The Progress and Promise of Lineage Reprogramming Strategies for Liver Regeneration

Abstract

The liver exhibits remarkable regenerative capacity. However, the limited ability of primary human hepatocytes to proliferate in vitro, combined with a compromised regenerative capacity induced by pathological conditions in vivo, presents significant obstacles to effective liver regeneration following liver injuries and diseases. Developing strategies to compensate for the loss of endogenous hepatocytes is crucial for overcoming these challenges, and this remains an active area of investigation. Lineage reprogramming, the process of directly converting one cell type into another bypassing the intermediate pluripotent state, has emerged as a promising method for generating specific cell types for therapeutic purposes in regenerative medicine. Here, we discuss the recent progress and emergent technologies in lineage reprogramming into hepatic cells, and their potential applications in enhancing liver regeneration or treating liver disease models. We also address controversies and challenges that confront this field.

Summary.

This review provides an overview of recent advancements in in vitro and in vivo hepatic lineage reprogramming strategies within the context of liver regeneration. Additionally, it discusses the controversies and challenges facing the field and proposes future directions for lineage reprogramming to treat liver diseases.

The liver is a pivotal organ that not only performs various critical metabolic functions essential for maintaining physiological homeostasis in mammals but also possesses an extraordinary regenerative capacity.¹ However, liver regeneration can be compromised by numerous pathogenic conditions such as metabolic disorders, hepatotoxicity, and viral infections. Without timely intervention, these liver pathologies may lead to acute liver failure or end-stage liver disease (ESLD), both of which have become leading causes of death worldwide and impose a significant health burden on millions of people each year. Treatment options for ESLD and acute liver failure are so limited that orthotopic liver transplantation often remains the only curative therapy.² However, a global shortage of donor organs restricts the widespread application of this treatment. Human hepatocyte transplantation presents an attractive alternative to liver transplantation, as its safety and efficacy have been demonstrated in animal models and in patients with several severe liver diseases.^3,4 Over the past decades, various strategies aimed at providing adequate functional human hepatocytes for cell-based therapy have been developed.5, 6, 7, 8 Among these, lineage reprogramming strategies raised particular interest in the field of liver regeneration due to their significant potential to generate hepatocytes both in vitro and in vivo (Figure 1), along with a reduced risk of tumorigenesis.⁹

Figure 1.

Historical view of the development of lineage reprogramming into hepatic cells. Selected advances in the development of hepatic lineage reprogramming are highlighted in different colors. Cyan indicates studies conducted in vivo, and green indicates studies conducted in vitro. GECs, Gastric epithelial cells.

In this review, we discuss the current advances in generating hepatocytes via transcriptional factor (TF)- and chemical compound-mediated lineage reprograming methods in vitro and examine the rapidly advancing in vivo lineage reprograming techniques in recent years. In particular, we focus on strategies that promote the endogenous generation of hepatocytes through partial reprogramming, clustered regularly interspaced short palindromic repeat (CRISPR)- and mRNA-based technologies, and the transformation of ectopic sites into liver-like organs, all within the context of liver regenerative medicine. We also highlight the controversies and challenges facing the field and propose future directions for lineage reprogramming in the treatment of liver diseases. Together, these advancements offer the potential for an optimal solution in regenerative medicine: reprogramming to address liver injuries and potentially cure liver diseases.

In Vitro Lineage Reprogramming to Hepatic Fate for Liver Regeneration

Unlike classical reprogramming, which converts initial cells into induced pluripotent stem cells (iPSCs), lineage reprogramming typically involves overcoming pre-existing epigenetic signatures in the initial cells and directly forcing the establishment of the desired cell identity, bypassing the iPSC stage.¹⁰ Due to the absence of the pluripotent stage, lineage reprogramming could theoretically reduce cancer risk after transplantation of reprogrammed cells. Two established methods, defined TF- and chemical compound-mediated lineage reprogramming have been most commonly used to generate hepatic cells from different initial cells in vitro (Figure 2).

Figure 2.

Different in vitro lineage reprogramming strategies to generate hepatic cells for liver regeneration. Different lineage reprogramming strategies allow for the generation of hepatocyte-like cells or hepatic progenitor cells directly or indirectly for liver transplantation. The direct lineage reprogramming is based on the use of reprogramming factors, such as TFs or small molecules to induce the initial cells, such as fibroblasts, primary hepatocytes, or other endodermal epithelial cells into hepatic cells directly. Although the indirect lineage reprogramming relies on the reprogramming factors to reach a proliferative plastic intermediate state, which is followed by conversion into hepatic lineages under the appropriate developmental cues.

TF-mediated Lineage Reprogramming in Vitro

Prompted by pioneering work on the reprogramming of fibroblasts to neuronal, cardiac, or hematopoietic differentiation via the overexpression of lineage-specific TFs,⁹ several groups initially employed different combinations of liver-enriched TFs, which control the development and maturation of hepatic lineage,¹¹ to induce mouse or human fibroblasts directly into hepatic cells.12, 13, 14, 15, 16, 17 Huang et al introduced genes encoding Gata4, Hnf1α, and Foxa3, along with the inactivation of p19^Arf,12 whereas Sekiya and Suzuki delivered Hnf4α along with one of Foxa1, Foxa2, or Foxa3, into mouse fibroblasts.¹³ Both studies successfully reprogrammed mouse fibroblasts into mouse-induced hepatocyte-like cells (miHeps). These miHeps acquired multiple hepatocyte-specific phenotypes and functions, and they reconstituted the injured liver of fumarylacetoacetate hydrolase knockout (Fah^–/–) mice, a model for human disease of hereditary tyrosinemia type I, after transplantation. Furthermore, Yu et al used a combination of Hnf1β and Foxa3 to convert mouse fibroblasts into mouse-induced hepatic stem cells, which could be stably expanded in vitro and differentiate bidirectionally into both hepatocytic and cholangiocytic lineages in vitro and in vivo.¹⁴ Notably, 2 groups have successfully extended liver-specific TF-based reprogramming from mouse fibroblasts to human fibroblasts. Huang et al reported that the forced expression of FOXA3, HNF1A, and HNF4A,¹⁵ whereas Du et al introduced a set of TFs (HNF1A, HNF4A, HNF6, ATF5, PROX1, and CEBPA),¹⁶ in human fibroblasts, allowed for the generation of human-induced hepatocyte-like cells (hiHeps). In both cases, the hiHeps displayed a remarkable array of liver metabolic activities, including inducible expression and activity of Cytochrome P450 Family 3 Subfamily A Member 4, comparable to that of primary human hepatocytes (PHHs). Additionally, both studies validated the repopulation potential of hiHeps in different acute liver failure models and revealed simultaneously that hiHeps have a lower repopulation capacity than PHHs upon transplantation.

To circumvent the proliferation arrest associated with terminal differentiation of these hiHeps, Xie et al developed a 2-step reprogramming strategy to successfully convert human fibroblasts into human hepatic progenitor-like cells (HPLCs).¹⁷ This strategy relied on the overexpression of a combination of 5 liver-specific TFs (HNF4A, HNF6A, GATA4, FOXA2, and HHEX) initially, followed by the selection and expansion of the converted cells in a hepatic expansion medium. These expanded human HPLCs could then be highly induced to differentiate into mature hepatocytes in vitro, closely resembling PHHs, and efficiently engrafted into injured mouse liver in vivo.

Aside from liver-enriched TF-driven lineage reprogramming into a hepatic fate, a study has also suggested the potential for establishing Yamanaka factor (OCT4, SOX2, KLF4, and c-MYC, abbreviated OSKM)-mediated reprogramming to hepatocytes. This method relies on transient overexpression of OSKM to erase epigenetic patterns of initial cells, achieving a plastic intermediate stage first, without reaching an iPSC state. The transitioning cells are then induced into a specific lineage under tailored induction conditions.¹⁸ The advantage of this approach lies in its ability to generate large numbers of cells before final differentiation into the desired cell type in vitro, while also bypassing the pluripotent state. Zhu et al cut short reprogramming to pluripotency through the temporal forced expression of OSKM in human fibroblasts to create an induced multipotent progenitor cell (iMPC) state, from which endodermal progenitor cells (iMPC-EPCs) were efficiently differentiated.¹⁹ The iMPC-EPCs could expand in vitro over 25 times and be induced to further differentiate into hepatocytes resembling primary human fetal hepatocytes.

Together, these TF-mediated lineage reprogramming approaches represent an exciting step toward the production of a large quantity of hepatocytes for drug development, disease modeling, and cell therapy. Notably, most of these studies claimed that the phenotypes of the iHeps are stable and exhibit no tumorigenesis after transplantation. However, a recent study observed that miHeps generated by forced expression of Foxa2/Hnf4a in mouse stromal cells, had an unstable hepatocyte-like phenotype. These cells showed a tendency to return to their mesenchymal phenotype of origin after several in vitro passages. Furthermore, when applied in vivo, these cells showed uncontrolled proliferation, widespread liver engraftment, and produced both hepatic and ectopic mesenchymal derivatives,²⁰ raising safety concerns about iHeps generated by TF-mediated lineage reprogramming. Concerns also arise from gene transductions that employ retrovirus and lentivirus vectors, which may lead to the genomic insertion of introduced vectors. The manipulation of oncogenes such as p19^Arf and Myc should also be carefully considered, given the potential risks for the oncogenic transformation of reprogrammed cells.

In addition to safety concerns, the efficiency of hepatic reprogramming is critical for the clinical utilization of this strategy. The currently low hepatic reprogramming efficiency is inadequate for producing a sufficient number of hepatic cells for clinical application. Significantly increasing the starting cell number for reprogramming and establishing methods for the purification or collection of reprogrammed hepatic cells will undoubtedly make the process more challenging or tedious. Thus, enhancing reprogramming efficiency to obtain a large quantity of reprogrammed hepatic cells is crucial for the clinical application of this strategy. Furthermore, the stage or state of reprogrammed hepatic cells significantly affects their therapeutic potential after transplantation. For cases of acute liver injury or insufficiency, transplanting reprogrammed hepatic cells at a mature stage may more quickly help compensate for and restore liver function. However, the terminal differentiation stage of mature hepatic cells might reduce their proliferation potential in vivo, adversely impacting the efficiency of liver repopulation. This situation necessitates transplanting a larger number of hepatic cells to quickly achieve sufficient liver function. Conversely, transplanting hepatic progenitor/stem cells could support rapid cell expansion and repopulation within the liver, requiring relatively fewer cells for transplantation. Nevertheless, since these immature cells must undergo maturation in vivo to function as mature hepatocytes, they are more suitable for treating subacute or chronic liver injuries. It is also important to note that transplanting hepatic cells at an immature stage may lead to unexpected differentiation in vivo, raising further safety concerns. Therefore, further optimization of the TF-mediated reprogramming protocols is essential for the efficient and accurate generation of hepatic cells with stable phenotypes and full differentiation capacity. A careful and deeper characterization of reprogrammed hepatic cells, with particular attention to safety evaluations and the selection of specific reprogrammed states for transplantation according to liver disease types, must be conducted before considering the cell therapy applications of these cells.

Chemical Compound-mediated Lineage Reprogramming from Fibroblasts in Vitro

The use of genetic manipulation in lineage reprogramming raises safety concerns for clinical applications, while one of the most promising potential solutions is the induction of cell fate conversion by small molecule compounds. Small molecules offer several obvious advantages over TFs for regulating cell fate: (1) they can be cell permeable; (2) they function through the regulation of proteins and signaling pathways without perturbing DNA sequences; (3) they are more cost-effective; and (4) they are more easily synthesized, preserved, and standardized. In particular, the effects of small molecules can be fine-tuned by varying their incubation time, concentrations, and combinations, thus providing a higher degree of temporal and spatial control over protein functions.²¹ Therefore, many studies have focused on generating hepatic cells by chemical compound-mediated lineage reprogramming strategies over the past few years.

Initially, 2 groups demonstrated that, under defined small-molecule combinations, a single TF of Hnf1a, Foxa1, Foxa2, or Foxa3 was sufficient to elicit the direct conversion of mouse fibroblasts into miHeps with high efficiency.^22,23 Subsequently, Li et al established a TF-independent chemical reprogramming strategy that could induce miHeps from mouse fibroblasts via an intermediate extra-embryonic endoderm-like (XEN-like) state.²⁴ However, this study did not demonstrate the metabolic functionality or the liver repopulation capacity of these chemically induced miHeps. More recently, Bai et al reported an optimized chemical approach that resulted in the reprogramming of functional hepatocyte-like cells independent of XEN-like cells. The resultant miHeps exhibit metabolic activity comparable to primary hepatocytes in vitro and the capacity to repopulate mouse livers in vivo, thereby providing a new paradigm for establishing mouse hepatocyte identity directly in fibroblasts.²⁵ Though there has been significant progress in chemical compound-driven conversion of mouse fibroblasts into miHeps, identifying a combination of small molecules that can reprogram human fibroblasts into hiHeps remains a major challenge.

Chemical Compound-mediated Lineage Reprogramming from Endodermal Epithelial Cells in Vitro

Cell fate conversion is relatively easier to achieve when there are more genetic or epigenetic similarities between the initial cells and the desired ones.26, 27, 28 The close relationship among endodermal organs makes the epithelial cells in the gastric-intestinal tract (G-I tract) a promising source for generating hepatocytes through reprogramming. Furthermore, abundant epithelial tissue of G-I tract can be obtained from patients via surgery or minimally invasive endoscopic biopsy, rendering this tissue a more attractive and pliable cell source than fibroblasts for progressing towards the goal of safe and efficient hepatic conversion using small molecule-based strategies. Our group has established a method that successfully converted human gastric epithelial cells (hGECs) to hiHeps via the human induced endodermal progenitor cells (hiEndoPCs) using a combination of 4 small molecules with the support of human gastric subepithelial myofibroblasts as feeders.²⁹ The subepithelial myofibroblasts play an indispensable role in reprogramming hGECs into hiEndoPCs. Specifically, FGF9, secreted by those myofibroblasts, activates the Wnt-signaling pathway, crucial for promoting hiEndoPC reprogramming. Notably, without the support of subepithelial myofibroblasts, the conversion of hiEndoPCs from hGECs does not occur, even with the defined small molecule cocktails alongside FGF9. This suggests that there may be other unknown events mediated by direct interactions between subepithelial myofibroblasts and gastric epithelial cells that also play a key role in the reprogramming process. Our reprogramming strategy ensures the safe and efficient generation of hiEndoPCs capable of differentiating into several clinically relevant endodermal cell types, including hepatocytes. Furthermore, these hiEndoPCs can be patient-derived, thus avoiding issues of immune rejection upon transplantation. Importantly, these cells have displayed no tendencies toward tumorigenesis. In terms of production efficiency, we can generate approximately 0.3 to 3 billion hiEndoPCs within 4 weeks from 0.1 to 1 million hGECs, obtained from small gastroduodenal specimens or endoscopic biopsies from patients aged 35 to 78 years. Subsequently, these hiEndoPCs can be induced to differentiate into functional hepatocytes (hiEndoPC-Heps) in about 2 weeks, a significantly shorter timeframe compared with the differentiation of embryonic stem cells or iPSCs. These hiEndoPC-Heps are capable of reconstructing liver parenchyma and rescuing Fah^-/- mice. Our reprogramming strategy, which particularly focuses on intra-germ layer conversions, provides a general paradigm for exploring stem/progenitor cell reprogramming. It holds promise for generating various endodermal functional cell types for personalized regenerative medicine, biological studies, and pharmaceutical industry.

Over the past few years, several other groups have also developed different chemical strategies to achieve intra-germ layer conversions without using TFs. These strategies successfully reprogrammed primary hepatocytes into expandable HPLCs. For instance, Katsuda et al and Wu et al identified distinct chemical cocktails capable of converting primary rat or mouse hepatocytes into proliferative HPLCs in vitro.^30,31 Subsequently, other groups developed independent chemically defined media that reprogramed adult or infant PHHs into proliferative HPLCs.32, 33, 34, 35, 36 These chemically induced HPLCs exhibited high proliferation potential and the capacity for functional hepatocyte differentiation in vitro, as well as liver repopulation potential upon transplantation in vivo. These protocols, relying solely on pharmacological treatment without any genetic manipulation, inherently avoid the risk of insertional mutagenesis, unlike approaches that use TFs expressed from integrating vectors. The human HPLCs generated by these methods do not undergo transformation even after prolonged in vitro expansion. Additionally, since both the starting (mature hepatocytes) and target cells (hepatic progenitor cells) in the reprogramming process are confined to the hepatic lineage, they do not produce unintended non-hepatic differentiated derivatives or tumors, despite variable liver repopulation efficiency following transplantation.32, 33, 34, 35, 36 In particular, recent work by Yuan et al addresses potential tumorigenicity concerns related to proliferating human hepatocytes (ProliHHs) generated via a chemical strategy, demonstrating that ProliHHs maintain chromosomal and genomic stability throughout culturing and are non-tumorigenic upon transplantation.³⁷ These findings collectively reinforce that chemical conversion of human hepatocytes into expanding HPLCs hold great promise for various applications, including liver disease modeling, drug testing, and regenerative medicine. However, it is crucial to recognize that the in vivo safety of these chemically derived human HPLCs has thus far only been validated in models treating acute or subacute liver injuries, such as Fah^−/−/Rag2^−/−/Il2rg^−/− mice,32, 33, 34^,37 Alb-TRECK/SCID mice,³⁵ NOD-scid IL2Rγ^-/- mice preconditioned by retrorsine treatment and partial hepatectomy (PHx),³⁴ and cDNA-uPA/SCID mice.³⁶ In these models, the injury is restricted to hepatocytes with liver architecture intact, without significant immune disturbances, fibrosis, or cirrhosis. Therefore, should these cells be considered for future clinical applications, it is imperative to transplant them into more hostile environments associated with chronic severe liver diseases like alcoholic liver disease, non-alcoholic fatty liver disease, hepatitis B or C virus infection, to thoroughly assess their safety, including any potential tumorigenicity.

Other Chemical Compound-mediated Generation of Hepatocytes in Vitro

In recent years, another technology, organoid culture has emerged as a novel tool for the robust generation and expansion of various primary cells in the form of mini-organs under chemically defined systems.³⁸ Using combination of small molecules, growth factors, and cytokines with the support of extracellular matrix, Hu et al and Peng et al have generated expandable 3-dimensional human and mouse hepatocytes organoids that retain features of primary hepatocytes.^39,40 Despite using the same initial cells, such as primary mouse or human hepatocytes, organoid technology-derived expandable hepatic cells differ from chemical reprogramming-derived expandable hepatic cells. The former resemble the unipotent proliferating hepatocytes that emerge after acute liver injury (eg, PHx), whereas the latter are similar to bipotential hepatic progenitors that occur during liver development or chronic liver injury. Additionally, 2 other groups have established different small molecular cocktails capable of maintaining primary human or mouse hepatocyte functions in vitro, without cell proliferation over the long term, providing an efficient platform for drug discovery.^41,42

These chemical strategies offer novel references that can help optimize chemical compound-based reprogramming to obtain a large quantity of human mature hepatocytes. However, it must be pointed out that the current chemically generated hepatocytes rely mainly on the use of human liver tissues or G-I tract tissues as starting materials, which are limited by the accessibility of human tissues. Overcoming these technical challenges is necessary before their broad application in clinic settings. Remarkably, the recent successful reprogramming of human fibroblasts to human iPSCs⁴³ and mouse fibroblasts to mouse hepatocyte-like cells solely using small molecule cocktails²⁵ provides encouraging proof-of-concept for the potential of chemical reprogramming to produce human hepatocytes from more accessible human fibroblasts in the future.

In Vivo Lineage Reprogramming for Liver Regeneration

In vivo lineage reprogramming refers to in situ conversion of resident cells into desired cells or inducing cells from a non-proliferative or post mitotic state into a proliferate or mitotic state. This approach helps replenish the loss of functional cells caused by organ disease or injuries. This strategy is suitable for restoring the regeneration of diseased/injured organs that lack, or have lost, the capacity to regenerate effectively. Examples include the pancreas, neuron, heart, or liver in cases of severe chronic disease.

Lineage-specific TF-mediated in Vivo Lineage Reprogramming

The classical in vivo lineage reprogramming methods also depend on the overexpression of lineage-specific TFs in the initial cells within their native tissue.⁴⁴ Zhou et al found that ectopic expression of Pdx1, Neurog3, and Mafa induced direct in vivo reprogramming of acinar cells into β-like cells.²⁸ Qian et al and Song et al made efforts to improve cardiac function by inducing scar-forming fibroblasts to become cardiomyocytes in the injured hearts of mice.^45,46 Overexpression of NeuroD1 directly converted reactive astrocytes and NG2 glia into induced neuronal cells in the adult mouse cortex.⁴⁷ These successes prompted the efforts to generate hepatocytes directly in vivo. Although hepatocytes have an outstanding regenerative capacity after acute injuries, their potential is significantly impaired when subjected to numerous chronic assaults. In liver fibrosis and cirrhosis, the hepatocytes dramatically decreased in number along with the extensive accumulation of myofibroblasts (MFs), a mesenchymal liver non-parenchymal cell type generated from hepatic stellate cells in response to liver injury. The MFs not only drive the development and progression of liver fibrosis but also impede the regeneration of hepatocytes. Song et al used a p75 neurotrophin receptor peptide-tagged adenovirus to specially target MFs and express 4 liver-enriched TFs, Foxa3, Gata4, Hnf1a, and Hnf4a in a liver fibrosis model.⁴⁸ Rezvani et al utilized a particular adeno-associated virus (AAV) vector 6 to express six liver-enriched TFs, Foxa1, Foxa2, Foxa3, Gata4, Hnf1a, or Hnf4a in MFs.⁴⁹ Both methods demonstrated comparable initial transduction efficiency in hepatic MFs (20%–30%) and produced similar yet low hepatic reprogramming efficiency (<1%). These MF-derived miHeps exhibited many aspects of the functional properties and gene expression profile of hepatocytes, and reduced the extent of liver injury, as well as fibrosis. In conclusion, direct in vivo lineage reprogramming of MFs to hepatocytes can replenish the lost hepatocytes and reduce fibrosis simultaneously, therefore opening a novel avenue for the effective treatment of liver fibrosis or cirrhosis otherwise incurable.

Small Molecule-mediated in Vivo Partial Reprogramming

Distinct from hepatic lineage-specific TF-mediated in vivo reprogramming that has been previously discussed, Lin P et al. developed a chemically induced revitalization strategy.⁵⁰ This strategy involves a combination of 5 compounds (5C): CID755673, GSK429286A, ETC-1002, Salidroside, and Forskolin. They demonstrated that systemic administration of 5C in mice with chronic liver injury leads to improved liver function and ameliorated liver fibrosis. They employed Col1a2-CreER;Rosa26-tdTomato mice to trace damaged hepatocytes in vivo, and showed that 5C treatment restores regenerative potential in these damaged hepatocytes in mice subjected to carbon tetrachloride (CCl₄)- and 3,5-diethoxycarbonyl-1,4-dihydrocollidin (DDC)-induced chronic liver injury. These beneficial effects likely contribute to the observed recovery of liver function and the reduction of liver injury. Mechanistically, the 5C treatment reverses injury-associated changes at the epigenetic level. This reversal was particularly notable among liver-enriched TFs, such as Foxa2, Hnf4a, and Foxa3, which ultimately reestablished the healthy hepatocyte-associated transcriptional networks. Overall, this study presents a promising strategy for liver regenerative medicine through the chemically induced revitalization of injured hepatocytes, aiming to repair liver tissues and resolve liver fibrosis. Despite these promising outcomes, the tracing system used does not completely rule out the possibility that some healthier hepatocytes might also be labeled post-injury, which could lead to misinterpretation of the 5C effects on damaged hepatocytes. Therefore, the specific condition and rejuvenation state of the Col1a2-traced damaged hepatocytes requires more careful characterization. Furthermore, to enhance the understanding of the 5C treatment effects on damaged hepatocytes in vivo, developing a more specific lineage tracing system for these cells is necessary in future studies.

Yamanaka Factor-driven in Vivo Partial Reprogramming

Recently, another in vivo reprogramming strategy called partial reprogramming, has emerged as a novel method to ameliorate aging processes or improve tissue regeneration.⁵¹ Partial reprogramming depends on short-term induction of OSKM to transiently and partially reprogram mature cells to a plastic state in vivo. This strategy targets cells with no or limited proliferation capacity to enable them to reenter mitosis or attain a more proliferative state. Chen et al reported that adult cardiomyocyte-specific expression of OSKM in vivo induced which to a state that resembles fetal cardiomyocytes, thereby re-conferring regenerative capacity to un-renewal adult hearts.⁵² Most recently, Hishida et al reported that transient in vivo expression of OSKM specially in hepatocytes induces partial reprogramming of adult hepatocytes to a progenitor state and has beneficial effects on regenerative capacity after acute liver injury,⁵³ suggesting that partial reprogramming represents a promising method for enhancing liver regeneration in situ. Although short-term induction of Yamanaka factors has been shown not to result in tumor formation, continuous expression of these factors often induces early lethality and teratomas in normal mice.^52,54, 55, 56 Therefore, the overexpression of OSKM in special tissues should be timely and precisely controlled. Given that the chronic liver injuries, such as fibrosis, alcoholic, and non-alcoholic fatty liver diseases are themselves the tumorigenic environment, more extensive and prolonged safety testing must be performed for in vivo reprogramming under these pathologies.

Bacteria-induced in Vivo Partial Reprogramming

A recent study introduced a natural in vivo model of Mycobacterium leprae (ML)-infected 9-banded armadillos for mammalian adult liver growth at the organ level without prior injury.⁵⁷ This research demonstrated that ML infection triggers in vivo partial reprogramming in the liver, leading to a significant increase in liver size accompanied by sustained function and architecture, without causing damage, fibrosis, or tumorigenesis during the establishment phase of infection. The ML adapts dynamic partial reprogramming to activate developmental and fetal mechanisms, thereby promoting de novo liver organogenesis and hepatocyte proliferation. This evolutionarily refined in vivo bacterial model—encompassing M. leprae and its natural animal host provides a unique example of adult liver growth devoid of adverse consequences. The bacteria-induced in vivo partial reprogramming of liver cells may help reveal endogenous pathways capable of effectively re-engaging liver growth, offering potential therapeutic implications for safer liver regeneration and rejuvenation.

Other Newly Developed Technologies for in Vivo Reprogramming

Beyond the application of lineage-specific TFs or Yamanaka factors, a novel reprogramming technique based on a modified CRISPR system was recently developed for reprogramming.⁵⁸ This system consists of a short guide RNA (gRNA) coupled to the mutating Cas9 that cannot cut DNA but is fused with transcriptional activators or repressors.59, 60, 61, 62 When directed to the desired regions by gRNA, the mutated Cas9 can either activate (CRISPRa) or inhibit (CRISPRi) the expression of targeted genes at the transcriptional or RNA transcript level. Two studies have demonstrated the successful generation of iPSCs from somatic cells by endogenously activating pluripotent TFs using CRISPRa system.^63,64 The CRISPRa system has also been used to activate endogenous cardiac or neural lineage-specific TFs, inducing the generation of cardiovascular progenitors or neuron from fibroblasts in vitro, respectively.^65,66 Interestingly, Moreno et al achieved in vivo reprogramming of rod cells into cone-like cells by knocking down a master regulator, Nrl, using the CRISPRi system, thereby preventing secondary cone cell loss.⁶⁷ Zhou et al subsequently applied a CRISPRi system, CasRx to reduce Ptbp1 expression at the mRNA level, effectively converting Müller glia into retinal ganglion cells in vivo, thus alleviating symptoms associated with retinal ganglion cells loss.⁶⁸ Relevant to liver reprogramming, Luo et al targeted Hnf4a with the CRISPRa system, inducing the expression of HNF4A at the protein level in hepatic stellate cells both in vivo and in vitro.⁶⁹ Although direct in vivo generation of hepatocytes using CRISPRa or CRISPRi systems has not yet been reported, the evidence of CRISPR-mediated gene editing in mice^69,70 suggests that these system could be feasible and promising methods for generating hepatocytes in vivo to treat liver diseases or promote liver regeneration.

Recently, RNA-based strategies have emerged as innovative reprogramming methods, attracting considerable attention due to their successful medical applications, such as the mRNA-based vaccines against COVID-19 utilized worldwide.⁷¹ Modified mRNA, known for its improved performance in stability and immunogenicity, has been harnessed to overexpress TFs instead of using DNA-encoding step for generating iPSCs in vitro.⁷² Most recently, the use of modified mRNA encoding 4 cardiac-reprogramming genes along with 3 reprogramming-helper genes induced the cardiomyocyte-like cells in vitro. The administration of this modified mRNA cocktail during myocardial infarction resulted in the presence of reprogrammed cardiomyocyte-like cells within the scar tissue, and significant improvements in cardiac outcomes were also noted.⁷³ The targeted delivery of Hnf4a mRNA to injured livers reduced fibrosis and cirrhosis,⁷⁴ further supporting the notion that mRNA-based overexpression of TFs is a promising avenue for hepatic lineage reprogramming in vivo.

In Vivo Delivery Systems

Realizing the promise of in vivo reprogramming requires the ability to deliver reprogramming agents, such as TFs, CRISPR system, and mRNA to target organs and tissues safely, precisely, and efficiently. Currently, delivery technologies, including viral vectors, lipid nanoparticles (LNPs), and virus-like particles, have been utilized for in vivo delivery. Their benefits and drawbacks are discussed in the literature.⁷⁵ In the context of liver-targeted delivery, AAVs are the most widely used viral vectors for delivering in vivo gene therapies in animal models of human liver disease⁷⁰ and in clinical trials.⁷⁶ This ubiquity is due to their natural hepatocyte tropism, low rates of genomic integration, and high biocompatibility. They have been used to deliver TFs for in vivo liver reprogramming in animal models of human liver disease as mentioned above, suggesting promising clinical applications of AAV vector-mediated reprogramming in treating liver diseases. However, reprogramming the liver in vivo efficiently and safely requires overcoming inherent challenges associated with AAVs, such as limited packaging capacity (<4.7kb), vector immunogenicity, and pre-existing immunity. Additionally, to enhance hepatocytes transduction or target liver non-parenchymal cells, particular modifications, such as attaching ligands that bind to specific receptors on the surface of hepatocytes or liver non-parenchymal cells to AAV capsids, should be performed.⁷⁷

LNPs, as non-viral vehicles, are also highly promising methods for in vivo delivery, particularly to the liver, due to their natural hepatocyte tropism mediated by Apolipoprotein E:low density lipoprotein receptor interactions between LNPs and hepatocytes.⁷⁸ Furthermore, LNPs offer several advantages over AAVs in delivering gene editing agents in mRNA form, which include unlimited package size, low immunogenicity, scalable production, and being biodegradable and non-toxic in vivo. LNPs have already been successfully employed in mouse models, primates, and in clinical trial to deliver mRNA-encoded TFs, CRISPR system-mRNA, and siRNA to hepatocytes for treating liver fibrosis and metabolic diseases.74, 79, 80, 81 Therefore, for in vivo liver reprogramming, LNPs are well-suited for delivering mRNA encoding CRISPR system or specific TFs to hepatocytes to reprogram injured/aging hepatocytes into a proliferation state or alleviate proliferation inhibition caused by liver diseases or injuries.

Collectively, these technologies have established the feasibility of in vivo lineage reprogramming for the treatment of liver diseases (Figure 3). However, the selection of the best reprogramming strategies, including target cells, reprogramming factors, and delivery methods depends on the specific tissues and diseases involved. For conditions like liver fibrosis or cirrhosis, reprogramming accumulated mesenchymal cells into hepatocytes may be suitable to overcome regenerative barriers caused by fibrosis. In cases of liver aging or impaired hepatocyte proliferation, which are characterized by a decreased regenerative capacity primarily in hepatocytes, reprogramming is considered effective for restoring a youthful or healthy regenerative state to aging/diseased hepatocytes.

Figure 3.

Different strategies for liver-targeting lineage reprogramming in vivo. TFs can be encoded by DNA- or mRNA-based sequences or be directly activated by CRISPR activation system. The AAVs and LNPs are the most frequently used vectors that deliver the reprogramming systems into liver because of their natural hepatic tropism. Furthermore, tailored combinations of small molecules can also be administered to the liver to induce cell reprogramming. Depending on the specific disease or injury, the targeting cells may be myofibroblasts or damaged hepatocytes. These cells are then converted into healthy hepatocytes or hepatic progenitors for the restoration of liver function, reducing the fibrosis or reversing aging-associated phenotypes. BD, Bile duct; CV, central vein; HA, Hepatic artery; PV, portal vein.

Endogenous Injury-induced Hepatic Epithelial Reprogramming and Implications for Liver Regeneration

In addition to the technology-induced in vivo hepatic reprogramming discussed earlier, certain types of liver injury can also trigger endogenous hepatic reprogramming. While the hepatocyte pool is primarily replenished through self-renewal of pre-existing hepatocytes during homeostasis and in injury conditions where hepatocyte proliferation is intact,82, 83, 84 endogenous hepatic reprogramming also plays a significant role in liver regeneration. Following PHx-induced acute liver injury, some adult hepatocytes transition to acquire the chromatin landscapes and transcriptomes of fetal hepatocytes, undergoing a rapid and transient adult-to-fetal reprogramming.⁸⁵ During N-acetyl-para-aminophenol (APAP) and CCl₄-induced acute pericentral injury, the hepatocytes around the injured area also transiently upregulate fetal programs at the transcriptional level when they are displaced towards pericentral area,^86,87 indicating the transcriptional reprogramming of these hepatocytes into hepatic progenitor cells (HPCs). These observations are supported by integrated transcriptional analyses in PHx and APAP-induced acute liver injury models, where the upregulation of HPC genes (Axin2, Tbx3, and Sox9) was observed in some hepatocytes.⁸⁸ It should be noted that in these cases of acute liver injury, the hepatocyte-to-HPC reprogramming is transient and occurs at the transcriptional level without an obvious and persistent change in cell identity. Moreover, the functional validation of this fetal reprogramming in liver regeneration following acute liver injury remains less explored and needs further in-depth research.

In various types of chronic liver injuries, especially DDC-induced chronic periportal injury, adult hepatocytes are found to convert to HPC or biliary epithelial cell (BEC)-like cells, exhibiting significant phenotypes of both hepatocytes and bile ducts.89, 90, 91, 92, 93 The NOTCH,⁸⁹ TGF-β,⁹³ and IL-6/STAT3⁹² signaling pathways, and Arid1a-mediated chromatin permissive chromatin state⁹¹ are involved in the hepatocyte-to-HPC reprogramming. These hepatocyte-derived HPCs could differentiate back into hepatocytes upon the cessation of injury, and subsequently contribute to approximately 25% of hepatocyte regeneration.⁹⁰ Following severe liver injuries, when hepatocyte proliferation is impaired, BECs can convert into functional hepatocytes via a transitional hepatic progenitor cell (THPC).94, 95, 96, 97, 98 These BEC-derived HPCs are bipotent and they either differentiate into hepatocytes or re-adopt a BEC fate, permanently incorporating into the liver parenchyma to mediate liver regeneration. Mechanistically, Notch and Wnt/β-catenin signaling orchestrate BEC-to-THPC and THPC-to-hepatocyte conversions, respectively.⁹⁸

Collectively, these spontaneously occurring cellular reprogramming events reveal a degree of plasticity between hepatocytes and BECs under certain regenerative stress, suggesting a crucial repair mechanism in liver regeneration. Therefore, therapies that promote this endogenous injury-induced hepatic epithelial reprogramming could open up new treatment avenue for certain liver diseases. It should be noted that, in chronic liver injuries, especially severe chronic liver injuries, the liver’s microenvironment is very harsh and already exhibits a certain degree of epithelial plasticity or a supportive reprogramming microenvironment. Therefore, using in vivo reprogramming strategies to treat these chronic liver injuries requires consideration of several important issues: Firstly, the chronic liver injury environment typically poses a higher risk of carcinogenesis for reprogrammed cells due to potential exogenous gene integration or mutations, such as those introduced by viral vectors carrying exogenous TFs. For this reason, non-integrative small molecule or mRNA technology-mediated in vivo reprogramming is more appropriate; Secondly, in these chronic liver injuries, the existing epithelial reprogramming tendencies and pro-reprogramming microenvironments may promote the exogenous reprogramming factors to function, thus it may be appropriate to reduce the dosage, intensity, and duration of use of exogenous reprogramming factors to avoid safety risks; Thirdly, in severe chronic liver conditions such as liver fibrosis or cirrhosis, where there is not only impaired proliferation in hepatocytes but also a significant BEC proliferation response, targeting impaired hepatocytes to reverse proliferation defects and targeting BECs to induce them towards hepatocytes would be beneficial for enhanced liver regeneration; Lastly, in liver fibrosis and cirrhosis, which severely limit the growth, integration, and function of reprogrammed cells, targeting proliferative fibroblasts or activated MFs to alleviate the hardened tissue microenvironment and efficiently induce their conversion to hepatocytes is also an important strategy to promote liver regeneration.

Transforming Ectopic Sites into Liver-like Organs

Orthotopic hepatocyte transplantation may not be feasible for all patients with liver disease, as the diseased liver often provides an inadequate and hostile environment for transplanted hepatocytes. Ectopic sites for transplantation could provide a healthy environment that enable successful hepatocyte engraftment. A common site for ectopic hepatocyte transplantation in rodents is under the kidney capsule. Utoh et al demonstrated that isolated hepatocytes transplanted under the kidney capsule of mice treated with monocrotaline and subject to PHx proliferated, recruited non-parenchymal cells from host tissues, and eventually formed an organ-sized, complex liver system with liver-specific architectures and functions.⁹⁹ Dong’s group has conducted extensive research on transforming the spleen into a liver-like organ. Initially, they translocated the spleen from its original site in the abdominal cavity to a subcutaneous area and remodeled the splenic tissue matrix through repeated injections of a tumor extract to support the growth of transplanted cells. Subsequently, they transplanted autologous, allogeneic, or xenogeneic liver cells into the remodeled spleen, enabling it to function like a liver.¹⁰⁰ In a recent study, they developed an improved strategy to transform the spleen into a liver-like organ without issues of immune rejection and dependence on exogenous source of seed cells. This involved stimulating splenic fibroblast proliferation and subsequently reprogramming the splenic fibroblasts into iHeps in situ.¹⁰¹ The transformed spleen was capable of compensating for liver function in a model of acute liver failure, suggesting this strategy could be a promising alternative to the liver transplantation.

Additionally, the intraperitoneal space, which includes the peritoneum and mesentery, has also been utilized for hepatocyte transplantation. A recent study showed that human iPSC-derived liver buds engrafted and developed well-vascularized hepatic tissue on the mesentery of immunodeficient mice, maturing into tissue resembling an adult liver and performing similar functions. Mesenteric transplantation of liver buds improved survival of mice after liver failure induced by either diphtheria toxin or ganciclovir in two different mouse models.¹⁰² Moreover, a landmark study reported that 8 children with acute liver failure were intraperitoneally transplanted with alginate-encapsulated PHHs. Four children fully recovered, and 3 were bridged to liver transplantation, demonstrating the potential of encapsulated PHHs to improve native liver regeneration or serve as a bridge to liver transplantation.⁴ Most recently, encapsulated reprogrammed proliferating human hepatocyte-derived liver organoids were intraperitoneally transplanted to treat liver failure animal, which increased the survival of mice with post-hepatectomy liver failure and ameliorated hyperammonemia and hypoglycemia by providing liver functions.³⁷ Furthermore, lymph nodes, which are well vascularized and allow the transport and rapid expansion of immune cells, have been discovered as a new site for ectopic liver organogenesis.¹⁰³ Eric Lagasse’s laboratory demonstrated that primary hepatocytes transplanted into lymph nodes in mice and large-animal model with tyrosinemia engraft, proliferate, and generate an ectopic liver, rescuing the animal from a fatal liver disease.104, 105, 106 They also showed that hepatocytes transplanted into the mesenteric lymph nodes in large animal models of surgically induced subacute liver failure could organize into complex ectopic liver tissues with normal cytoarchitecture, such as liver lobules and microvascularization.¹⁰⁷ Most recently, an experimental therapy that grows miniature livers inside a person’s lymph nodes has been used in a patient with liver failure for the first time, but it will be months before it is known if it fully replaces their liver function. Collectively, the generation of auxiliary liver tissue using the lymph nodes as hepatocyte engraftment sites represents a potential therapeutic approach to supplement declining hepatic function in the treatment of liver disease.

Conclusions

Together, these publications highlight that producing functional hepatocyte-like or hepatic progenitor-like cells via lineage reprogramming both in vitro and in vivo are feasible and an exciting step toward for generating a large supply of hepatocytes for therapies targeting liver diseases. Nevertheless, several significant obstacles must be addressed before lineage reprogramming can be applied to treat liver diseases and injuries in humans. First, all current in vivo validation of reprogrammed hepatic cell functionality is limited to animal models, and how these technologies will transition to human patients and their efficacy in the human body remain uncertain. Secondly, achieving full functional maturation of in vitro or in vivo converted hepatocytes to match that of normal endogenous hepatocytes remains a critical challenge, and substantial work is needed to optimize current protocols for producing ideal hepatocytes with high purity and fidelity. Thirdly, the development of better delivery vectors (eg, optimized AAV vectors and LNPs delivery system) for reprogramming agents to improve the reprogramming efficiency and liver-targeting is essential. Fourthly, it is crucial to thoroughly evaluate the potential benefits and limitations of various transplantation sites for the generated healthy hepatocytes based on the specific context of liver disease. Finally, for therapeutic applications, scalable and safe approaches that avoid genetic integration, tumorigenesis, and the use of animal-derived factors are necessary. In this respect, employing small molecules to develop chemically defined approaches and mRNA-based technologies seem promising for the clinical application of induced hepatocytes. Despite the significant progress made by current studies, further long-term fundamental research is required to test the functionality, safety, and efficiency of reprogramming strategies in liver pathophysiology that is clinically relevant.