In vitro construction of liver organoids with biomimetic lobule structure by a multicellular 3D bioprinting strategy

Abstract Liver disease is one of the serious threats to human life and health. Three‐dimensional (3D) liver models, which simulate the structure and function of natural liver tissue in vitro, have become a common demand in medical, scientific and pharmaceutical fields nowadays. However, the complex cellular composition and multi‐scale spatial arrangement of liver tissue make it extremely challenging to construct liver models in vitro. According to HepaRG preference and printing strategy, the formulation of bioink system with opposite charge is optimized. The sodium alginate‐based bioink 1 and dipeptide‐based bioink 2 are used to ensure structural integrity and provide flexible designability, respectively. The HepaRG/HUVECs/LX‐2‐laden liver organoids with biomimetic lobule structure are fabricated by a multicellular 3D droplet‐based bioprinting strategy, to mimic the cell heterogeneity, spatial structure and extracellular matrix (ECM) features. The liver organoids can maintain structural integrity and multicellular distribution within the printed lobule‐like structure after 7 days of culture. Compared with the 2D monolayer culture, the constructed 3D organoids show high cell viability, ALB secretion and urea synthesis levels. This study provides a droplet‐based and layer‐by‐layer 3D bioprinting strategy for in vitro construction of liver organoids with biomimetic lobule structure, giving meaningful insights in the fields of new drugs, disease modelling, and tissue regeneration.


| INTRODUCTION
The liver, as the largest metabolic organ of human body, is responsible for performing a variety of complex functions such as biosynthesis, secretion, storage, metabolism and detoxification. Unfortunately, orthotopic liver transplantation (OLT) is still the only effective treatment for end-stage liver disease. However, the problems of donor organ shortage, matching type and immune rejection greatly limit its wide application. 1,2 Over the past few decades, liver modelling using human cells in petri dishes has contributed greatly to a variety of important research related to liver disease treatment and liver regeneration. [3][4][5] Organoid models have gradually matured and become prominent in recent years. [6][7][8][9] Compared with traditional monolayer culture and animal models, organoids are able to partially reproduce key physiological features of the native human organs, including cell heterogeneity, spatial structure and microenvironment, and provide important cell-cell and extracellular matrix (ECM) interactions. 10,11 These can promote hepatocyte proliferation, differentiation, expression of specific genes and proteins, and response to exogenous stimuli. 12 3D bioprinting is a promising technique that enables spatial patterning of multiple cells and matrix materials in a preset manner, offering unprecedented potential for creating organoids that mimic the microscopic structure of native organs. 5,13,14 Bioprinting of hepatic and endothelial cells to produce structured spheroids has showed better performance including long-term culture, cell viability and high MRP2, albumin (ALB), and CD31 expression levels when compared to the non-structured spheroids. 15 A hepatic construct with hepatocyte progenitor cells having a hexagonal structure has been developed by a two-step bioprinting strategy, which showed significantly improved hepatocyte phenotype and function. 16 An array of liver lobules prepared using an extrusion-based bioprinting method demonstrated that the multicellular types with spatial patterns in bioinks play a role in cell organization and function. 17 In addition, a liver model constructed by extrusion bioprinting showed clinically relevant dose-dependent hepatotoxicity. 18 Moreover, in a bioprinting study, the multicellular composition of hepatic parenchymal, stellate and endothelial cells was shown to be necessary to construct a liver fibrogenesis model. 19 However, due to the complex cellular composition of liver tissue and the multi-scale spatial arrangement between cell and ECM, the challenges for 3D bioprinting of liver models are mainly limited by the performance of bioink materials. The characteristics of bioink materials not only need to meet the printability requirements, but also need to meet the bionic needs of different cell types for specific ECMs. So far, few engineered liver models have been able to simultaneously mimic multicellular composition, spatial patterns, and ECM characteristics of natural livers.
Hence, this study aims to in vitro construction of liver organoids with biomimetic lobule structure by a multicellular droplet-based 3D bioprinting strategy. According to the literature, 19 the multicellular composition of the 3D liver model consists of hepatic parenchymal HepaRG cells, stellate cells (LX-2) and human umbilical vein endothelial cells (HUVECs). The selected multi-material bioinks, including sodium alginate (SA), hyaluronic acid (HA) and dipeptide (with preferred sequence by HepaRG cells), are deposited in a layer-by-layer way eventually to form a self-standing construct with defined structure. The multicellular 3D printing and bioink material design strategies developed in this study have great potential in the development of 3D organoids that mimic the native microenvironment of organs.

| Cell culture
The used cells in this study, including HepaRG, LX-2, HUVECs, green-fluorescent HepaRG (HepaRG-GFP), bule-fluorescent LX-2 (LX-2-BFP), and red-fluorescent HUVECs (HUVECs-RFP), were cultured in DMEM containing 4.5 g/L D-glucose, supplemented with 10% (v/v) FBS (Gibco) and 1% (v/v) penicillin-streptomycin at 37 C with 5% CO 2 in an incubator. In the experiment using 2D culture as a control, the 2D monolayer culture group maintained the same HepaRG cell number as the 3D model. During cell culture, half of culture medium was replaced by the fresh one every 2 days.
The electrostatic interaction between negatively charged bioink 1 and positively charged bioink 2 ensures in-situ gelation to preserve the fabricated structure. For the cell-laden bioinks, HepaRG cells were encapsulated in bioink 1 at a final density of 7.5 Â 10 6 cells/mL. LX-2 and HUVECs were respectively embedded in bioink 2 to achieve a final density of 1 Â 10 7 and 2.5 Â 10 6 cells/mL, respectively. The used ratio of HepaRG/HUVECs/LX-2 in constructed hydrogel scaffold was maintained approximately 10:8:3 during the printing process.

| Characterizations of the prepared bioinks
The prepared bioinks were characterized by zeta potential measurements, atomic force microscopy (AFM) imaging and rheological analysis. The bioinks were diluted for 20-fold by Milli-Q water to zeta potential measurements on a Zetasizer Nano-ZS (Malvern) at 25 C.
For AFM imaging, 5 μL of bioink samples were deposited on mica sheets. After air-dried, samples were observed using FASTSCANBIO AFM (Bruker) in a tapping mode, and the obtained images were processed using the nano-scope analysis 1.9 software (Bruker). The rheological analysis including amplitude and frequency sweeps was performed on a MCR 302 rheometer (Anton Paar) equipped with a 25 mm diameter parallel plate at a 0.5 mm gap. The test temperature was maintained at 25 ± 3 C. The amplitude sweep was carried out in the strain range of 0.01%-100% with a fixed frequency of 1 Hz. The frequency sweep was performed in the range of 0.1-100 Hz at a constant shear strain of 1%.

| Model designing and 3D bioprinting
The liver model with biomimetic lobule structure consisted of 12 hexagons, three cell types, and two formulated bioinks. A layer-by-layer bioprinting strategy was conducted by a droplet-based Digilab CELLJET™ 3D bioprinter (Thermo Fisher Scientific) equipped with a 190 μm inner-diameter nozzle. Prior to printing, the prepared bioinks were sterilized for 30 min under UV light and then loaded into the dispensing system by the syringe pump through the tip orifice. The dispense positions were programmed as coordinates by AxSys™ software, and the dispense volume for each drop was optimized as 50 nL. The spacing distance between droplets is 375 μm.
The diameter of each droplet formed in the constructed hydrogel scaffold is approximately 294 ± 11 μm.

| Swelling and degradation of hydrogel scaffold
The swelling and degradation of hydrogel scaffolds were studied by gravimetric and rheological methods. The initial weight of the fabricated hydrogel scaffold was marked as W i . During test, printed hydrogel scaffolds were immersed in PBS solution and maintained at 37 C. At different time points (Day 1, 5 and 10), the liquid on hydrogel surface was gently removed by filter paper, and the wet weight of the hydrogel scaffold was measured as W t . The swelling and degradation of hydrogel scaffold were evaluated by wet weight gain% = At each time point (Day 0, 1, 5 and 10), the elastic modulus (G 0 ) of hydrogel scaffold was determined by the rheological method described above.

| In vitro analysis of hepatic functions
For hepatic function measurements, the ALB secretion and urea synthesis were quantified using the corresponding assay kits following the protocol from the manufacturer. At days 1, 4, and 7 during culture, the supernatant was collected and stored at À80 C. The insoluble matter of the sample was centrifugally removed prior to testing. ALB and urea secretions in the supernatant were determined using the ELISA kit (ab179887, Abcam) and urea assay kit (ab83362, Abcam), respectively. ALB and urea contents were assayed by absorbance at 450 nm and 462 nm on a microplate reader (Thermo Scientific), respectively.

| Statistical analysis
All tests were repeated three times in parallel, and all data in this study were presented as the mean value ± standard deviation. The statistical analysis was performed using one-way ANOVA test to determine the significant levels at **p < 0.01 and *p < 0.05.

| Bioink formulation and characterizations
In the liver organoid construction, two bioinks were required to load HepaRG and LX-2/HUVECs cells, respectively. The bioink 1, composed of SA/HA/RGD, was used to encapsulate HepaRG cells. The dipeptide material was used as bioink 2 to respectively load LX-2/ HUVECs cells. The design inspiration for this bioink system is as follows: in view of the complex structure to be constructed and the regionalized spatial distribution of the three types of cells, it seems necessary to use polymeric materials in combination with supramolecular materials. As shown in Figure S1 Figures 2F and 3A). During the construction of liver organoids, the layer-by-layer printing process was shown in Figure 2D. As can be seen from Figures 2E and S4 Figure 2F. It can be seen that three fluorescent cells were spatially well-organized in the printed structure. Figure 3A showed the cellular organization and distribution of the liver organoids with biomimetic lobule structure, after 7 days of culture. During culture, the three cells spontaneously organized together. Cell proliferation ( Figure 3B) assayed by absorbance at 450 nm was in accordance with the live-dead stain results ( Figure 3C). Since 2D culture was used as a control for 3D models, the initial cell density was relatively high compared to conventional 2D monolayer culture. As demonstrated in Figure 3B, the number of cells in 2D culture remained dynamically stable during the first 4 days, but on Day 7, a large number of cells died due to contact inhibition ( Figure 3B,C). Within 7 days, the cells embedded in the liver organoids showed a steady proliferation and maintained viability ( Figure 3B,C).

| ALB and urea secretions
The ALB secretion and urea synthesis of the HepaRG/HUVECs/LX-2-laden liver model were determined by immunofluorescence staining and quantitative assay kits. Figure 4A showed that the ALB secretion in 3D organoids was significantly higher than that in 2D culture. As can be seen from Figure 4B,C, ALB and urea secretions in both liver models were time-dependent and increased significantly with the prolonged culture during 7 days. On Day 7, the total content of ALB secreted in the supernatant from 2D culture and 3D organoid models supramolecular hydrogels have the potential to become novel bioink materials. 20,32 They usually have good biocompatibility, ordered and reversible nanostructures, and stimulus responsiveness to pH, temperature, light, and so forth. In this study, dipeptide materials were used as bioinks based on the preferred sequence by HepaRG cells and the printing strategy.
The liver model with lobule-like structure was conducted by a droplet-based and layer-by-layer bioprinting strategy. 33