**3. Results**

### *3.1. System Testing for E*ffi*ciency of Sphere Capture and Transfer*

To test the ability of the vacuum pump system and print head to pick up spheroids, glass beads similar in diameter to the average spheroids that would be used with this printing setup were used.

These beads, according to the supplier's specifications, had a diameter of 700 ± 70 μm on average (Figure 3A), which is very similar to the ~800 μm diameter of the spheroids that would be used. The efficiency of the system was determined by using two separate containers (Beads Only and Beads + PBST) of spheres which were picked up by turning on the vacuum pump system for 10 s at a time, removing the print head (Figure 3B) from the container, and then counting how many spheres had been picked up. After collecting the beads (Figure 3C) various times (n = 28 for Beads Only; n = 18 for Beads + PBST), the efficiencies of sphere capture were determined to be 98.3 ± 0.5% for Beads Only and 98.4 ± 0.6% for Beads + PBST (Figure 3D).

**Figure 3.** Vacuum and print head capture efficiencies. A substitute was chosen for spheroids because of the non-sterile conditions and multiple tests that needed to be performed. (**A**) Zirconia beads with a diameter (700 ± 70 μm) equivalent to that of the spheroids were used. (**B,C**) The print head (B, seen from bottom) was used to pick up (C) a 4 × 4 array of beads in a dry setup. After the print head was placed in the desired container of beads, the vacuum was turned on for 10 s. At the end, the number of beads captured was counted, and (**D**) the efficiencies were calculated. (**E**) Then, the optimal time for the vacuum to be on at which most beads would be captured was determined to be 5 s (\* p < 0.05, n = 13 for each); (**F**) how the number of beads in the container affected the precision of the bioprinter was also determined. Scale bars: (**A**) 1 mm, (**B**), and (**C**) 2 mm.

To further optimize the printing process, the time during which the print head remained in the bead container was varied (1 s, 2 s, 5 s, and 8 s) to determine the optimal time needed to pick up most of the beads. For each time, the corresponding efficiencies (n = 13 for each time point) were calculated

(Figure 3E). From this test, it was determined that 5 s was the shortest and most e fficient time, with an efficiency of 97.8 ± 0.9% which was significantly di fferent (p < 0.05) from those measured for times of 1 s (91.4 ± 2.3%) and 2 s (92.6 ± 1.7%) but did not change significantly when the time increased to 8 s (97.5 ± 1.1%) or 10 s.

In addition, due to the fact that as spheroids are removed, more empty space in the spheroid container would be present, a test was done to determine if the number of beads within the container affected the capture e fficiency. Beads were added to the dish at amounts of 100, 500, 1000, 2500, and 5000, estimated on a mass basis. The print head was used to capture beads a number of times (n = 10 for each group), and the subsequent e fficiencies were determined (average e fficiency: 97.6 ± 0.2%) (Figure 3F). There was no significant di fference between any amounts of beads tested, showing that as the number of beads decreases in the container, the e fficiency does not change.

The repeatability of the bioprinter ability to place the beads in the same area with each subsequent layer was tested to ensure that the spheroids could accurately be placed on the needles when generating tissues. This was determined by placing a full layer of beads into a pre-determined area on a layer of silicone multiple times. After every single layer was placed, an image was taken of the beads, with each placed layer given a di fferent color (Figure 4A). A composite image of each layer aligned together was generated to demonstrate that the bioprinter could place the spheroids in the same area, made apparent by the overlap in colors (see yellow color) (Figure 4B). To estimate the degree of overlap between the layers, a colocalization tool in conjunction with ImageJ was used to calculate the overlap coe fficient for each combination of any two layers (Figure 4C). The overlap coe fficients between any two layers were greater than 0.99, showing that each layer closely overlapped with the others.

**Figure 4.** Bead layer-by-layer overlap coe fficient determination. To determine the precision of the bioprinter, ( **A**) layers of beads were added to a predefined area (upper panels), and each run was assigned a color (lower panels). (**B**) Images from individual runs were overlapped using ImageJ. The yellow color indicates the overlap of the beads. ( **C**) The overlap coe fficient was determined (Colocalization Finder for ImageJ) using the images, and the average overlap coe fficient was found to be 0.997. Scale bars: 2 mm.

### *3.2. Alginate Bead Formation and Print Head Testing*

Small beads made from alginate were chosen as a surrogate for cellular spheroids for testing the bioprinter's ability to print onto the needle arrays correctly. Alginate beads were synthesized using a dropwise technique (Video S2) and separated according to their diameter. Once these beads were separated, they were collected into a single dish and measured using phase-contrast microscopy (Figure 5A). The average diameter of the beads (n = 27) was determined to be 851 ± 18 μm, which is similar to the desired diameter of 800 μm. The beads were then picked up with the print head (Figure 5B) and transferred to the needles (Figure 5C) using the bioprinting process (Video S3). Placing a single layer of beads took approximately 45 s.

Additionally, to determine how e ffectively the bioprinter could print multiple layers of spheroids, two layers were printed onto the needles (Figure 5D; Video S4). From these data, it was seen that the bioprinter could successfully print the alginate beads in multiple layers. Our results clearly show that multilayer printing of spheroids is possible using our newly designed bioprinter.

**Figure 5.** Bioprinter testing with alginate beads. Soft alginate beads were used as a spheroid surrogate as they could easily be punctured and placed on the bioprinter needles. ( **A**) The beads were synthesized at an average diameter of 851 ± 15 μm, which was similar to the spheroid diameter. (**B**–**D**) Picture (**B**) showing beads aspirated onto the end of a 4 × 4 arrangemen<sup>t</sup> print head and ( **C**) printed onto the needles in a single layer. ( **D**) Two layers were also tested to ensure that multiple layers could be printed. Scale bars: ( **A**) 500 μm; (B) 2 mm; ( **C**) and ( **D**) 5 mm.

### *3.3. Proof-Of-Concept Testing with hiPSC Spheroids*

We confirmed the potential of our bioprinter to place cellular spheroids onto the needles in a layer-by-layer fashion. For this, hiPSC spheroids were grown to a similar diameter of, on average, 718 ± 77 μm (Figure 6A, Image S1) and added to the round-bottom container. The bioprinter was started, and a single layer of spheroids was picked up and placed onto the needles. The top (Figure 6B,C) and side (Figure 6D,E) view images show a single layer of spheroids on the needles. These data clearly show that the bioprinter was effectively able to aspirate and transfer a single layer of cellular spheroids onto the needles.

**Figure 6.** Proof-of-concept testing with human induced pluripotent stem cells (hiPSC) spheroids. hiPSC spheroids were cultured using a rotating culture flask method. (**A**) The spheroids were grown to an average diameter of 718 ± 77 μm. Once enough spheroids were available to be used with the bioprinter to create a single layer, they were loaded into the bioprinter and placed onto the needles. (**B**,**C**) Side view of the hiPSC spheroid layer after placement onto the needles. (**D**,**E**) Top view of the hiPSC spheroid layer. The ability to print cellular spheroids onto the needle array confirms the efficiency of the layer-by-layer printing technique presented here.
