Synthetic human livers for modeling metabolic diseases

Abstract

Purpose of review

In this review, we will explore recent advances in human induced pluripotent stem cell (iPSC)-based modeling of metabolic liver disease and biofabrication of synthetic human liver tissue while also discussing the emerging concept of synthetic biology to generate more physiologically relevant liver disease models.

Recent finding

iPSC-based platforms have facilitated the study of underlying cellular mechanisms and potential therapeutic strategies for a number of metabolic liver diseases. Concurrently, rapid progress in biofabrication and gene editing technologies have led to the generation of human hepatic tissue that more closely mimic the complexity of the liver.

Summary

iPSC-based liver tissue is rapidly becoming available for modeling liver physiology due to its ability to recapitulate the complex three-dimensional architecture of the liver and recapitulate interactions between the different cell types and their surroundings. These mini livers have also been used to recapitulate liver disease pathways using the tools of synthetic biology, such as gene editing, to control gene circuits. Further development in this field will undoubtedly bolster future investigations not only in disease modeling and basic research, but also in personalized medicine and autologous transplantation.

INTRODUCTION

The liver plays a critical role in maintaining homeostasis by regulating the metabolism of carbohydrates, proteins, and lipids; synthesis of serum proteins, bile acids, and urea; detoxification of drugs, and metabolic waste products; and storage of vitamins and carbohydrates [1]. The vast majority of these processes are primarily mediated by hepatocytes, parenchymal cells that serve as the functional metabolic unit of the liver. Hepatocytes make up about 70% liver mass, with biliary epithelial cells, sinusoidal endothelial cells, hepatic stellate cells, and Kupffer cells making up the remaining mass [2].

Chronic liver disease affects more than 4.5 million people and is the 11th leading cause of death in the United States [3,4]. Disease can result from hepatitis viral infections, excessive alcohol intake, fatty liver disease, inherited metabolic diseases, autoimmune liver diseases, drug-related liver diseases, as well as malignancies, and idiopathic causes [5]. Because the liver performs vital regulatory functions, liver disorders are often lethal or require lifelong management. Currently, transplantation is the only definitive therapy for end-stage liver failure [6].

One of the biggest challenges in understanding the mechanisms behind the pathogenesis of liver disease and in developing novel therapies for the treatment of liver disease is the lack of model systems that accurately reproduce the physiological functions of the human liver in vitro. The degeneration from healthy-functioning organ to end-stage liver disease is a dynamic process that involves cellular damage, diminished function, and adaptation to inflammation; these changes may occur before or after transplantation. Failure of recovery and progressive deterioration are conditioned by environmental and genetic factors that govern the balance between injury and repair incited by cytokines, adipokines, bacterial products, and metabolites. Immortalized or cancer cell lines display metabolic enzyme profiles that differ from primary mature hepatocytes, exhibit dysfunctional apoptotic pathways, and lack the genotypic variability seen in the human population [7-9]. While the complex processes underlying disease have been elucidated in part through detailed analysis of inbred and genetically engineered animals, no animal model yet identified can faithfully represent the manifold genetic and environmental factors that decide the course of disease. Thus, while animal models have helped to detail numerous factors that contribute to development of disease, they have failed to replicate multiple aspects of human physiology. Simply put, mice are not humans. Because of this, many drugs initially shown to be effective in cell lines or animal models have been found to be ineffective or even toxic in humans during subsequent clinical trials [10-13].

Primary human hepatocytes have emerged as the gold standard for studying the biochemical and cellular mechanisms of liver disease and identifying novel pharmacological therapies [14]. However, these cells are not readily available because they can only be isolated from donor livers not used for transplantation, explanted livers after transplantation, or liver resection specimens. In addition, conventional two-dimensional in-vitro hepatocyte cultures are also limited by the model’s failure to duplicate the liver architecture and cellular heterogeneity, the lack of cell–cell and cell–extracellular matrix (ECM) interactions, the limited proliferative capacity of hepatocytes, and the rapid loss of liver-specific gene expression and function. These limitations make two-dimensional in vitro primary hepatocyte cultures unsuitable for a number of disease studies or drug screens [15-17].

HUMAN PLURIPOTENT STEM CELLS AS A MODEL FOR LIVER DISEASES

The advent and rapid improvement of induced pluripotent stem cell (iPSC) technology has paved the way for studies that describe the use of iPSCs for modeling a variety of diseases [18-20]. iPSCs are ideal for disease modeling and large-scale screening studies because they can be generated from a small sample of tissue or blood obtained from virtually any patient, including those with rare genetic disorders; have the potential to differentiate into almost any cell type; have unlimited proliferative capacity; and with proper handling, retain genomic and epigenetic stability [21]. Over the years, the development of techniques to differentiate human iPSCs into hepatocytes and other liver nonparenchymal cells using stepwise protocols that mimic in-vivo organogenesis in cultures has resulted in cells that closely resemble human liver cells [22-26,27^■■,28]. Using these techniques, hepatocyte-like cells (HLCs) have been produced and used to model a variety of metabolic liver diseases [28-34,35^■,36,37,38]. Below, we discuss some of these studies, specifically focusing on metabolic liver diseases that were extensively modeled and are highly prevalent in the population.

The classical form of alpha-1-antitrypsin deficiency (ATD) is a congenital liver disease resulting from the homozygous PiZ mutation in SERPINA1 which promotes spontaneous polymerization and intracellular accumulation of antitrypsin in the endoplasmic reticulum (ER) of hepatocytes. This buildup of mutant antitrypsin leads to cellular overload, hepatic cell death, fibrosis, and liver failure. A study aiming to model antitrypsin deficiency showed successful differentiation of antitrypsin deficient iPSCs into cells that displayed many features of hepatocytes. Most importantly, these antitrypsin deficient HLCs exhibited retention of the mutant antitrypsin within the endoplasmic reticulum, a key feature of this disease. Significantly, the study observed variability in the degree of antitrypsin accumulation among antitrypsin deficient HLCs obtained from different patients, a finding that was attributed to either variability in the hepatocyte differentiation efficiency or to differences in liver disease severity [25]. Factors, such as incidence, age of onset, and liver disease severity are considerably variable among antitrypsin deficient patients. Therefore, a study that employed biochemical and ultrastructural analyses of iPSC-derived HLCs taken from several patients that represented the varying degrees of liver disease severity was initiated. The study found that HLCs derived from antitrypsin deficient patients who presented with no liver disease were able to degrade and secrete mutant antitrypsin relatively efficiently. In contrast, HLCs obtained from patients with severe liver disease displayed delayed degradation of mutant antitrypsin resulting in the accumulation of the mutant protein in pre-Golgi compartments. Moreover, analysis of subcellular morphology showed the presence of globular inclusions only in HLCs from patients with severe liver disease. The study, therefore, demonstrated the utility of iPSC-derived HLCs in predicting disease susceptibility and outcome, which could be useful for future personalized clinical interventions [29]. In another study, zinc finger nuclease and piggyBac technology were used to genetically edit the antitrypsin PiZ mutation in antitrypsin deficient iPSCs. This footprint-free genetic correction restored the structure and function of the antitrypsin protein and prevented the accumulation of the mutant protein in the differentiated HLCs. This study provided proof of concept that genetic correction of disease-causing mutations in iPSCs in vitro can rescue disease phenotypes in the resulting HLCs [30]. Subsequently, similar studies using transcription activator-like effector nucleases have been reported. These studies opened the possibility of using genetically corrected iPSCs to generate HLCs that can be used for autologous transplantation [31,32].

Wilson disease is caused by mutations in the ATP7B gene that encodes a transporter protein responsible for copper export into bile and blood. The ATP7B protein is localized in the Golgi of hepatocytes where it regulates the transport of copper for incorporation into ceruloplasmin. However, in response to copper overload, it is shuttled to the endosome–lysosome compartment to transport copper into the bile canaliculi [33]. Wilson disease can have a wide range of hepatic and neural clinical phenotypes depending on the specific mutation but the mechanisms are not fully understood. A study that aimed to model Wilson disease described the generation of Wilson disease iPSCs bearing the R778L mutation in the ATP7B gene. Wilson disease HLCs exhibited typical features of the disease, namely cytoplasmic localization of the mutant ATP7B protein and abnormal copper transport. Copper transport in these Wilson disease HLCs was rescued, following either genetic correction of the mutated gene using lentiviral vectors or treatment with the chaperone drug curcumin. These findings highlight the model’s potential utility in drug screening and developing novel therapies for Wilson disease [34]. In another study, Wilson disease iPSCs carrying the H1069Q mutation were differentiated into HLCs. In contrast to previous reports from overexpression studies in cell lines, about one third of the mutant ATP7B was surprisingly localized in the Golgi and could freely move to the endosome–lysosome compartment during copper accumulation. Further analysis of mRNA and protein levels showed that while mRNA levels were normal, there was an 80% reduction in the protein levels of the mutant protein. The study concluded that ER associated degradation plays a major role for the loss of ATP7B function due to the H1069Q mutation [33]. A more recent study differentiated Wilson disease iPSCs with the H1069Q mutation into HLCs that exhibited canalicular-sinusoidal polarity in two-dimensional culture. Analyses of these polarized HLCs revealed that, similar to the previous study, the H1069Q mutation does not prevent the trafficking of the ATP7B to the Golgi. Instead, the mutation impeded the redistribution of the protein to the bile canaliculi after copper overload. This suggests that pharmacological chaperones alone would not serve as an effective treatment for this specific mutation [35^■].

The use of human iPSC-derived HLCs for modeling metabolic liver disease has showcased its extraordinary potential for understanding mechanisms of disease, identifying potential therapeutic drugs, and developing novel treatment strategies. However, the main gap that prevents advancement of the field is the lack of an ideal human organ model platform that recreates human liver cell interactions, immunological responses, and disease progression. Furthermore, ideal organ models should mimic the microstructure, extracellular cues, and vascularization of a human organ. These necessary components all play vital roles in the proliferation, differentiation, gene and protein expression, polarity, drug metabolism, migration, response to stimuli, and mechanics (elasticity and shape) of cells [39,40].

BIOFABRICATION OF SYNTHETIC HUMAN LIVER TISSUE

Biofabrication is a multidisciplinary field that combines the principles of engineering, biology, and material science to generate biologically functional products using cells, proteins, and biomaterials [41]. There has been great progress in biofabricating multicellular platforms to recapitulate human liver structure and functions. Two different approaches have been taken towards developing human liver tissue platforms: models in which the initial physical and cellular structure mimics the organization of cells in the liver acinus by design (bioprinting, micro liver microphysiology systems, natural liver extracellular matrix) [26,27^■■,42^■■] and models in which organoids are stimulated to self-assemble into liver-like structures [43^■■]. A recent study explored a novel approach in modeling hepatic, biliary, and pancreatic organogenesis in vitro by separately differentiating iPSCs into foregut and midgut spheroids. Afterwards, the foregut and midgut spheroids were fused and spontaneously formed an organoid comprised of cells expressing hepatobiliary and pancreatic markers without the addition of extrinsic factors. By using this multicellular model, researchers were able to replicate some aspects of cellular interactions and multiorganogenesis and model human developmental biology. Challenges remain, such as immaturity of some functions and other characteristics associated with mature human hepatocytes, and lack of flow through vascular channels within the liver organoids. However, the presence of heterotypic cells enables novel translational applications for understanding cell–cell interactions in disease modeling [44^■■,45^■].

Mini human livers have also been fabricated in recent reports [27^■■,42^■■]. Genetically engineered human iPSCs were differentiated into liver cells and, together with supporting nonparenchymal cells, were grafted into empty rat liver scaffolds and cultured in organ bioreactors under constant flow. The system mimicked many aspects of human fatty liver disease. The human iPSC-derived mini liver developed macrosteatosis, an acquired proinflammatory phenotype, and shared a similar lipid and metabolic profile to human livers with nonalcoholic steatohepatitis and terminal liver failure. In a following report, our group also generated mini human livers that contained hepatocytes, vascular, bile duct, and supportive (fibroblast, mesenchymal cells) cell types generated from human iPSCs. The organ-like microenvironment allowed for further maturation of human iPSC mini livers and the microstructure had similarities to human liver. Moreover, the human iPSC-derived mini livers were transplanted in immunodeficient rats and survived for up to 4 days. This approach highlighted the importance of vascular flow, anatomical structure, and cell interactions for liver maturation and function.

Progressive developments in biofabrication technology have led to the generation of three-dimensional models. In comparison with current two-dimensional systems, these three-dimensional models more closely resemble the complexity and heterogeneity of tissues. Biofabricated mini livers generated from human iPSCs have been shown to replicate many characteristics of the liver, including tissue architecture, vasculature, hepatocyte gene expression, and function. Future developments in the field would certainly lead to improvements in mature hepatocyte function, long-term survival, and long-term maintenance of hepatocyte phenotype in these systems.

The evolution of human tissue biofabrication has been based on the parallel development and application of transdisciplinary biotechnologies, including bioprinting, matrix materials, patient-derived primary cells, and iPSCs as well as synthetic biology tools, to edit genes to generate an optimized human tissue platform. Clustered regularly interspaced short palindromic repeats (CRISPR)/ CRISPR-associated protein 9 (Cas9) editing has proven to be extremely efficient for making targeted changes to the genome and has now been widely adopted to produce a variety of mutations, deletions, and genetic rearrangements in numerous biologic systems [46-49]. More importantly, CRISPR/Cas9, in conjunction with the use of large guide RNA (gRNA) libraries, can be used for genome-wide interrogation of gene function. For instance, CRISPR/Cas9’s powerful editing capability was demonstrated in a recent publication [50^■■] where human hepatocytes with a monogenic disease in ammonia metabolism (ornithine transcarbamylase [OTC]), were genetically corrected ex vivo, by deleting a mutant intronic splicing site. This was achieved with more than 60% editing efficiency and successfully restored the urea cycle in vitro. The corrected hepatocytes were also able to engraft and expand after transplantation in a xenograft animal model, suggesting that CRISPR/Cas9 editing did not affect proliferative capacity [50^■■].

It is well recognized that genetic variations affect the severity of disease phenotypes, as shown in antitrypsin deficiency [51]. To date, more than 125 single nucleotide polymorphisms have been reported for SERPINA1, making this gene highly pleomorphic. Given the high number of genetic variants that are being identified and their role in disease presentation, the utilization of CRISPR-Cas9 in conjunction with gRNA libraries represents an attractive approach. A recent report elegantly explored this approach by designing genome-wide CRISPR/Cas9-based screens using human mesenchymal precursors cells; these cells were differentiated from human embryonic stem cells carrying the pathogenic mutations that cause diseases associated with accelerated aging. The authors identified deficient genes that alleviated cellular senescence by deep sequencing and demonstrated the therapeutic effect of the identified hits. This effective approach demonstrated that CRISPR/Cas9-based genetic screening is a robust method for systematically uncovering genes, which may represent a therapeutic target [52^■■]. Future scenarios may utilize gene editing to generate specific patient iPSCs under different genetic backgrounds. Then, liver cells can be differentiated in vitro to fully examine putative genetic modifiers using gRNA mini-libraries to study the response to candidate drugs for the treatment of ATD.

CONCLUSION

Human in vitro models using cancer cell lines fail to recapitulate key aspects of liver diseases. Two-dimensional cultures of primary human hepatocytes are also limited by the lack of donor livers and the inability to replicate liver architecture, cellular heterogeneity, and cell–cell and cell–ECM interactions. Human iPSC-based genetic engineering and generation of mini livers have emerged as novel tools that enable researchers to model the complexity of liver diseases. For example, perturbations in the interactions between different liver cell types and changes in the cellular microenvironment and architecture, that would otherwise be impossible to reproduce in two-dimensional cultures, are key features in these models. In addition, mini livers can serve as a platform for drug development and testing. Figure 1 provides a schematic of the potential application of iPSC-derived mini livers. The next-generation of synthetic human liver tissue derived from patient cells will incorporate controlled variation of experimental parameters, extensive use of real time imaging modalities, metabolomics, and a wide array of real-time cell function readouts that should be correlated with disease progression to enable the identification and validation of prognostic and therapeutic markers. Improvements in hepatic differentiation protocols, generating synthetic scaffolds, and incorporating synthetic biology tools, make it likely that future iPSC-mini livers will soon resemble human livers, facilitating their future use in clinical regenerative and precision medicine.

FIGURE 1.

Biofabricated liver tissue and organoids derived from human induced pluripotent stem cells are novel tools that recapitulate key components of human liver structure and function. Utilization of clustered regularly interspaced short palindromic repeats/clustered regularly interspaced short palindromic repeats-associated protein 9 technology and gRNA libraries allows for genetic correction or screening of mutations in these tissue or organoid cells. Deep sequencing of their genomes using NGS may systematically identify potential therapeutic targets or genetic modifiers that can be used to improve clinical trial designs. Future clinical trials can use these genetically modified induced pluripotent stem cell-mini livers to improve precision medicine and treat specific patient populations. gRNA, guide RNA; NGS, next-generation sequencing.

KEY POINTS.

• iPSC-derived hepatocytes serve as a useful platform for studying the pathophysiology and potential therapies for metabolic liver diseases.

• Biofabricated mini livers and organoids from human iPSCs replicate the tissue architecture, cellular heterogeneity, and cell–cell interactions of liver tissues.

• Clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 functional analysis offers a powerful tool for establishing connections between genetic variations and biological phenotypes.

• Biofabricated mini livers generated using CRISPR/CRISPR-associated protein 9-edited human iPSCs can be a powerful tool for disease modeling and drug testing.