non-planar layer by layer. The planar scaffold was designed to have a final length of 15 mm, width of 25 mm and height of 0.75 mm (X, Y, and Z dimensions, respectively) with square pores of 750 µm by the side as is presented in Figure 2(a). The non-planar scaffold preserves the square pores and dimensions. Still, it is curved in two sections, with peaks and valleys with a 15° degree of inclination considering the XY plane [Figure 2(b)], resulting in a porous non-planar structure as shown in Figure 2(c). In Rhino 6, a mesh with the size and curved geometry desired was modelled and then copied twice with a Z offset of 250 µm from each mesh, Figure 2(d). This approach to modelling scaffolds offers advantages for structures comprising a limited number of layers, especially those with inner holes or irregular shapes within specific layers. After modelling the scaffold, the layers were linked to Grasshopper and converted to curves and meshes, which were transformed into paths in XYZ coordinates. The paths were then converted to numbers and concatenated in conjunction with a start gcode and an end gcode that have the initial and final instructions so the bioprinter can work properly.

Before testing the bioprinting capacity of the device with embedded living material, we performed a 3D printing assessment. We evaluated the extrusion resolution and the capability to follow the designed planar and non-planar trajectories of the conditioned 3D printer. In this process, we formulated hydrogel without living material and a rigid support for the non-planar trajectories. We quantitatively evaluated the printed scaffolds in terms of their ability to retain the designed curved geometry after being detached from the rigid support.

We formulated a hydrogel containing 5% (w/v) sodium alginate (Sigma-Aldrich) and 5% (w/v) gelatin (porcine skin type A, 90–100 g bloom; Sigma-Aldrich), a typical formulation for bioinks (Webb and Doyle, 2017; Yenilmez et al., 2019; Faulkner-Jones et al., 2015; Gao et al., 2025) and deeply studied by Abedi et al. (2024). After hydrogel formulation, its viscosity µ (cP) was determined using a Brookfield RVDVE viscometer by triplicate in a shear rate range $γ˙$ of 2.09–20.9 s⁻¹ at 37°C. The obtained viscosity was used to calculate the volumetric flow rate Q (mL/s) as in equation (1) (Chen, 2019):

where R_t (mm) is the inner radius of the tip, L_t (mm) is the length of the tip, n (dimensionless) is the power law index, K (Pa·sⁿ) is the power law consistency index and P (kPa) is the extrusion pressure, which was set as 5 kPa, considering that it should be less than 10 kPa to avoid damage of living materials in the subsequent bioprinting (Blaeser et al., 2016). To prove that the applied pressure to extrude the hydrogel was below 10 kPa, an injectability evaluation (Gloria et al., 2020) was conducted. For this, a universal testing machine (Univert CellScale, 200 N) was used. A 3 mL syringe, with a 25 G plastic tapered nozzle was filled with hydrogel and fixed at the testing machine, where the plunger was compressed under the displacement control mode. The resultant forces were recorded to identify the pressure at which continuous cylindrical strands were obtained. Finally, the storage modulus (G’) was obtained based on the methodology presented in (Campbell et al., 2021).

Finally, the printing velocity v_p (mm/s) was computed using equation (2):

where D (mm) is the desired filament diameter, which in theory should be equal to the tip diameter. The printing velocity was then incorporated into the g-code.

After obtaining the printing parameters, two planar extrusion tests were carried out by depositing the hydrogel ink at 37°C over a planar substrate (plastic Petri dish, diameter 60 mm) using a 25 G plastic tapered nozzle and a 3 mL syringe. The first test, to prove the printer’s ability to extrude continuous cylindrical filaments, consisted of following the trajectory depicted in Figure 3(f). Once the extrusion was completed, the hydrogel was ionically crosslinked by depositing droplets of 300 mM CaCl₂ (Li et al., 2016), covering the entire surface and leaving it for two minutes. The resolution of the filaments was evaluated in ImageJ by measuring the filament diameter and calculating its error with the theoretical one (inner diameter of the nozzle) and by proving that no discontinuities were present.

The second extrusion test consisted of printing the previously designed square porous trajectory [Figure 2(a)], with the aim of evaluating the printer resolution. As in the first test, the extruded hydrogel was ionically crosslinked. The pore resolution was evaluated using the printability ratio P_r (dimensionless), as given in equation (3) (Zhang et al., 2018):

where L (mm) is the pore perimeter and A (mm²) is the pore area. The closer P_r is to 1, the better resolution the scaffold has. To calculate P_r, an image of the printed scaffold was taken and analysed in the ImageJ software, where the perimeter and area of all pores (180) were obtained and averaged.

Non-planar extrusion test was performed to assess the ability of the conditioned printer device to follow continuous curved trajectories at different scales without collision. The previously formulated hydrogel and the same printing parameters as in the planar tests were used.

Curved solid templates were fabricated to serve as support during the non-planar deposition of the hydrogel, maintaining the curved shape before the crosslinking. Firstly, a support template (Template 1) was designed following the same curved trajectory as the previously designed porous curved scaffold [Figure 2(b)], with the same dimensions in x − y (15 × 25 mm), but without the pores, as shown in Figure 3(a). Template 1 was scaled in the x − y dimensions, resulting in Template 2: 6 × 10 mm and Template 3: 4.5 × 2.5 mm, as shown in Figure 3(c)–(e). These templates allowed us to test different ratios of curvature; Template 1: 0.2919 µm⁻¹, Template 2: 0.5047 µm⁻¹ and Template 3: 2.0188 µm⁻¹. The three templates were fabricated by stereolithography using the Form2 printer (Formlabs, Black Resin).

The hydrogel was extruded following the non-planar trajectory of the curved porous scaffold, which was scaled accordingly to each template. After the complete hydrogel deposition, the printed scaffolds were ionically crosslinking following the same procedure as for the planar extrusion tests.

To ascertain whether the scaffold retained its curvature once crosslinked and removed from the template, the average curvature κ of the superior lateral edge of each template was calculated using equation (4). This theoretical curvature was derived from a curve within the plane, defined as a relationship between rectangular coordinates expressed as y = f (x), i.e.:

We calculated κ by numerical approximation of the derivative using function diff in Matlab (v2020a, MathWorks), considering a set of 17 pairs of points x, y from the CAD template file to form the 2D curve, as observed in Figure 3(b). To estimate the curvature on the printed scaffolds and to compare it with the theoretical one, a b-spline curve was fitted on the superior edge of the scaffold using the Kappa-curvature analysis toolbox of the image analysis software Fiji (Schindelin et al., 2012).

After the 3D printing capacity assessment, we assessed the 3D bioprinting capability of our device. Three bio-printing conditions were considered: a planar scaffold using a bacteria bioink formulation and a planar and non-planar scaffold using a mammalian cells bioink formulation. For the condition of the non-planar scaffold, we performed a deeper analysis of the effect of different curvatures on cell viability as this is a current challenge to constructing bioprinted tissue structures (Flores-Jimenez et al., 2023; Kainz et al., 2024).

The bacteria bioink preparation was adapted from Ning et al. (2019). Autoclaved bacteriological agar (6 g/L) and trypticase soy (TS) medium (50 g/L) were combined. Then a 6% (w/v) alginate solution, previously treated with UV light, was dissolved into the prepared agar/trypticase medium; we obtained a hydrogel whose viscosity was determined using a Brookfield viscometer (DV2TRV) at room temperature (23 ± 1°C) and a range of rotation speeds (12, 60, 80 and 100 rpm) employing spindles 0.4, 0.5 and 0.6 (Espinosa et al., 2019). This characterization was performed by triplicate before embedding the bacteria. As a living material, we used Lactococcus lactis (ATCC 19435), which was cultured in TS broth medium at 37°C with continuous shaking (100 rpm). Subsequently, 200 µL (1 × 10⁶ CFUs/mL) of this culture was transferred to a fresh 30 mL TS medium and maintained at 37°C. After 8 h of incubation, biomass from the bacterial culture was collected by centrifugation (10,000 rpm, 4°C, 5 min); optical density spectrophotometry was used to measure the bacterial concentration (λ = 595 nm) (Fusieger et al., 2022). We added bacteria culture into the prepared TS-alginate ink to obtain the bioink formulation; this bioink was used exclusively in planar bioprinting testing.

To prepare the mammalian cell bioink, the same hydrogel previously formulated and characterized was used (5% alginate – 5% gelatin) and sterilized for subsequent mammalian cell incorporation. For this, the UV-sterilized alginate and gelatin powders were dissolved in a sterile cell culture medium (DMEM F12, Thermofisher) and the obtained solution was filtrated with a 200 µm filter (Lorson et al., 2020).

Human hepatocellular carcinoma cells (HEPG2, ATCC HB-8065) were cultured in DMEM medium F12, supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich), according to the ATCC, and 1% of penicillin-streptomycin (ThermoFisher). Cells were incubated at 37°C in a humidified atmosphere with 5% CO₂, the media was changed every other day until an 80% of confluent was reached. To subculture cells, a 0.05% Trypsin-EDTA solution (ThermoFisher) was used, and cells at passage eight were used for all 3D bioprinting experiments. A total of 33 million cells were sus-pended in 300 µL of supplemented cell media and then slowly transferred to a syringe containing 3 mL of the previously prepared alginate-gelatin bioink, resulting in a concentration of 10 million cells per mL, following the protocol outlined in (VKJB, 2021). This bioink was used for both planar and non-planar bioprinting tests. As a growth and viability control for mammalian cell experiments, cell cultures (2D) were performed on microscope cover glass slides pre-treated with FBS and incubated for 1 h at 37°C in a humidified atmosphere with 5% CO₂ to facilitate cell attachment. HEPG2 cells were seeded onto the slide surface at a density of 50 × 10³/cm². Cell media was changed every day.

After the formulation of bioinks, we configured the bioprinter to extrude bioinks following the designed planar and non-planar trajectories described in Section 2.2 and using the support templates as in Section 2.3.3. In every experiment, the bioprinter was disinfected using 70% ethanol and UV irradiation for 2 h. The subsequent bioprinting procedures were carried out inside a BSL 2 Bio-safety cabinet. The cartridge (syringe) was filled with the prepared bioink in aseptic conditions and installed in the modified extruder. Ionic crosslinking occurred after completing each extrusion trajectory by drop deposition of 300 mM CaCl₂, covering the printed structures for 2 min. After crosslinking, all scaffolds were removed from the substrates (Petri Dish for planar and support templates for non-planar), immersed in complete cell media and incubated at 37°C in a humidified atmosphere with 5% CO₂.

After the biopriniting procedure, L. lactis viability within the printed scaffold was evaluated by fluorescence staining using the LIVE/DEAD™ BacLight™ Bacteria Viability Kit (Invitrogen), as reported in (VKJB, 2021). The staining process involved simultaneously incubating the printed bacteria scaffolds with SYTO-9 and propidium iodide at room temperature for 15 min. High-resolution imaging of bacterial scaffolds was conducted through confocal laser scanning microscopy (CLSM) using the LSM 800 system (Zeiss, Germany) and the image capture software Zen 2.3 system. Specific wavelengths were used for fluorescence excitation and emission. SYTO-9 fluorescence was induced at 488 nm, with subsequent detection limited to wavelengths exceeding 550 nm. Similarly, propidium iodide was excited at 555 nm, followed by emission detection beyond 550 nm (Rodriguez-Padilla et al., 2021). To assess the growth of L. lactis within 3D printed scaffolds, the fabricated structures were incubated and retrieved at 0 and 8 h. A 40× objective was used with a scan area of 638.9 × 638.9 um and a pixel size of 1.25 um. The printed scaffolds were visualized using a z-study with a slice thickness of 7 µm.

After the bioprinting procedure, a live/dead assay was performed as reported in Lee et al. (2023) staining solution of 2 µM Calcein AM and 4 µM Ethidium homodimer-1 was prepared in 10 mL of phosphate-buffered saline (PBS) according to the manufacturer protocol (Invitrogen). Scaffolds were rinsed three times with PBS to remove the cell culture media and then submerged in the live/dead solution for 30 min at 37°C before imaging. The same procedure was applied to the 2D control glass slides. Scaffolds and 2D controls were visualized at five time points: 24 h, 48 h, 72 h, 96 h and 120 h, using the Zeiss Confocal Micro – scope LSM 800 and the image capture software Zen 2.3 system, using the 20× objective and a scan area of 638.9 × 638.9 µm. A z-stack was captured with a total thickness of 440 µm and slide thickness of 7 µm. The non-planar bioprinted scaffolds were visualized at different curvature sections across the geometry, as shown in Figure 3(d) for the ones printed with Template1 and Template2, and in the marked sections in Figure 3(e) for Template3. While for the planar bioprinted trajectories, the points specified in Figure 3(f) were visualized. For all cases, the percentage of cell viability was quantified using the ImageJ software.

The design of the new extruder is shown in Figure 4(a), whereas in Figure 4(b), we show the manufactured extruder by 3D printing using resin and installed into the bioprinter. The new printer head uses an 8 mm ACME thread to transfer the movement to the bioprinting cartridge. This design allows the nozzle plunger to move up and down so the extrusion system can deposit the material and do retractions. The carriage was designed for modularity, so the part that locks the bioink cartridge in place has three different working volumes: 3 mL, 5 mL and 10 mL. In addition, an end-stop switch is included; thus, when the extruder reaches the end of the path, it stops, preventing possible accidents to the machine or the cartridge.

The bioprinter render is shown in Figure 5(a), including the designed extruder. Figure 5(b) shows the assembled bioprinter with a 3D printed extruder as is shown in Figure 4(b). Figure 5(a) shows the new interface with display, SD card and USB flash drive to load the files to be bioprinted. To provide thermal regulation during deposition, the original heated bed was kept. The current use of the bioprinter inside the biological safety cabinet is shown in Figure 5(c). The motherboard controlling the bioprinter was installed below the printing bed, attached to a 3D-printed ABS bracket. All the electronic components were allocated to be coupled to the machine’s structure with 3D-printed ABS parts.

Once the Marlin Firmware V2.0.5.3, was downloaded and modified in Visual Studio Code with the Platform IO extension, it was customized according to the modified printer’s characteristics, including bed shape, extruder and linear advance. The thermistor for the printing bed was selected to maintain the thermal regulation within a temperature range of 5°C-50°C, with a PID controller tuning performed after the new extruder was assembled into the bioprinter. Motors X, Y and Z were calibrated to characterize step/mm in each direction. Bioprinter dimensions limits for X, Y and Z were set to 215.5 mm, 200 mm and 55 mm, respectively. Options of sensor-less homing, step interpolation as well as motor’s fine-tune movement capability (0.1 µm) were configured. After the calibration protocol, the motor steps for a 3 mL cartridge were determined to be 680 steps and were added to the firmware. The configuration was saved as a base for further adjustments. The extruder backlash was reduced to 200 µL and compensated in the firmware.

In Figure 6, we presented a comparative non-planar scaffold design with the software Slic3r, in which the slicing process produced staggered transitions in the non-planar parts, see Figure 6(a), and with the proposed slicing process using Rhinos/Grasshopper obtaining continuous deposition paths, as shown in Figure 6(b). This approach improves resolution and avoids material waste due to path interruptions and crossings.

After the rheological characterization, the 5% alginate – 5% gelatin hydrogel viscosity (µ) resulted in an average of 2,200 cP at 37°C, which is a value within the range of viscosity reported for alginate-based inks used for extrusion bioprinting (Tirella et al., 2009; Guillotin et al., 2010). This viscosity value resulted in the power law index (n) of 0.787 and the consistency index K of 18.66 (Pa · sⁿ). With an extrusion pressure, obtained from the injectability test, of 5 kPa, the volumetric flow rate (Q) calculated with equation (1), resulted in 0.87 mL/s. Therefore, the printing velocity (v_p), calculated with equation (2) resulted in 17.8 mm/s, which was introduced in the g-code. For its part, the storage modulus G’ resulted in 53 kPa for the 5% alginate – 5% gelatin ink, and in 64 kPa for the 6% alginate – agar ink.

After performing the first planar extrusion test, by depositing continuous hydrogel filaments and crosslinking them, it was obtained an error of 2.8% compared with the expected filament diameter (0.5 mm) and no discontinuities were observed, as shown in Figure 7(a). For the case of the second planar extrusion test, a rectangular scaffold with square porous was obtained, as shown in Figure 7(b). The resolution of the square pores was evaluated with equation (3), resulting in a printability (P_r) value of 0.88.

The non-planar extrusion tests for the curved square-pore scaffolds were performed using the support templates. Figure 7(c) shows an example of a non-planar printed scaffold over the Template1 before the crosslinking process, whereas Figure 7(d), shows an example of the non-planar scaffold after crosslinking and removed from Template1, where it can be observed that the non-planar shape was kept. The average experimental curvature radio was calculated using equation (4), by analysing the curve formed at the top of the scaffold as shown in Figure 7(d), which resulted in 0.2542 µm⁻¹. The same procedure was performed for the hydrogel extruded over each scaled support template (Template 2, 3), resulting in the curvature radios of 0.3157 and 0.24193 µm⁻¹, respectively. The errors of the crosslinked hydrogel scaffolds, compared with the theoretical curvature of each rigid support template, after detachment were 12.91%, 37.44% and 88.015%. These values depend entirely on the crosslinked structure properties; further analysis can be performed to study and optimize the consistency of these semi-solid structures and their implications over the embedded living material response.

The bioink composed by L. lactis biomass, trypticase soy (TS) medium (50 g/L) and 6% (w/v) alginate solution was successfully bioprinted. Figure 8(a) shows a confocal microscopy analysis demonstrating the initial and sustained viability of L. lactis within the planar scaffold at 0 h, and in Figure 8(b) at 8 h after bioprinting process, in both cases a green fluorescence is indicative of live bacteria in the bioprinted matrix, dead bacteria, identified by red colour were not appreciated at this time. This observation suggests that our scaffolds support bacterial survival and facilitate their proliferation, being a potential platform for studying bacterial consortia in semi-solid phases. Further work will be performed in this area.

The cell viability percentages, resulting after following the live/dead staining protocol and the imaging points previously described in Figure 3(e) and (f), are presented in Figure 9. As can be observed in Figure 9(a), cell viability was maintained at high levels for the first two days and was at an average of 71.8% on day 3 (72 h) in all sections. Comparing the non-planar, 3D planar and 2D cell culture strategies, the latest holds the highest values in cell viability. However, non-planar strategies retain the same cell viability as 3D planar bioprints, see Figure 9(a).

Differences in cell clustering definition were observed between 2D controls and cells encapsulated in the bioink, as shown in Figure 10, an expected result since the hydrogel matrix in 3D restricts cell growth and proliferation. In Figure 11, samples from different curvature sections of Template 2 and 3 are presented. It is appreciated that cell viability and distribution are similar across the entire scaffold, proving that the effect of curvature is minimal at these scales. This platform can be used for further experiments to test other curve geometries.

We successfully created a 3D bioprinter conditioning a 3D printer and designing and manufacturing an extruder for bioinks materials based on hydrogels. Moreover, by upgrading its hardware and firmware, the constructed bioprinter could follow extrusion paths in the XY Z plane, making it feasible to bioprint planar and non-planar structures, following curved trajectories for bioink deposition, with a maximum of 15° degree of inclination. It was also proven that the conditioned 3D bioprinter device was capable of extruding non-planar trajectories at different scales, in the curvature radius range of 0.29–2.02 µm⁻¹. Printability of hydrogel, considering its viscoelastic properties and non-Newtonian flow behaviour, using this bioprinter was confirmed, as well as the viability of living materials (bacteria and mammalian cells) in planar and non-planar porous scaffolds after bioprinting. Moreover, it was proven that the non-planar scaffolds preserve their geometry after being bioprinted, crosslinked and removed from the printer substrate. This device and methodology can be applied to study different mechanobiology effects and an alternative for complex curve geometries in tissue engineering.

This work has received funding from Tecnologico de Monterrey, Institute of Advanced Materials for Sustainable Manufacturing under the grant Challenge-Based Research Funding Program 2022, Grant no. I006-IAMSM004-C4-T2-T.