Liver ductal organoids reconstruct intrahepatic biliary trees in decellularized liver grafts

Three-dimensional scaffolds decellularized from native organs are a promising technique to establish engineered liver grafts and overcome the current shortage of donor organs. However, limited sources of bile duct cells and inappropriate cell distribution in bioengineered liver grafts have hindered their practical application. Organoid technology is anticipated to be an excellent tool for the advancement of regenerative medicine. In the present study, we reconstructed intrahepatic bile ducts in a rat decellularized liver graft by recellularization with liver ductal organoids. Using an ex vivo perfusion culture system, we demonstrated the biliary characteristics of repopulated mouse liver organoids, which maintained bile duct markers and reconstructed biliary tree-like networks with luminal structures. We also established a method for the co-recellularization with engineered bile ducts and primary hepatocytes, revealing the appropriate cell distribution to mimic the native liver. We then utilized this model in human organoids to demonstrate the reconstructed bile ducts. Our results show that liver ductal organoids are a potential cell source for bile ducts from bioengineered liver grafts using three-dimensional scaffolds.


Introduction
Liver transplantation is currently the only curative option for patients with end-stage liver disease. However, the demand for liver organs greatly exceeds the supply of donor livers. To address this challenge, approaches such as cell transplantation, bioartificial organs, and liver support devices have been explored [1,2]; however, none have yet been established as therapeutic alternatives.
Decellularization and recellularization, in which an extracellular matrix (ECM) is prepared from its native organs, retaining the inherent structure and biological properties, followed by recellularization with new cells to create transplantable functional organs, are promising techniques for tissue engineering [3]. Since the first report of a decellularized liver scaffold [4], recellularized liver graft models have been investigated from some liver cell sources. In addition to hepatocytes, a major functional cell type to be recellularized, the source cells include cholangiocytes and endothelial cells, as well as other stromal cells. Hepatocyte cell sources have been widely reported, and include primary hepatocytes, fetal hepatocytes, induced pluripotent stem cells (iPSC)-derived hepatocyte-like cells, and direct reprogramming of fibroblasts [5][6][7][8][9]. In contrast, the study of cholangiocyte cell sources has been rare owing to the difficulty of culturing primary cholangiocytes. Mouse immortalized cholangiocytes [10], normal rat cholangiocytes [11], and iPSC-derived cholangiocytes [9] have been studied as sources to recellularize the decellularized liver tissue ECM. Although there is proof-of-concept to show that external cells can repopulate decellularized liver tissue, further physiological cholangiocyte candidates are needed for cellular sources.
Recent advances have enabled the culture of tissue stem cells as three-dimensional (3D) organoids, which self-organize into 3D structures mimicking the original organs [12]. This organoid culture technique has also been applied to liver stem cells. A bile duct fragment embedded in Matrigel self-organized into liver ductal organoids with bipotential differentiation capacity into both hepatocyte and cholangiocyte lineages [13,14]. Liver ductal organoids have been reported as a potential cell source for hepatocyte regeneration [15].
Therefore, we expect that the liver ductal organoids may also be a cell source for the regeneration of cholangiocytes, and the proliferating and self-organizing ability of the organoids provide a means to obtain cell networks for tissue engineering.
Here, we have described liver ductal organoids as a potential bile duct cell source for a bioengineered liver graft. We characterized the biliary properties of liver ductal organoids in vitro and those of repopulated bile ducts in a bioengineering liver graft ex vivo. The appropriate cell placement and structure of the recellularized liver also demonstrated advantages for future clinical progress.

Mouse liver ductal organoids exhibit characteristics of functional cholangiocytes
Although liver ductal organoids are derived from intrahepatic bile ducts, these cells have mainly been investigated as a resource for differentiated hepatocytes [13,14,[16][17][18] rather than for cholangiocytes [19,20]. We investigated to what extent the ductal organoids are characteristic of cholangiocytes. Liver ductal organoids showed rapid proliferation when cultured in Matrigel ( Fig. 1A-C). Liver ductal organoids under maintenance culture were analyzed by RT-qPCR and immunofluorescence to determine the expression of markers specific to cholangiocytes and hepatocytes. In the RT-qPCR analyses, the expression of the cholangiocyte-specific markers Krt19 and Sox9 was 0.68-fold lower and 2.0-fold higher in ductal organoids than extrahepatic bile ducts, respectively ( Fig 1D). In contrast, the expression of hepatocyte-specific markers, Alb, Hnf4a, and Cyp3a11, was quite low compared with that in primary hepatocytes.
Immunofluorescence analyses revealed that ductal organoids exhibit cystic structures expressing cholangiocyte lineage markers, such as KRT19, SOX9, ASBT, and CFTR, whereas ALB was absent (Fig. 1E). These findings indicate that liver ductal organoids under maintenance culture conditions, which reportedly possess bipotent stemness, exhibit characteristics of cholangiocytes but not of hepatocytes.
We subsequently characterized liver ductal organoids by focusing on their cholangiocyte function. The multidrug resistance protein 1 (MDR1) transporter is expressed in normal biliary epithelia and is involved in the efflux of a broad ranges of substrates into the lumen [21]. We thus evaluated the ability of organoids to efflux rhodamine 123, which is mainly transported by MDR1 transporters. Liver ductal organoids were incubated with the dye rhodamine 123 ( Fig. 2A), and the dye was transported into the lumen of the cystic organoids. In contrast, in the presence of 20 µM verapamil, which inhibits MDR1 function, rhodamine 123 did not accumulate in the lumen ( Fig. 2A). Quantification of the fluorescence intensity showed that verapamil significantly blocked the transportation of rhodamine 123 (P value < 0.0001) (Fig. 2B), indicating that this fluorescent dye was actively transported by MDR1.
In addition, to assess the function of another transporter expressed in normal biliary epithelia, cystic fibrosis transmembrane conductance regulator (CFTR), we conducted a forskolin-induced swelling assay. This assay evaluates the increase in CFTR function induced by the activation of cAMP pathway, which is stimulated by forskolin, and leads to cyst swelling [22,23]. Liver ductal organoids were incubated with forskolin in the absence or presence of CFTR inh -172, a specific CFTR inhibitor (Fig. 2C). Organoids increased the cyst size by 2.66-fold after forskolin treatment (Fig. 2D). This forskolin-induced swelling ratio was significantly reduced to 1.55 by CFTR inh -172 (P value < 0.0001), indicating that liver ductal organoids have CFTR functional activity. The inhibition of CFTR was not associated with cell death (P value > 0.4) (Fig. 2E). These findings suggest that liver ductal organoids have the properties of functional cholangiocytes.

Mouse liver ductal organoid cells repopulate the decellularized rat liver ECM to reconstruct biliary trees
As the ductal organoids exhibit the characteristics of cholangiocytes to certain extent, we then investigated if the ductal organoids were a potential cell source for bioengineered livers. We have previously described a decellularized liver graft technique that offers a bioengineered scaffold with a physiological ECM [6,8] (Supplementary Fig. S1). In addition to the native ECM, decellularized liver also retains the vascular and biliary network frame structure ( Supplementary Fig.  S2). Histological analysis confirmed the absence of cells in the decellularized liver scaffold (Supplementary Fig. S3). After confirming that the biliary structure was preserved in the decellularized liver, liver ductal organoid cells were then injected into the biliary network of a decellularized rat whole liver scaffold via the common bile duct (Fig. 3A). The recellularized scaffold was cultured in an ex vivo perfusion culture system for 3-5 days (Fig. 3B); then, the tissue was formalin-fixed and paraffin-embedded for histological analysis. The ductal organoid-derived cells engrafted along the bile duct walls, forming a monolayered structure lining the lumens (Fig. 3C). RT-qPCR analyses revealed that recellularized bile ducts expressed cholangiocyte marker genes, including Krt19, Sox9, Cftr, and Hnf1β at comparable levels with those in the extrahepatic bile duct in vivo (Fig. 3D), whereas hepatocyte markers (Alb, Cyp3a11, Hnf4a) remained at a low level. Moreover, the repopulated organoid cells sustained the expression of stemness markers (Lgr5, Prom1). In the immunofluorescence analyses, the repopulated organoid cells also exhibited key biliary markers (KRT19, SOX9, ASBT, CFTR) (Fig. 4A). PCNA, a proliferation marker originally positive in ductal organoids, was maintained in some areas even after recellularization (Fig. 4B), indicating that the engrafted cells retained proliferative ability in the bile duct structure. From these findings, it was determined that repopulated liver ductal organoids were engrafted along the bile duct ECM, maintaining cholangiocyte properties.
Next, to demonstrate the 3D structure of recellularized bile ducts, ductal organoid cells expressing GFP were inoculated into a decellularized liver. The cells spread to the periphery of the liver, displaying a branched tree-like network ( Fig. 5A). In addition, the confocal microscopy images showed engrafted cells on the decellularized bile duct ECM that formed luminal structures ( Fig. 5B and Supplementary video S1). As shown by these results, the intrahepatic bile ducts were recellularized efficiently by the injection of ductal organoid-derived cells via the common bile duct.

Liver ductal organoid cells retain cholangiocyte characteristics during simultaneous recellularization with primary hepatocytes
Hepatocytes are the main functional cellular unit in the liver. We evaluated whether the co-recellularization of repopulated liver ductal organoids and hepatocytes affected their engraftment or differentiation properties. As the optimal culture medium for hepatocytes and cholangiocytes differs, we first injected 5 × 10 6 liver ductal organoid-derived cells via the common bile duct and the recellularized liver was perfused with expansion medium (EM) containing 10 μM forskolin (Supplementary Table S1), a bile duct organoid EM, via the portal vein for 5 days. Then, freshly isolated mouse primary hepatocytes (5 × 10 7 cells) were injected via the common bile duct, followed by perfusion culture with HCM ™ (Lonza Sales Ltd, Basel, Switzerland), a hepatocyte culture medium, via the portal vein for 2 days (Fig. 6A). Histological analyses of this co-recellularized liver revealed the appropriate cell distribution of hepatocytes into the parenchymal space and of liver ductal organoid cells into the bile duct ( Fig. 6B, C). Through the immunofluorescence analyses, repopulated primary hepatocytes were shown to express ALB and HNF4a. Repopulated liver ductal organoid cells expressed KRT19 and SOX9 but not ALB or HNF4a (Fig. 6D), showing that the repopulated cells maintained biliary lineage and did not differentiate to hepatocyte lineage when co-cultured with primary hepatocytes.

Human liver ductal organoid cells are capable of repopulating decellularized rat liver ECM
These findings must be extended to human cells to establish a transplantable human liver graft in the future. Human liver ductal organoids have bipotential capacity to differentiate into both hepatocytes and cholangiocytes, similar to mouse liver ductal organoids [14]. We generated three liver ductal organoid lines from residual liver specimens from patients undergoing hepatectomy in response to liver tumors (Table 1). Human liver ductal organoids cultured in Matrigel (Fig. 7A, B) expressed key biliary markers (KRT19, SOX9) but not ALB (Fig. 7C). Next, we injected the human organoid cells into rat decellularized liver and achieved successful recellularization (Fig. 7D), similar to the mouse organoids, in a rat decellularized liver. The RT-qPCR analysis demonstrated that KRT19 and PROM1 were upregulated in all cases evaluated and SOX9 was upregulated in two of three cases (Fig. 7E). In contrast, the gene expression of hepatocyte markers (ALB, CYP3A4) remained low (Fig. 7E). In the immunofluorescence analyses, the recellularized bile ducts were shown to maintain the expression of KRT19 and SOX9 (Fig. 7F). Therefore, human liver ductal organoids maintained cholangiocyte properties after recellularization.

Discussion
Here, we have reported that liver ductal organoids can be a useful cholangiocyte cell source for bioengineered bile ducts. Luminal structures with a single cell layer lining were successfully reconstituted in the 3D scaffold decellularized from the liver which expressed appropriate bile duct markers. In addition, the co-recellularization with both primary hepatocytes and liver ductal organoids enabled the proper cell placement similar to the native liver, maintaining specific markers for each cell type.
Since the liver ductal organoid was first introduced as a bipotential stem cell culture [13], hepatocyte lineage differentiation from the ductal organoids has been well studied [17,18]. Very recently, several papers investigating the cholangiocyte characteristics of this bipotent organoid were published [19,20,24]. The induction of biliary lineage differentiation in liver ductal organoids upregulated cholangiocyte-related genes and downregulated stem cell markers such as Lgr5 and Ki67 [19]. However, in this study, we found that the liver ductal organoids already possess cholangiocyte-like properties prior to differentiation, including representative cholangiocyte markers and functions. In addition, the repopulated liver ductal organoid-derived cells remained proliferative and retained the expression of stem cell markers Lgr5 and Prom1 as well as key biliary markers (Krt19 and Sox9). This characteristic appears to be a favorable feature as a source for tissue engineering because maintaining the proliferating potential is important to enable engrafted cells to expand and self-organize in a decellularized liver graft ex vivo, or even after transplantation. Liver ductal organoids morphologically consisted of flat cells forming a cystic structure in vitro, meanwhile the recellularized bile ducts showed luminal structures in the form of a simple columnar epithelium. We also confirmed that repopulated liver ductal organoids were not differentiated into hepatocyte lineage, despite their bipotential capacity and perfusion with hepatocyte medium in the co-recellularized liver.
With regard to cholangiocyte cell sources for recellularization, a similar approach was recently reported [20], in which engineered extrahepatic bile duct models in vitro were established using human decellularized extrahepatic bile ducts repopulated with bile duct-derived organoids. They concluded that among three types of bile duct-derived organoids, namely liver ductal organoids, extrahepatic bile duct-derived organoids, and bile-derived organoids, extrahepatic bile duct-derived and bile-derived organoids repopulated decellularized extrahepatic bile ducts efficiently in vitro. While the liver ductal organoid was a promising extrahepatic bile duct cell source, our study demonstrated that liver ductal organoids from both mice and humans successfully repopulated intrahepatic bile ducts in decellularized rat liver. Given the regional differences in the cell characteristics and gene profiles between extra and intrahepatic bile duct organoids [24,25], it is reasonable to use region-specific organoids for each target of intra and extrahepatic bile ducts to establish an engineered liver graft. To avoid the immune response, an autologous cell source that does not require immunosuppression therapy is a potential solution. As liver ductal organoids can be generated from a small piece of a patient liver without gene editing, engineered bile ducts can be reconstructed using autologous cells. Moreover, the liver ductal organoids are genetically stable during long-term expansion [14], which allows the supply of a large number of cells. Because liver ductal organoids are also envisioned as a hepatocyte cell source owing to their bipotential capacity, large numbers of the organoids will be needed for transplantation as the hepatocyte cell source. To supply large numbers of cells, a highly efficient method of culturing liver ductal organoids has been reported [16].
Given these advantages, liver ductal organoids have the potential to be a cell source for bioengineered liver grafts.
We demonstrated the cholangiocyte properties of the repopulated liver ductal organoids, although functional interactions between recellularized bile ducts and hepatocytes remained ambiguous. The integrated function of the bile efflux and transportation are important in recellularization. Primary hepatocytes quickly lose their functions in vitro [2,26]. Likewise, it is difficult to maintain recellularized primary hepatocytes in perfusion culture for a long time, whereas repopulated liver ductal organoids were viable for more than a week. Such a limited viability restricts the ability of recellularized hepatocytes to construct the functional structures. Thus, in future studies, it is important to improve hepatocytes as a resource for recellularization, as well as to evaluate the cellular integration between recellularized bile ducts and hepatocytes upon experimental transplantation.
In conclusion, we demonstrated that liver ductal organoids are a useful cell source for the reconstruction of intrahepatic bile ducts in a 3D scaffold of decellularized liver and to enable the appropriate cell distribution of recellularized hepatocytes and bile ducts in bioengineered liver grafts. It would be of great interest to develop liver ductal organoids as a hepatocyte source as well as to generate functional recellularized liver grafts in vivo for the clinical application of these cells in the future.

Preparation and maintenance of human liver ductal organoids
Human liver specimens (0.5-1.0 cm 3 ) were obtained from the non-tumorous part of the resected liver from patients who underwent hepatectomy. The specimens were subjected to the same procedure as the mouse organoid culture described above. For passaging, the application time of TrypLE Express was 10 min. The organoids were cultured in human EM (Supplementary Table   S1) [14].

Isolation of mouse primary hepatocytes
Primary hepatocyte isolation was performed using two-step collagenase perfusion technique, as previously described [27]. Briefly, a C57BL6/J mouse was anesthetized with isoflurane (Wako) and the inferior vena cava was

Cell viability assay
After 3,000 single cells dispersed from liver ductal organoids were embedded in Matrigel, the organoids were cultured in EM for 5 days. Organoid viability was quantified every day from aliquot of the organoid using CellTiter-Glo ® (Promega, Madison, WI) and GloMax Discover Microplate Reader (Promega).

Histological analysis
Organoids were retrieved from Matrigel, re-embedded in Cellmatrix Type I-A (Nitta Gelatin), and fixed in 10% formalin (Wako) overnight at room temperature.
Recellularized liver grafts were fixed in 4% paraformaldehyde (Wako) for 24 h at 4°C or in 10% formalin overnight at room temperature, and then embedded in paraffin.
Paraffin-embedded sections (4 μm) were dewaxed, rehydrated, and subjected to either hematoxylin and eosin staining or immunohistochemical staining. For immunostaining, antigen retrieval was performed by autoclave for 15 min at 121°C, and the sections were by incubation in PBS containing 10% donkey serum and 0.1% Triton X-100 (Nacalai Tesque Inc., Kyoto, Japan). All antibodies were diluted with PBS with 5% donkey serum and 0.1% Triton X-100.
The sections were incubated with primary antibody overnight at 4°C, followed by incubation with secondary antibody for 1 h at room temperature, and then mounted using ProLong™ Gold Antifade Mountant with DAPI (Invitrogen). The stained sections were visualized using an Olympus BX50F4 microscope (Olympus Optical, Tokyo, Japan). The antibodies used for immunostaining are listed in Supplementary Table S2.

Quantitative real-time PCR
Before RNA extraction, liver ductal organoids in Matrigel were collected and washed twice with PBS. To extract RNA from the recellularized liver, the recellularized liver after perfusion culture was incubated with RPMI (Wako) containing 10 mg/mL collagenase type II (Gibco) for 15 min at 37°C. The solution of dissolved liver was centrifuged, the supernatant was removed, and the pellet was processed for RNA extraction.
Total RNA was prepared from organoids and engineered liver grafts using

CFTR functional assay
The CFTR functional assay was performed based on a previous report [23].
Briefly, 1000 individual cells were embedded in a 5 µL Matrigel drop and incubated in EM for 4 days. Then, the gels were preincubated with EM containing DMSO or 60 µM Inh-172 (Cayman, Michigan, USA), a CFTR-specific inhibitor, for 3 h. The gels were subsequently incubated with EM containing 10 µM forskolin (Wako) with DMSO or Inh-172 for 24 h before imaging with a Leica DMi8 (Leica Microsystems) using the Leica Application Suite X (LAS-X) software.
The swelling rate of the organoids was calculated by comparing the total area of each organoid before and after the stimulation, using ImageJ.

Harvest, decellularization, and recellularization of rat whole liver
Harvest and decellularization. Harvest and decellularization of the liver were performed as previously reported [8]. Under general anesthesia with isoflurane Recellularization. Liver ductal organoids were dissociated into single cells and 3-5 × 10 6 cells suspended in 5 mL of EM were administered into the scaffold through the bile duct at a flow rate of 1 mL/min. Before perfusion, the recellularized liver was incubated at 37°C in EM supplemented with 50 µg/mL gentamicin (Gibco), 2.5 µg/mL of amphotericin B (Gibco), and 10 µM forskolin for 3 h. We applied our previously described recellularization protocol for primary hepatocytes [8]. In total, 5 × 10 7 hepatocytes were suspended in 30 mL of HCM ™ (Lonza, Sales Ltd, Basel, Switzerland) and injected via the bile duct at a flow rate of 1 mL/min. The recellularized liver was incubated at 37°C for 3 h before starting perfusion culture.

Perfusion culture
After incubation for 3 h following recellularization, the recellularized liver was placed in the circulation culture system, and the cannula of the portal vein was connected. The perfusion culture was conducted to continuous flow at a rate of 0.7-1.0 mL/min at 37°C.

3D bile duct imaging
The liver with repopulated bile ducts of GFP-expressing cells was cultured in the circulation culture system for 5 days and imaged using confocal microscopy (Leica Microsystems). A reconstructed 3D image was prepared from Z-stack images using ImageJ (National Institutes of Health).

Statistical analyses
A p value of < 0.05 was considered to indicate statistical significance; analyzes were performed by unpaired t-test using GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA) (Fig. 2B, D, E).