Liver Organoids: Formation Strategies and Biomedical Applications

Abstract

The liver is the most important digestive organ in the body. Several studies have explored liver biology and diseases related to the liver. However, most of these studies have only explored liver development, mechanism of liver regeneration and pathophysiology of liver diseases mainly based on two-dimensional (2D) cell lines and animal models. Traditional 2D cell lines do not represent the complex three-dimensional tissue architecture whereas animal models are limited by inter-species differences. These shortcomings limit understanding of liver biology and diseases. Liver organoid technology is effective in elucidating structural and physiological characteristics and basic tissue-level functions of liver tissue. In this review, formation strategies and a wide range of applications in biomedicine of liver organoid are summarized. Liver organoids are derived from single type cell culture, such as induced pluripotent stem cells (iPSCs), adult stem cells, primary hepatocytes, and primary cholangiocytes and multi-type cells co-culture, such as iPSC-derived hepatic endoderm cells co-cultured with mesenchymal stem cells and umbilical cord-derived endothelial cells. In vitro studies report that liver organoids are a promising model for regenerative medicine, organogenesis, liver regeneration, disease modelling, drug screening and personalized treatment. Liver organoids are a promising in vitro model for basic research and for development of clinical therapeutic interventions for hepatopathy.

Introduction

The liver is an important metabolic organ located in the abdominal cavity. It performs more than 500 functions including metabolic, synthetic, immunologic, and detoxification processes [1]. The liver exhibits a high regenerative capacity. The unique regenerative mechanisms of the liver ensures that the liver-to-bodyweight ratio is always maintained which is indispensable for body homeostasis [2]. Liver diseases (fatty liver, liver cancer, genetic inherited disorders and viral hepatitis) account for approximately 2 million deaths per year worldwide due to the shortage of donor livers and poor understanding of the mechanisms of liver pathology [3]. Therefore, studies on liver development, regeneration and pathogenesis play significant roles in regenerative medicine and disease therapy.

Currently, several studies have explored a variety of cell and animal models[4]. Traditional two-dimensional (2D) culture is limited by long-term and stable expansion[5]. In 2D culture, primary hepatocytes (PHs) gradually undergo morphological changes and specific function loss within a short period of time, and eventually die [6, 7]. In addition, 2D culture does not mimic intricate three-dimensional (3D) architecture and cellular heterogeneity of the liver tissue. Moreover, it does not exhibit cell–cell and cell–extracellular matrix interactions that are essential for maintaining in situ phenotypes and biological functions, and tissue-specific cellular processes [8–10]. Establishment of animal models is labor-intensive, resource-intensive and takes a long time [11]. Furthermore, animal models are limited by interspecies phenotypic and genetic characteristics and animal ethics [10].

Advances in studies on extracellular matrix (ECM) biology, have led to better understanding of signaling pathway regulation of stem cell niches and differentiation and have allowed exploring mechanism of self-organization and organoid culture systems [8, 12, 13]. Organoids are described as 3D cell culture systems, which refers to formation of 3D cell condensates or organ analogues by self-organization to mimic the spatial structural and physiological functional characteristics of an organ in vitro [14, 15]. After several generations, the structure, physiology and genetic integrity of organoid can be stably maintained over several months in in vitro culture [16]. This study uses the liver organoid definition reported by Huch and Takebe, that liver organoids are derived from single type cell culture and multi-type cell co-culture [15, 17]. Primary liver tumor cells can also be used for establishing liver tumor organoid [18].

Development of organoid technology has significantly changed liver-related research [19]. In this study, formation strategies of liver organoid (single type cell culture and multi-type cells co-culture) and the wide range of applications in regenerative medicine, organogenesis, liver regeneration, disease modelling, drug screening and personalized treatment are explored and summarized.

Three factors of liver organoid formation

Liver organoid technology mimics the structural and physiological characteristics and basic tissue-level functions of liver tissue in a dish [4]. The process of organoid formation mainly involves self-organization of cell population [5]. Cell population can self-organize and spatially rearrange to ordered structure under the uniform signaling environment. This process significantly mimics liver development process [20]. During the liver organogenesis, bipotent liver progenitors (hepatoblasts) with the potential of differentiating into hepatocytes and cholangiocytes, undergo morphogenetic processes (cell shape changes, cell proliferation and migration) to form liver buds in the signaling environment, including hepatocyte growth factor (HGF), Wnt, bone morphogenetic protein (BMP) and fibroblast growth factor (FGF) [12, 21, 22]. The culture environment, signal and starting cell types play indispensable roles in formation of liver organoid [5].

Culture environment provides 3D extracellular matrix (ECM) with characteristics of the body [23]. ECM comprises a network of extracellular molecules which provides 3D microenvironment and biochemical signaling to support cell adherence and growth [24]. Hydrogels are commonly used to support liver organoid culture [23]. Commercial Matrigel, a natural ECM extracted from Engelbreth-Holm-Swarm (EHS) mouse sarcoma, is the most widely used matrix for generating liver organoids [25]. Moreover, ECM hydrogel derived from decellularized tissues (DT), such as decellularized liver and small intestine matrix have been examined to support liver organoid culture [26–28]. Natural hydrogels have several advantages such as mimicking the structure and microenvironment of ECM, and maintaining a milieu of proteins, growth factors and cytokines, which provides mechanical, biophysical, and biochemical cues to cells [29, 30]. However, complex and variable constituents are not chemically defined, making it difficult to control the consistency and reproducibility of large-scale organoid production [26]. Furthermore, Matrigel is derived from EHS mouse sarcoma, thus limiting the progress of downstream organoid development processes and clinical applications [23]. Chemically defined hydrogels have been introduced to support intestinal, cerebral and liver organoid culture [31, 32]. Synthetic hydrogels have several advantages, including controllable mechanical properties, consistency, reproducibility and Good Manufacturing Practices (GMP)-compliance. However, they are limited by lack of natural matrices and biological signals [33]. Gjorevski et al. reported that Polyethylene glycol (PEG) hydrogel, can mimic physical and biochemical characteristics of liver microenvironment by regulating the mechanical properties thus supporting intestinal organoid generation [31]. PEG hydrogels are easily modifiable with collagen, fibronectin, laminin and RGD peptide to improve biological functions [34].

During the process of liver organoid formation, the starting cell population begins to assemble itself in the specific signaling environment [8]. However, exogenous signals that are associated with liver development should be provided to trigger self-organization [9]. Liver bud is formed from the foregut endoderm by induction of Wnt, FGF, BMP and Activin/Nodal signals [12]. Overactivation of Wnt pathway promotes anterior endodermal progenitors to adopt hindgut fates, resulting in failure of liver generation [35]. Moreover, different signals can activate or inhibit the signaling pathways of organ development to direct the differentiation direction of pluripotent stem cells (PSCs) in vitro [36].

The starting cell types determine the characteristics of the final liver organoid. Liver organoids are derived from PSCs, adult stem cells (ASCs), PHs and primary cholangiocyte (PCs) or iPSC-derived hepatic endoderm cells (iPSC-HEs) co-cultured with mesenchymal stem cells (MSCs) and umbilical cord-derived endothelial cells (HUVECs) (Fig. 1) [15, 17, 37–39]. In addition, liver tumor organoids can be generated from primary liver tumor cells [40]. The different starting cells follow different paths, therefore, choosing the appropriate starting cell types is important in biomedical applications.

Fig. 1.

Strategies for formation of liver organoids in vitro. The strategies for formation of liver organoids can be divided into single type cell culture and multi-type cells co-culture. In the single type cell culture, iPSCs, ASCs, PHs and PCs are embedded into Matrigel. Liver organoids can be formed after several days in presence of factors such as Rspo1, EGF, FGF, HGF and Nicotinamide. In multi-type cells co-culture, iPSC-HEs, are cocultured with HUVECs and MSCs, then plated onto dishes with pre-solidified Matrigel to form LBs after 48 h

Single type cell culture

Liver organoids formed from single cell type ensures proliferation and self-organization of homogeneous cell populations. iPSCs, ASCs, PHs and PCs are used as starting cells to form liver organoids. Liver organoids formed from iPSCs and ASCs often should be further differentiated into mature hepatocyte or hepatobiliary organoids.

iPSCs

In 2006, Yamanaka transferred Oct3/4, Sox2, c-Myc and Klf4 into differentiated fibroblasts and successfully obtained reprogrammed pluripotent stem cells with high similarity to embryonic stem cells in terms of gene and protein expression profiles, proliferation and differentiation characteristics [41, 42]. iPSCs has unlimited self-renewal and can be differentiated into a variety of cell types using specific differentiation protocols. Therefore, iPSCs are considered to be the best effective source of donor cells in regenerative medicine [43]. In the process of embryonic development, gastrulation is important. The gastrula further differentiates into three germ layers such as endoderm, mesoderm and ectoderm under mediation of specific signal factors (Wnt, transforming growth factor-beta ligands and FGF). Liver progenitors arise from endoderm [12, 44]. In addition, iPSCs can be differentiated into definitive endoderm (DE) to mimic liver development [45].

Mun et al. successfully established liver organoids by iPSCs with long-term expansion and stable mature hepatic characteristics. iPSCs differentiated into endoderm within 25 days and further exhibited hepatic maturation in alternate hypoxia and normoxia conditions. Liver organoids were then formed over 2D monolayers of mature hepatocytes. The floating cysts were collected and embedded in Matrigel for culture and amplification. In this stage, they retained properties of stem cells and showed hepatic characteristics, therefore, the liver organoids were successively passaged and expanded in the expansion medium (EM). For further maturation, the liver organoids were incubated in differentiation medium (DM), After differentiation, the organoids showed more mature characteristics of functional hepatocytes [36]. Several studies report EpCAM as a marker of human liver stem/progenitor cells [16]. Akbari et al. generated liver organoids using iPSC-derived EpCAM⁺ endodermal cells as the starting cells. Liver organoids were generated within 2 weeks. On maintaining the expansion efficiency and differentiative capacity, the liver organoids were expanded over 16 months. The liver organoids were transferred into DM for differentiation into mature hepatocytes. Analysis after differentiation showed that they had characteristics of functional hepatocytes in vitro and after intrasplenic injection into NSG mice with liver damage, the symptoms associated with the damage were alleviated and human albumin (hALB) was detected at day 32 [46].

Ogawa et al. established cholangiocyte organoids by human iPSC-derived hepatoblasts. Hepatoblasts cocultured with OP9 stromal cells secreted NOTCH protein that mimicked JAG1/Notch signaling thus promoting efficient differentiation of cholangiocytes and 3D aggregates formation in Matrigel-coated plate through use of developmentally relevant cues (such as NOTCH signaling). The 3D aggregates were then embedded in a mixture of type I collagen and Matrigel (containing epidermal growth factor (EGF), HGF and TGFβ1, etc.) to generate organoids. The cholangiocyte organoids exhibited 3D ductal and/or cyst structures [47, 48]. In addition, Sampaziotis et al. established cholangiocyte organoids from iPSC-derived cholangiocyte progenitors (CPs). CPs were embedded in Matrigel to generate functional cholangiocyte organoids [49].

Previous studies report that homogeneous hepatocyte-like cell population derived from DE does not accurately mimic regulation of complex signals during liver development [50]. Takebe et al. performed co-culturing of multi-type cells to mimic cell–cell interactions in liver development, which generated the bud-like structure [51]. These findings will be further discussed under LBs. Asai et al. reported that although organoids cannot be generated without cell–cell surface contact, however, mesoderm-derived paracrine signals can support maturation of hepatocytes and improve hepatic functions [52]. Guan et al. reported that iPSCs orderly differentiated into DM, foregut and hepatoblasts to form liver organoids. After further maturation, the liver organoids showed the functions of mature hepatocytes and were surrounded by bile duct-like structures [53]. Wu et al. successfully generated hepatobiliary organoids. Studies on mice biliary tubulogenesis reported that biliary specification can be regulated by the NOTCH signaling which is activated by endothelial cells of mesoderm, and TGF-β signaling. iPSCs were simultaneously induced to form endoderm and partly to mesoderm by incubating them in media with 25% mTeSR (one of key factors; agonists of TGF-β signaling pathways), which delayed hepatic differentiation, but activated the biliary specification. The hepatoblasts were bidirectionally differentiated into hepatocyte-like cells and cholangiocyte-like cells to form hepatobiliary organoids in the DM (containing uniquely prepared cholesterol⁺ MIX). The mature hepatobiliary organoids exhibited hepato-biliary functions in vitro. The hepatic structure was not complete, however, biliary structures were complete 8 weeks after transplantation into nude mice [21].

ASCs

Small cells near the bile duct tree are activated into hepatic oval cells (this reaction is called oval reaction in rodent models but ductal reaction in primate models) when mature liver is chronically affected by noxious agents (such as toxins and viruses) [54, 55]. Hepatic oval cells characterized by little cytoplasm and small oval nuclei, and Lgr5 (leucine-rich-repeat-containing G protein-coupled receptor 5; the receptor for the Wnt agonists R-spondins) are expressed on the cell surface [54, 55], and are known as liver Lgr5⁺ stem cells (biliary epithelial-derived progenitor cells). Lineage tracing shows that this type of cell population mainly resides in the terminal branch of the bile duct tree in the liver with the potential of bidirectional differentiation [15, 56–59]. During liver repair, it can differentiate into hepatocytes and bile duct epithelial cells. Interestingly, liver Lgr5⁺ stem cells are absent in healthy adult liver, whereas they are significantly increased by Wnt-driven regeneration near the bile duct [56, 59–61].

Liver Lgr5⁺ stem cells of bidirectional differentiation potential play an important role in study of liver organs [16]. Huch et al. embedded Lgr5⁺ liver stem cells into Matrigel containing EGF, Rspo1 (the Wnt agonist), FGF10, HGF and nicotinamide to form and expand liver organoids. Liver organoids were expanded for more than 12 months. Differentiated liver organoids with mature hepatocyte markers (Cyp3a11, Fah, G6pc and Alb) were obtained by changing the culture conditions and inhibiting Notch and TGFβ signal transduction. The liver organoids were transplanted into fumarylacetoacetate hydrolase (Fah⁻/⁻) mice model which accumulates toxic metabolites in liver causing persistent liver damage, analysis showed that the liver damage was alleviated [15]. Schneeberger et al. established large-scale culturing methods of liver organoids. Lgr5⁺ human liver stem cells were expanded and differentiated in spinner flasks containing Rspo1 and 10% Matrigel. After 6 weeks, a total amount of 10¹⁰– 10¹² cells were acquired from a starting number of 10⁶ cells [62]. Clevers et al. successfully achieved long-term expansion of human EPCAM⁺ ductal cells using the previously established in vitro culture system of single mouse-derived Lgr5⁺ hepatic stem cells. The embedded EPCAM⁺ ductal cells formed organoids in a Matrigel containing EGF, RSPO1, FGF10, HGF, Nicotinamide, Forskolin (FSK; a cAMP pathway agonist). The liver organoids could be passaged for over 6 months. After 3 months of culture, cloning liver organoids showed stable structure of chromosome and low rate of single base changes. When transplanted into nude mice administered with CCl₄-retrorsine to induce acute liver damage, liver organoids immediately became functional hepatocytes. Furthermore, organoids from α1-antitrypsin (A1AT) deficiency and Alagille syndrome patients could be expanded in vitro and mimicked the in vivo pathology [16].

PHs and PCs

Hepatocytes are the smallest functional unit in adult liver, and account for about 80% of liver parenchymal cells and are easily accessible cell populations from livers. These characteristics of hepatocytes make them suitable to be used as starting cells for formation of liver organoids [63]. Studies report that after liver damage, mainly acute liver damage such as partial hepatectomy (PHx), hepatocytes are mainly involved in liver regeneration rather than stem cells [64]. A study on murine liver regeneration reports that only hepatocytes are involved in liver regeneration after liver damage [65]. Clevers et al. isolated and embedded mouse Axin2-positive and human mature PHs in Matrigel to form liver organoids with a series of small molecule inducers such as RSPO1, EGF, FGF7, FGF10, HGF and TGF-β inhibitor. Wnt/Rspo1 and HGF signaling pathways are implicated in regulation of PHs expansion. Analysis showed that the mature hepatocyte derived organoids (Hep-Orgs) had a ‘‘bunch-of-grapes’’ appearance and albumin (ALB) secretion was only two–fourfold lower compared with that of PHs. Hep-Orgs had compact structure and retained key functions and gene expression profiles of hepatocytes compared with cholangiocyte derived organoids (Cholo-Orgs). Furthermore, after more than 2–3 months of culture, the rate of expansion of Hep-Orgs derived from murine and adult biopsies became significantly slow, and was associated with telomeres becoming shorter [38]. In 2018, Peng et al. reported that tumor necrosis factor (TNF)-α, an injury-induced inflammatory cytokine, may be implicated in expansion of liver organoids. Isolated mice PHs were embedded in growth-factor reduced (GFR) Matrigel with EM containing EGF, HGF and TNFα. At 2 weeks after seeding, organoids were formed. The organoids exhibited stable characteristics, and could serially be passaged for over 6 months [66]. Primary hepatocytes were recently isolated from different stages of nonalcoholic steatohepatitis (NASH) mouse model for establishment of NASH liver organoids. The study reported that expression of lipid metabolism-related genes was not upregulated, whereas expression of fibrosis-related genes was upregulated. These findings implied that the NASH liver organs reflected fibrosis rather than lipid metabolism. Moreover, the NASH liver organs expressed some liver-specific markers [67]. Tysoe et al. isolated PCs from common bile duct (CBD) to establish cholangiocyte organoids. The PCs were embedded in Matrigel with Rspo1, EGH and DDK-1 (a canonical WNT/β-catenin pathway antagonist; to prevent amplification of ASCs). Combination of Rspo1 and DKK-1 inhibits canonical and improve non-canonical WNT signaling, thus allowing large-scale, long-term generation of functional and genetically stable cholangiocyte organoids [68].

Multi-type cells co-culture

Liver organoids formed from multi-type cells co-culture benefits from vascular structure, self-condensation and self-organization of multi-type cell populations. LB are derived from hepatic endodermal cells with endothelial cells surrounding septum transversum mesenchyme population. Multi-type cells co-culture ensure cell–cell contact and interaction during liver development, which accurately mimics the complex regulation of signals.

iPSCs, HUVECs and MSCs

To accurately simulate the niche of hepatocytes, interaction between parenchymal cells and non-parenchymal cells in liver physiology and pathophysiology should be mimicked. However, it is difficult for liver organoids to generate a complex three-dimensional vascularized structure in dishes. Takebe et al. co-cultured iPSC-HEs with HUVECs and MSCs (iPSC-HEs: HUVECs: MSCs = 10:7:2) and plated them into dishes with pre-solidified Matrigel to form LBs. After 48 h, these multi-type cell populations spontaneously self-organized into macroscopically visible 3D cell aggregates. The LBs showed the endothelial network and expression of hepatic specific marker genes. They were connected with the FGF and BMP pathways activated by paracrine of cell–cell and stromal interactions. After further expansion, the entire LBs quickly connected with recipient vasculature to form functional vascular networks within 48 h of transplantation, which stimulated maturation of LBs. ALB was detected in the bloodstream at 10 days. Moreover, LBs showed therapeutic abilities after implantation into mice with drug-induced lethal liver failure [17]. Takebe et al. reported that a variety of cells such as liver, kidney, lung, intestine, brain, heart and cancer (HepG2) cells can be co-cultured with HUVECs and MSCs to generate complex vascularized organ buds [13]. The study reported that MSC-driven contraction and substrate matrix stiffness were important for self-condensation. During the process of MSC-driven contraction, cells movement is driven by active myosin II (MII); the contractive force produced by actomyosin cytoskeleton, which directs self-condensation. In addition, mm-sized large LBs can form on the substrate with E ~ 10 kPa, whereas on the too soft or too stiff substrate they form smaller aggregates or 2D sheets, which are connected with the relation of cell–cell and substrate force [10]. Takebe et al. developed large-scale LBs production platform for clinical applications. The microwell array plate (the dimple is triangularly arranged at 30 μm, the aperture diameter is 500 μm and the depth is 400 μm), can be up to 20,000 micro spots. The platform can produce vascularized and functional LBs reaching clinical scales (> 10⁸) [69]. Zhang et al. regenerated iPSC-derived CDX2⁺ hindgut endoderm cells (PGECs) based on robust propagating potential of endoderm progenitors. Subsequently, HUVEC and MSC were co-cultured with PGECs to produce PGEC liver bud (PGER-LBs). PGER-LBs exhibited vascularized network and successfully rescued liver-failure mice after transplantation [70].

Hepatocytes and stromal cells

Fresh PHs have several advantages in cell characteristics and functionality, therefore, they are the best choice for regenerative medicine, study of liver disease and drug screening. Takebe et al. co-cultured human fetal liver cells (hFLCs) with HUVECs and MSCs to generate liver organoid for recapitulating liver organogenesis in vitro. The hFLCs, HUVECs and MSCs (hFLCs: HUVECs: MSCs = 2:1:2) were plated into nontreated dishes. After 4 days of culture, a macroscopically visible cell cluster (100–200 μm) was observed [71]. Leite et al. generated a novel coculture liver organoid model to explore liver fibrosis. In this organoid model, human hepatic stellate cells (HSCs) were isolated and cocultured with HepRG cells. During the 21 days culture, the liver organoids showed stable morphology and hepatic function implying that they can be used for long term toxicity testing [72]. A large-scale (> 10¹⁰) and rapid (24 h) protocol for generation of porcine hepatocyte organoids from porcine PHs coculturing with HUVECs (PHs: HUVEC = 100:1) was developed using rocked suspension culture techniques. Analysis showed that over 94% of organoids were around 100 μm diameter, cell viability of organoid was more than 98% and metabolism, synthesis and detoxification of organoids was higher compared with that of PHs. These findings showed the crucial role of HUVECs in supporting PHs [73]. Takebe reported that when hFLCS were cocultured with HUVECs and MSCs, hFLCs were wrapped by HUVECs at several hours after seeding [71]. Furthermore, Asai et al. reported that when HUVECs are cocultured with iPSC-HEs, HUVECs quickly form tubular structure surrounded by iPSC-HEs. In addition, HUVECs may induce hepatic differentiation and promote hepatic functions by secreting paracrine soluble factors. Coculturing MSCs with iPSC-HEs shows similar capacity as HUVECs [52]. Recent organogenesis studies report that mesenchymal and endothelial cells play indispensable role in liver development. These cells regulate the fate of liver progenitors [12]. Rat hepatocyte organoid was established by coculturing rat PHs with MSCs onto dishes with pre-solidified decellularized liver ECM hydrogel. At 48 h after seeding, PHs and MSCs at a ratio of 5:1 efficiently formed liver organoids by MSC-driven condensation. Studies report that decellularized liver ECM hydrogel can contribute to self-condensation of organoids and survival and hepatic functions of organoids can be maintained for at least 20 days [27].

Primary liver tumours

Primary liver cancer (PLC) is a highly aggressive malignant tumor. PLC can be classified into two major classes including hepatocellular carcinoma (HCC) or cholangiocarcinoma (CC) and combined hepatocellular-cholangiocarcinoma (CHC) subtype [74, 75]. Currently, common tumor models include patient-derived cancer cell lines (PDC) and patient-derived xenografts (PDX) [76]. PDC models are 2D models, therefore, they lose the heterogeneous characteristics of tumor cells when cultured in vitro. Moreover, 2D culture cannot mimic 3D tissue structure and simulate the tumor microenvironment, resulting in differences in drug screening in vivo [77]. PDX models have limitations including low engraftment efficiencies of patient tumors, high cost, taking a long time to be developed and are labor intensive. Furthermore, PDX derived from immunodeficient mice, may undergo mouse-specific tumor evolution and may not be effective for high-throughput screening. Advances on the organoid technique can be used to circumvent these limitations [78]. Patient-derived tumor organoid (PDO) models are 3D cell culture model in vitro, which highly simulates the structure and functions of tumor tissue of body and demonstrate cell-to-cell and cell-to-matrix interactions. PDO has similar pathophysiological characteristics to differentiated tumor tissue in vivo. PDO model have recently shown a great prospects in drug screening and personalized medicine, including colon, stomach, pancreas, lung, breast and taste buds cancers [18, 40, 76].

A study by Huch et al. is the most representative research in liver cancer organoid development. Broutier et al. successfully established long-term human PLC organoids with three subtypes including HCC, CC and CHC. PLC cell populations were derived from liver resection by digestion. Growth of noncancerous cells was inhibited by isolation medium excluding Rspo1, Wnt3a and Noggin, but supplemented with dexamethasone and Rho kinase inhibitor. PLC organoids were formed at least 2 weeks after embedding onto Matrigel and further expanded for 1 year. Morphology of PLC organoids was different from healthy liver organoids but similar to that of liver cancer tissue in vivo. Furthermore, these organoids stably maintained the histological characteristics and gene profile of parental tumors in long-term culture and showed generalized intertumoral heterogeneity. A total of 11 specific biomarkers were identified in the three subtypes of PLC and the ERK inhibitors were validated using the xenotransplantation model [79]. Li et al. established several human PLC organoids derived from a liver cancer resection tissue which was cut into several tissue slices. Study of drug effects of organoids showed that intratumor heterogeneity affected the drug response of the organoids [80]. Saito established PDO derived from Biliary tract carcinomas (BTCs) and used it for drug screening. The organoids which were stably expanded for over 1 year and histopathology characteristics and gene profiles were maintained [81].

Biomedical applications of liver organoids

Organoids models have gained exponential interest as alternative models for biomedical research. Liver organoids have broad prospects in basic research and therapeutic applications including regenerative medicine, organogenesis, liver regeneration, disease modelling, and drug screening and personalized therapy fields. Different liver organoids have been generated for different applications (Fig. 2).

Fig. 2.

Biomedical applications of liver organoids. Liver organoids are ideal seeds for liver organoids transplantation in regenerative medicine. Liver organoids are combined with BAL for treatment in vitro. Liver organoids and scaffolds are used to construct engineered implantable tissues. Liver organoids are promising models for understanding organogenesis and liver regeneration. Organoids derived from hepatopathy can be used for disease modelling in vitro. Furthermore, liver cancer organoids can be used for drug screening and development of personalized treatment

Regenerative medicine

Liver disease is a major health burden, which accounts for approximately 2 million deaths per year worldwide [3]. Currently, orthotopic liver transplantation is the only available treatment for acute liver failure (ALF) and end-stage liver disease (ESLD) such as liver cirrhosis [82]. However, shortage of transplantable organs limits liver transplantation. Furthermore, after transplantation, patients have to undergo long time immunosuppression therapies [83]. Alternatives therapy approaches to orthotopic liver transplantation have been sought through development of regenerative medicine. PHs and stem cell transplantation, bioartificial liver (BAL) devices and engineered implantable tissues are alternatives for treatment of liver disease patients. Livers have high innate capacity of regeneration, therefore, they are suitable for regenerative therapy [84]. Furthermore, combination of liver organoids with other therapeutic approaches shows better therapeutic prospects. For instance, liver organoids are combined with BAL in vitro and liver organoids and scaffolds are used to construct engineered implantable tissues.

A study by Huch et al. reported that mice Lgr5⁺ liver stem cell-derived organoids can be transplanted into Fah⁻/⁻ mutant mice (tyrosinemia type I liver disease models) to explore the therapeutic effect of organoids. The Fah⁻/⁻ mutant mice died when not administered with NTBC (2-(2-nitro-4-trifluoromethylbenzoyl)-1,3-cyclohexanedione). The liver organoids underwent expansion and differentiation into mature hepatocytes in vitro. Although the survival of mice has increased 2 months after transplantation, the Fah⁺ nodules merely occupied 1% of liver mass. The low engraftment capacity can affect treatment efficacy [15]. Further, Huch et al. transplanted human EPCAM⁺ ductal cell-derived organoids into nude mice treated with CCl₄ to induce acute liver damage. hALB and a-1-antitrypsin were detected in serum of mice 7 days after transplantation. The level of hALB and a-1-antitrypsin was stably maintained for 120 days. Notably, the levels of hALB and a-1-antitrypsin were approximated within one month compared with human PHs transplantation 16]. Peng et al. recently reported that mice PH-derived organoids have high engraftment capacity. The Fah⁺ cell population increased up to 80% 103 days after transplantation into the Fah ⁻/⁻ mutant mice[66]. Hu et al. established human PH-derived organoids which were transplanted into immunodeficient Fah⁻/⁻ NOD Rag1⁻/⁻ Il2rg⁻/⁻ (FNRG) mice. The organoids graft showed rapid expansion 30 days after transplantation at a similar rate with that of PHs. The hALB of serum had over 200 mg/ml on average at 90 days after transplantation [38].

ASCs and PHs are harvested by invasive methods such as biopsy from adult liver. Use of iPSCs can circumvent this problem, because iPSCs are derived from clinical samples through noninvasive methods and have high expansion capacity in vitro [10]. Human ALB⁺ hepatocytes showed liver parenchyma cells and inter lobular veins 32 days after transplantation of iPSC-derived hepatocytes organoids into nude mice with acute liver damage [46]. Multiple duct-like structures were detected 6–8 weeks after transplantation of iPSC-derived cholangiocytes organoids into the mammary fat pad of NOD-SCID-IL2rγ ⁻/⁻ (NSG) mice [47]. In addition, biliary duct-like structures and large hepatic clusters are observed 4 weeks after iPSC-derived hepatobiliary organoids were transplanted into nude mice [21]. Takebe et al. established LBs which were mechanically stable and could be manipulated physically, thus ensuring modification of the transplant procedure and ectopic transplants. LBs showed consistent maturation and vascularization in vivo after transplantation into nude mice with liver injury. These characteristics ensures the viability of LBs and development of hepatic functions of LBs. LBs showed hepatic cord-like structures similar to those found in adult liver after 60 days and improved survival of mice [17]. Porcine liver organoids for spheroid reservoir bioartificial liver (SRBAL) were established for treatment of ALF monkeys administered with α-amanitin. Survival of ALF monkeys improved after treatment with liver organoids SRBAL compared with nontreated ALF monkeys. After 6 h treatment, blood ammonia and total bilirubin were reduced and albumin levels were increased. Porcine endogenous retrovirus were not identified in monkey liver or blood and the liver of treated monkeys showed timely liver regeneration leading to recovery [73]. Tysoe et al. generated functional bioengineered biliary tissue by combining cholangiocyte organoids and polyglycolic acid (PGA) scaffolds for biliary reconstruction. The cholangiocyte organoids were seeded into PGA scaffolds to establish bioengineered biliary tissue for 2–4 weeks [68]. These findings show that liver organoids have potential uses in regenerative medicine.

Organogenesis and liver regeneration

Embryonic organogenesis and adult liver regeneration are regulated by interaction of several signaling pathways [2, 12]. Understanding the mechanisms of these signaling pathways is crucial for alleviating abnormal liver development and for overcoming obstacles of liver regeneration [84]. Animal model studies report several significant mechanisms of liver development and regeneration, however, these mechanisms do not reflect those in humans due to inter-species differences [10].

LBs generated from the posterior foregut endoderm by hepatoblasts during liver development are induced to undergo morphological changes, proliferation and migration by Wnt, FGF, HGF and BMP signals [8]. Moreover, hepatoblasts further proliferate and differentiate to generate hepatocytes and biliary epithelia. Hepatic fibroblasts and stellate cells are generated by mesoderm-derived mesenchyme. Mesenchyme and endothelium assemble a complex structure and undergo hepatic maturation to form liver tissue. Liver organoids can reconstruct these major phases of organogenesis in experimental set ups [4]. Development of iPSCs that differentiate into hepatocyte or cholangiocyte-like cells in 2D, enables iPSC-derived liver organoids (a 3D model) to reconstruct the fundamental processes of liver development. iPSCs are often induced to differentiate into DE based on the knowledge of endoderm organogenesis to mimic liver development. Furthermore, iPSC-derived liver organoids undergo differentiation into mature hepatocytes, cholangiocyte and hepatobiliary organoids by exposure to specific signalling pathways. These processes mimic maturation of liver in vivo [10]. Takebe et al. established LBs by co-culturing iPSC-HEs with HUVECs and MSCs to mimic liver development. During the process of self-organization, the endothelial network formed and was homogenously distributed in the LBs. Cell-to-cell and stromal interactions can activate FGF and BMP pathways that are important in liver development [17]. However, from developmental perspective, elaborate communication among specialized endothelial and mesenchymal cell populations are curial for the formation of LBs. To accurately simulate the normal liver development in vitro, hepatic endoderm, mesenchymal and endothelial progenitors were derived from iPSCs to form LBs [69]. The LBs were transplanted into mice and maturation of LBs was stimulated through haemodynamic process [13, 85]. Therefore, LBs strategy is a valuable method for studies on organogenesis in vitro.

After PHx and chemical or infectious injury, the liver shows spontaneous regenerative capacity [86]. Biliary epithelial-derived progenitor cells respond to the chronic liver damage but mature hepatocytes respond liver damage from causes such as partial PHx [87, 88]. Huch et al. isolated Lgr5⁺ liver stem cells and generated long-term expanded liver organoids which retained bidirectional differentiation activity. The study reported that Rsop1, EGF, FGF and HFG are important for formation and expansion of liver organoids. Notably, liver organoids and liver tumor organoids were generated from healthy human primary tissue and PLC through this protocol [15, 16, 79]. Hu et al. explored importance of Rspo1 and HGF in PH-derived liver organoids whereas Peng et al. reported that TNFα plays key roles in liver regeneration [38, 66]. Tysoe et al. reported that DKK-1 is important for long-term expansion of cholangiocyte organoids [68]. These findings show that liver organoids strategy is a valuable and convenient method for understanding organogenesis and for exploring the process of liver regeneration.

Disease modelling

Liver is implicated in several metabolic functions such as oxidation, storage of hepatic glycogen and secreted protein synthesis. Metabolic abnormalities lead to liver diseases [89]. Establishment of a disease model in vitro can be used to effectively understand pathogenesis of disease and to explore effective treatments. Liver organoids model play important role in understanding pathogenesis of liver diseases and provide avenues for development of effective therapies.

A1AT deficiency is an inherited metabolic disease characterized by deficiency of A1AT in serum [90]. Alagille syndrome is characterized by lack of interlobular bile ducts, which leads to severe cholestasis and progression to cirrhosis [91]. Huch et al. successfully established liver organoids from liver biopsies of patients with both A1AT deficiency and Alagille syndrome. Theses organoids mirrored the in vivo pathology. Organoids from A1AT-deficiency patients secreted high levels of ALB and showed intake of low density lipoprotein whereas organoids from Alagille syndrome patients showed defective structure of the biliary tree [16]. Wilson’s disease is an uncommon inherited copper storage disease [92]. A previous study reported that canine liver organoids from COMMD1-deficient dogs had a higher intracellular accumulation of copper, which was highly similar to Wilson's disease in vivo [93]. Citrullinemia type 1 (CTLN1) is a urea cycle disorder, caused by defects in argininosuccinate synthetase (ASS) enzyme. Akbari et al. generated liver organoids from patient-specific iPSCs. The organoids showed ammonia accumulation which is related CTLN1 disorder. Moreover, organoids rescued the phenotype of CTLN1 by ectopic expression of ASS1 enzyme [46]. Cystic fibrosis (CF) associated cholangiopathy, is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene [94]. Ogawa et al. and Sampaziotis et al. reported that patient-specific iPSC-derived cholangiocyte organoids could model CF in vitro [47, 49]. A1AT deficiency, Alagille syndrome, Wilson's disease, CTLN1 and CF are hereditary liver diseases [89].

Steatosis has become one of the most common chronic liver diseases in many countries sue to increasing prevalence of type 2 diabetes and obesity. Severe steatosis can progress to steatohepatitis [95]. Mun et al. established iPSC-derived liver organoids to model steatosis by over-treatment with oleate and palmitate [36]. With the organoids were incubated with free fatty acid, and after 4 days the iPSC-derived liver organoids showed lipid droplets. Organoids for modeling non-alcoholic fatty liver disease have been validated through molecular and genetic analysis [37]. Liver organoids, containing hepatocyte-, stellate-, and Kupffer-like cells, can be generated from iPSCs. Liver organoids on exposure to free fatty acids show the steatosis-, inflammation-, and fibrosis-like pathology [95]. A number of studies have successfully generated liver organoids to model liver fibrosis, viral hepatitis, cholestasis and alcoholic liver injury [19, 37, 73, 96].

Broutier et al. established liver tumor organoids (HCC, CC and CHC) from PLC. The organoids retained the histological characteristics, gene expression and genomic landscape of the original tissue. This finding implies that organoids exhibit intertumoral and intratumoral heterogeneity. In addition, organoids transplanted into nude mice showed characteristics of PLC implying that these features were preserved in vivo [79].

Drug screening and personalized treatment

Currently, only few models are available that accurately reproduce the pathophysiology of primary tumors in vitro, thus limiting development and application of clinical drugs [97]. The FDA reports that more than 90% of the safe and effective drugs effective in animal experiments fail in human clinical trials. The field of drug research is characterized by high input and low output implying that only a few drugs pass clinical trials [98]. Emergence of organoid technology is a promising approach in drug development. It is gradually used for preclinical drug screening and predicting different treatment responses of individual patients.

Liver organoids generated from iPSCs and ASCs are able to model various of disease in vivo, therefore it is possible to use organoids for drug screening and personalized treatment. Moreover, studies are building biobanks of healthy and diseased human organoids by cryopreservation for biomedical applications, which can be used to replace traditional 2D cell lines with PDX for drug screening [4]. Li et al. successfully established 27 liver cancer organoid lines from PLC. These organoids were tested using 129 cancer drugs and findings showed that they generated 3483 cell survival data points. This study reported that the most of drugs were ineffective, however, a small group of drugs were pan-effective [80]. Small scale and personalized organoids can be generated for personalized treatment through development of liver tumor organoids from PLC. Huch et al. reported that PLC-derived organoids can be used to identify potential novel therapeutic targets and can be used as an in vitro model for drug testing and personal medicine [8].

Conclusion

Several studies show that liver organoids have great application potential in the field of liver basic research and medicine. Organoid approach has several advantages as a new experimental model in vitro that fully reflects the specificity and complexity of human system compared with 2D cell culture and animal models [5]. Formation modes of liver organoids have diversified through establishment of long-term cultivation of hepatocytes in vitro, which can be simply summarized into two kinds of culture systems, namely single type cell culture and multi-type cell co-culture. In single type cell culture systems, iPSCs and ASCs are mainly used to form liver organoids. These cells are then differentiated into a high proportion of mature hepatocytes with high liver functions, such as albumin, urea synthesis and cytochrome P450 activity. Moreover, liver organoids can be generated from PHs and PCs with mature hepatic functions. Multi-type cell co-culture systems form vascularized and transplantable organoids and maintains the signaling and interaction between mesenchymal cells and parenchymal cells. Multi-type cell co-culture systems are used to create an intracellular niche for hepatocytes. Hepatocytes can be cocultured with stromal cells (HUVECs or MSCs) to establish liver organoids.

Currently, liver organoids are mainly used in basic research, and there are still some clinical limitations. For example, poor control of morphology and composition, significant differences between individuals and batches of organ samples and the risk of rejection by immune system. Furthermore, the modeling cycle is time-consuming and has low efficiency. Hepatocytes transplantation trials report that clinical cure requires at least need 10⁸ hepatocytes [69]. Therefore, there is need to optimize the culture system to achieve large-scale culture. PDO is a novel model recently reported and has become popular among researchers. The duration for the first passaging of liver PDO is about 4 weeks, and the time for preparation of sufficient amount of PDO by amplification for drug screening can be up to 12 weeks [79]. However, the time window for clinical neoadjuvant therapy and postoperative chemotherapy is usually 2–3 weeks. Therefore, optimizing the culture system of PDO and improving utilization efficiency of primary tumor cells to rapidly generate large-scale PDO are prerequisites for clinical application of the therapy. Further studies should explore these problems to promote development of culture technology of liver organoids to meet clinical needs. Although the culture systems of hepatic are not perfect, it is important to culture models for the study of liver development and related diseases. The findings of these studies show that liver organoids have great potential in liver research and treatment of liver diseases.

Author contributions

Xinglong Zhu, Bingqi Zhang and Ji Bao designed the research; Xinglong Zhu, Bingqi Zhang and Yuting He contributed to literature selection and data extraction; Xinglong Zhu and Bingqi Zhang drafted the original manuscript; All authors have read the final article and approved the publication of the manuscript.

Compliance with ethical standards

Conflict of interest

The authors have no conflicts to disclose.

Ethical statement

There are no animal experiments carried out for this article.