**1. Introduction**

The rise of three-dimensional (3D) printing in the last three decades has permitted the arrival of a new manufacturing technology called 3D bioprinting for organ and tissue engineering (TE) [1–4]. This technology aims to deposit multiple biomaterials, growth factors, and living cells with precise control over their compositions, spatial distribution and architecture [5]. Since the appearance of the first bioprinting studies in 2003 introduced by Wilson and Boland [6], the field has experienced a growing interest by the scientific community in the last decade. The rapid increase in the number of related publications provides evidence of this tendency (Figure 1).

Today, allograft organ transplantation is still the only therapy effective against organ failures, but relatively simple implantable tissue constructs have been printed and successfully transplanted into animal models [7]. These works bring great hope for those patients who are looking for alternatives to the organ transplantation methods. Looking forward, the challenge remains of how to reproduce the complex cellular organization and micro-environment of an entire solid organ. This is still well beyond the capabilities of currently available bioprinting technologies [8].

**Figure 1.** Evolution of bioprinting and 3D printing publications measured by the number of Google Scholar hits. The keywords utilized for the searching included: "3D printing" OR "additive manufacturing", "3D printing" and "tissue engineering" and "bioprinting".

Three main bioprinting technologies have been developed: inkjet-, laser- and microextrusion-based bioprinting (MEBB) systems. Each of them has been utilized in several biological applications, offering different features in terms of cell viability, deposition speed, print resolution, scalability, cost or materials compatibility [9]. MEBB is the most extended technology because of its versatility and fast deposition of a wide range of bioinks, which enables the rapid generation of large-scale constructs [10–12]. Although excessive printing pressures could reduce cell viability, it is an excellent method for depositing high cell densities in several candidate bioinks [13]. The bioinks can be defined as formulation of biomaterials, biological molecules and cells processed using bioprinting technologies [14,15]. Most of the studies in 3D bioprinting have traditionally been limited to the use of one or two bioinks at one time, which is perhaps an oversimplification that limits the structural, material and biological potential of this technology [16].

Employing multiple building and sacrificial biomaterials and cells types in a single biofabrication session seems to be the right way of addressing the complexity of organ engineering and producing outstanding advances in the field [17–19]. Multi-material bioprinters have recently been developed by several research groups [7,11,12,20–22]. These bioprinting systems normally incorporate up to three or four printheads to perform multi-material extrusion like the open-source solution utilized by the authors in this study. To the best of our knowledge, advances in multi-material bioprinting will enable researchers to integrate intricate perfusable channels inside of complex shape constructs, and create constructs with several different cell densities, among other advantages. A more detailed study in multi-material bioprinting [8], using stem-cell-laden bioinks, alongside a network of reinforcing poly(*ε*-caprolactone) (PCL), led to the biofabrication of so-called developmentally inspired templates of bone tissue microfibers.

All of this cannot be accomplished without answering fundamental questions such as the ideal properties of the bioinks and the relationships between the bioprinting process parameters and the print resolution and fidelity [13]. In the case of MEBB, some previous research studies have correlated bioprinting parameters and printed outcomes. Wang et al. showed that optimized printing parameters such as bioink concentration, nozzle speed and extrusion rate produced poly(lactic-*co*-glycolic acid) (PLGA) scaffolds [23]. Mixtures of Gel-Alg were investigated by He and his coworkers to find the optimal values of air pressure, feed rate, and layer height to assure proper printing quality [13]. Suntornnond et al. used poloxamers to develop a mathematical model to correlate print resolution with process parameters [24]. Similarly, a prediction model was obtained by Trachtenberg et al. to determine the suitability of poly-propylene fumarate for MEBB [25] while Ting et al. examined the effect of PLGA composition and printing parameter on print resolution [26]. However, today, there is no a definite method to calibrate multi-material 3D bioprinters as well as to determine their final print resolution. Understanding how parameters such as printing speed and nozzle height affect the print resolution is vital not only for the shape of the printed constructs but also for their mechanical properties. When encapsulating cells, selecting the optimal printing parameters will reduce the adverse effect of the viscoelastic stresses on the cell viability [27,28].

In this paper, we advance in the development of the multi-material 3D bioprinting by proposing a method that analyzes the influence of the main printing parameters and accurately controls the print resolution. We anticipate that a significant increase on printing speed and quality of the constructs using the multi-material bioprinter is due to the use of an automatic calibration system. Poloxamer 407 (P407) hydrogels with different fluorescent inks were printed into different complex constructs for finding the optimal printing parameters. This allowed us to emulate the bioprinting of four materials, but, at the same time, also remove other secondary factors such as excessive swelling or temperature dependence. The proposed method was also tested printing a mixture of gelatin-alginate (Gel-Alg), a more cell-friendly bioink. Cell-laden Gel-Alg and P407-based bioinks were printed in a single session. After printing, cell viability of stem cells embedded in the Gel-Alg was analyzed to verify the effects of the calibration. The results demonstrated that our proposal has huge potential to help in creating large multi-material 3D constructs and potential vascular networks for tissue engineering.

### **2. Materials and Methods**

### *2.1. Bioprinting System Incorporating Four Printheads*

The experiments were performed using a desktop open-source 3D printer Witbox 2 (BQ, Madrid, Spain) modified for extruding hydrogels at 24 ◦C (Figure 2a). The mechanical resolution of the 3D printer is up to 20 μm according to the manufacturer's specifications. The Witbox 2 movements follow a Cartesian dimensional coordinate system, in which the printheads are moving across the *xy* horizontal plane while the printing platform only moves vertically (*z*-axis). The Witbox 2 was modified by substituting the standard fused deposition modeling nozzle in the *x*-carriage for four pneumatic-based MEBB printheads (Figure 2b,c). The four printheads' movements are controlled using open-source Rumba electronics (Reprap Universal Mega Board with Allegro driver; RepRapDiscount, Hong Kong, China).

The printing pressure of the four printheads can be independently adjusted using individual air pressure regulators (ARP20K-N01BG-1Z; SMC, Tokyo, Japan). Hydrogel deposition in each printhead is controlled by opening and closing the solenoid valve (VT307-6DZ1-01F-Q; SMC, Tokyo, Japan) connected to the metal-oxide-semiconductor field-effect transistor (MOSFET) terminals of the Rumba controller board.

**Figure 2.** (**a**,**b**) general and detailed view of the modified Witbox 2 3D printer with four bioprinting printheads; (**c**) 3D CAD design of the 3D printer *x*-carriage; (**d**) 5 mL syringe barrels loaded with dyed P407 hydrogels.

#### *2.2. G-Code Generation and Printing Software*

The software employed to control the modified bioprinter and the bioprinting process was comprised of several tools. First of all, a modification of Marlin firmware (v1.1) was loaded into the 3D printer main board [29]. The modified firmware allowed us to manage and coordinate all the activities of the 3D printer, including the movement of the four printheads and the deposition of the bioinks.

A computer-aided design (CAD) software (SolidWorks; Dassault Systems, v2016) was utilized to create the 3D models for bioprinting and generate the final stereolithography (STL) files. The opensource slicing software Slic3r (v1.2.9) [30] was utilized for G-code generation. Slic3r is mainly utilized in FDM and therefore it is not designed to operate pneumatic printheads. For that reason, custom post-processing Perl scripts were required to transform the original G-Code to the particular characteristics of the multi-material 3D bioprinter used. The four printheads moved according to G-code instructions, depositing biomaterials where they were initially programmed. Finally, the G-code was sent to the bioprinter using Repetier-Host (v1.6.2) software [31], which was also in charge of monitoring the bioprinting process.

#### *2.3. Multi-Material Bioprinting Procedure and Calibration*

Figure 3 describes the procedure to prepare a 3D model for the multi-material bioprinting process. This procedure starts opening the STL files containing the original geometry with the slicing software. In case a multi-material printing process is desired, several STL files should be generated, each of them assigned to the particular printhead that will print that part of the geometry. The assigning operation is performed in Slic3r using the "Settings" button. Each STL file will be displayed in a list on the left-hand side of the window and assigned to a specific printhead (Figure 4a).

When several printheads are assigned, the 3D model visualization will appear with a different color for each printhead (Figure 4). If only one printhead is utilized, a single STL will be required. Once the printing settings are introduced (deposition speed, infill pattern, number of perimeters, etc.), the G-code is generated and sent to the 3D printer through the Repetier-Host.

*xy* offsets of the 3D printer utilized were configured according to Figure 5a. When using multiple printheads, the original offset coordinates of the first printhead (P1) are set to zero (*x* = 0, *y* = 0). Then, the *xy* offset coordinates of every additional printhead must be determined with respect to

the coordinates of P1. Every offset must be entered in the slicing software to compensate for the misalignments between the printheads. Depending on the particular printer used and the configuration of its printheads, the values of the offset coordinates can be very different.

**Figure 3.** Flowchart of the procedure for the G-code generation and the interaction with the 3D bioprinter interface.

**Figure 4.** Multi-material printheads assignment using Slic3r software. (**a**) porous structure with four layers stacked, each one assigned to a different printhead; (**b**) 3D model of a heart section composed of four parts; (**c**) printhead trajectories calculated by the slicing software using a porous infill.

**Figure 5.** Schematic representation of the (**a**) *xy*-offsets and the (**b**) *z*-offset of four printheads P1/P4.

*z*-offset between various printheads also represents a crucial point for multi-material calibration (Figure 5b). A *z*-homing push button was installed in the 3D multi-material bioprinter to perform the automatic calibration of the *z*-offsets. This configuration allows us to use nozzles of different types and heights. In addition, this automatic system reduces drastically the time required to start the printing process because there is no necessity to perform any manual adjustment and the whole calibration process is done at once.

#### *2.4. Hydrogel Preparation*

Poloxamer 407 (P407, Pluronic® F127; Sigma-Aldrich, Madrid, Spain) was prepared at 40 wt % by weighing the quantity of polymer required and mixing in cold Milli-Q water at 4 ◦C. P407 powder was added gradually to MilliQ water to facilitate the dilution and stirred vigorously for 3 h using a magnetic stirrer at 4 ◦C. Once the solution was homogenized, it was centrifuged and stored overnight at 4 ◦C to remove air bubbles. P407 prepared solutions were always stored at 4 ◦C until further use.

Gelatin from porcine skin (G1890; Sigma-Aldrich) and sodium alginate from brown algae (A0682; Sigma-Aldrich) were dissolved in phosphate buffered saline (PBS) without salts at 10 wt % and 4 wt % respectively. A solution of 5%Gel-2%Alg was prepared by blending. The pH of the solution was adjusted to 7.2–7.4. Solutions were mixed using vortex and centrifuged at 1000 rpm for 1 min to remove air bubbles.

Four different fluorescent dyes (see clear differences in fluorescence under UV light at the Figure 2d) were utilized to improve the visualization of P407 and Gel-Alg (except in the case of using cells to avoid cytotoxicity) bioinks: orange (1:100; IFWB-33; Risk Reactor, Santa Ana, USA), clear blue (1:500; IFWB-C0; Risk Reactor), yellow-green (1:1000; IFWB-C8; Risk Reactor) and red (1:1000; IFWB-C7; Risk Reactor).
