Cell Therapy and Bioengineering in Experimental Liver Regenerative Medicine: In Vivo Injury Models and Grafting Strategies

To describe experimental liver injury models used in regenerative medicine, cell therapy strategies to repopulate damaged livers and the efficacy of liver bioengineering. Several animal models have been developed to study different liver conditions. Multiple strategies and modified protocols of cell delivery have been also reported. Furthermore, using bioengineered liver scaffolds has shown promising results that could help in generating a highly functional cell delivery system and/or a whole transplantable liver. To optimize the most effective strategies for liver cell therapy, further studies are required to compare among the performed strategies in the literature and/or innovate a novel modifying technique to overcome the potential limitations. Coating of cells with polymers, decellularized scaffolds, or microbeads could be the most appropriate solution to improve cellular efficacy. Besides, overcoming the problems of liver bioengineering may offer a radical treatment for end-stage liver diseases.

Different cell therapies and bioartificial livers have been attempted and used not only for advanced cirrhosis but also for acute and acute-on-chronic liver failure, inborn errors of metabolism, chronic cholestatic and autoimmune diseases, and non-alcoholic fatty liver disease (NAFLD) [1•, 2, 3].
Hepatocyte transplantation represents the proof of concept of liver cell therapy [4]. Sources of liver regenerative medicine that have already been applied in the clinical setting include human hepatic stem cells (hHpSCs), human biliary tree stem cells (hBTSCs), mesenchymal stem cells (MSCs), and macrophages. A recent multicenter phase-II open-label controlled trial of hematopoietic stem cells that administered repeat autologous infusions of G-CSF-mobilized CD133+ cells to patients with advanced cirrhosis (versus conservative management or treatment with G-CSF alone) found no impact on liver function or fibrosis [5]. Tissues are highly informative, especially when clinical results are weak or absent [6]. In fact, studies have shown that the role of mesenchymal-derived cells does not depend on repopulation, but on the production of factors and cytokines with multiple effects [7,8].
Clinical studies on liver regenerative medicine highlight the importance of solid preclinical evidence in this field. Moreover, preclinical studies should be tailored to address the questions raised by the clinical trials; e.g., the assessment of factors originating from outside the liver, such as from the gut (e.g., intestinal permeability, dysbiosis), sarcopenic muscles, or inflamed adipose tissue [6]. Importantly, liver repopulation and the proliferation and differentiation of transplanted cells should be investigated systematically in tissue over different long-term timepoints. Different experimental models should be evaluated to investigate specific etiopathogenetic features that may influence cell therapy outcomes. For example, models of liver fibrosis are the best candidates to study intrahepatic factors associated with the interactions and effects of exogenous cell transplantation, while NAFLD/ nonalcoholic steatohepatitis (NASH) models may reveal potential systemic factors that influence the effects of exogenous cells transplanted into the liver [6]. In the future, the in-depth study of stem/progenitor cell therapy effects could reveal antiinflammatory, antioxidant, and immunomodulatory effects that are now only studied in macrophages [8].
This review aims to provide an informative and helpful tool for designing preclinical investigations into liver regenerative medicine, including cell therapy and liver bioengineering. Although this is not a systematic review, experimental liver injury models already used in the setting of experimental liver cell therapy and other suitable models tested in rodents have been analyzed and reported, so key features can easily be appreciated, e.g., injury modality, time from injury to rescue treatment, serum tests, and histologic features (Table 1).

Acute Liver Failure models
Acute liver failure (ALF) is a condition that can arise from a broad spectrum of causes. It is characterized by the loss of hepatic, metabolic, and immunological functions [9,10].
The administration of acetaminophen (APAP) to mice is the most common ALF model since the APAP doses that cause toxicity are similar in mice and humans (≥150 mg/kg) [11][12][13]. Liu et al. recently reported a study in which BALB/c mice treated with APAP were used as a model. They observed an effect of intravenously administered human umbilical cordderived mesenchymal stromal cells (hUCMSCs) through their ability to reduce hepatic necrosis/apoptosis and enhance liver regeneration [14].
A useful tool to study TNFα-mediated apoptotic signaling mechanisms [15,16] and inflammatory-mediated liver injury [15,17] is the D-galactosamine/endotoxin (Gal/ET) model. D-Galactosamine inhibits protein synthesis by depleting uridine triphosphate pools, causing early generation of reactive oxygen species and finally apoptosis [18,19], while lipopolysaccharides increase the release of proinflammatory cytokines (TNFα) [20]. Zhang et al. used this common ALF model by injecting lipopolysaccharides and D-GalN aminoglycan into mice [21] and observed a therapeutic effect of TNFα pretreated umbilical cord mesenchymal stem cellderived exosomes (T-Exo). They demonstrated that T-Exo ameliorates conditions of ALF by inhibiting the activation of the NOD-like receptor protein 3 (NLRP3)-related inflammatory pathway [22].
A relevant limitation in MSC transplantation therapy is its poor efficacy in liver colonization and viability [23]. In this context, Ma et al. used a carbon tetrachloride (CCl 4 ) ALF model in nude mice to study the effect of genetically modified MSCs expressing CXC receptor 4 (CXCR4). Their work showed greater colonization of the failing liver by CXCR4-MSCs, leading to reduced mortality and improved liver regeneration [24].

Chronic Liver Disease Models
Chronic liver injury is associated with fibrosis leading to cirrhosis, a disease characterized by high levels of proinflammatory cytokines [25,26], an abnormal lobular architecture, and the formation of intrahepatic vascular shunts [27]. It has been demonstrated that carbon tetrachloride (CCl 4 ) given at low doses (0.5-0.8 mL/kg) causes persistent liver injury in rats with inflammation and fibrosis [28,35] (Fig. 1). Recent studies clearly demonstrate that bone marrow mesenchymal stem cell microvesicles (BM-MSC-MVs) and MSC small extracellular microvesicles (MSC-sEVs) possess anti-fibrotic, anti-inflammatory, and pro-angiogenic properties that can promote the resolution of CCl 4 -induced liver fibrosis in rats [29,40].

Liver Regeneration Following Surgical Partial Hepatectomy
Intense regeneration follows a partial hepatectomy involving a 70% resection in rats and pigs [30][31][32][33]. Makino et al. established 90% hepatectomy as the safety limit for murine hepatectomy and as a model for liver regeneration [34]. Eguchi et al. proposed an alternative model in rats involving resection of the two anterior lobes (68%) and ligation of the right liver lobes (24%), stimulating cytokine release following ischemic/reperfusion injury [36].
Several ALF models, such as those using hepatectomy or drug toxicity, have a high mortality rate in the acute phase [37][38][39], rendering them not perfectly suitable for regeneration studies. Inomata et al. standardized a new treatment using retrorsine (RS) together with partial hepatectomy (PH) in pigs to overcome these limitations. In 2019, Tsuchida et al. studied the engraftment of rat liver organoids derived from fetal livers in an RS/PH rat model. Organoids were injected through the portal vein, which led to liver regeneration and reduced ductular reaction. Remarkably, no translocation to other organs was observed [41].

Genetic Models of Inborn Metabolic Errors
Several genetic disorders can lead to liver failure, so it is important to use appropriate animal models mimicking human conditions. Hickey et al. developed the first genetically  [42].
The Gunn rat is a natural model for bilirubin encephalopathy because it inherently lacks all glucuronidation activity catalyzed by the UDP glucuronosyltransferase 1 (UGT1) isoform [45]. For this reason, these rodents are used as models to study type I Crigler-Najjar (CN) syndrome [46]. In this context, several studies have investigated the effect of humaninduced pluripotent stem cells (iPSCs) in cell therapy. In 2015, it was demonstrated that iPSCs reprogrammed from human skin fibroblasts could differentiate into hepatocytelike cells (iHeps). These cells were transplanted into the livers of Gunn rats and induced a 30-60% decline in serum bilirubin. Moreover, the excretion of bilirubin glucuronides indicated that transplanted iHeps expressed UGT1A1 activity [47].
Another genetic rat model is the Long-Evans Cinnamon rat, whose mutation mimics human Wilson's disease in terms of excessive copper accumulation in the liver, low levels of serum ceruloplasmin, and low excretion of copper into the bile [48].

Cholestasis and Cholangiopathy Models
Cholestasis refers to impairment in bile formation or excretion. This can be due to defects in intrahepatic bile production or transmembrane bile transport, or to mechanical bile flow obstruction. Cholangiocytes are hepatic cells that regulate the fluidity and alkalinity of bile through the secretion of osmolytes, such as Cl − and HCO 3 − . Primary damage to the biliary epithelium causes several chronic cholestatic disorders (cholangiopathies), which are always observed with coexisting cholangiocyte death and proliferation and various degrees of portal inflammation and fibrosis.
The most common experimental rodent model of intrahepatic cholestasis involves the hepatotoxicant molecule known as α-Naphthylisothiocyanate (ANIT) [49,50]. Recently, FVB.129P2-Abcb4 tm1Bor mice (FVB.Mdr2 −/− ) were used as a model for human primary sclerosing cholangitis (PSC) (Fig. 1). These rodents are characterized by the complete inability of the liver to secrete phospholipids into the bile, which results in the spontaneous development of progressive chronic biliary injury and fibrosis [51]. In this context, extracellular vesicles (EVs) isolated from human MSCs have been proposed as a possible effective therapeutic strategy to treat PSC patients [51]. Another mouse model of cholangitis can be obtained by injecting 2-octynoic acid coupled with bovine serum albumin (2OA-BSA) [52]. Fan et al. used this model to prove that human umbilical cordderived MSCs (UC-MSCs) injected intravenously could ameliorate liver inflammation [52].

Autoimmune Hepatitis Models
Autoimmune hepatitis (AIH) is a chronic inflammatory disease of the liver characterized by the loss of self-tolerance leading to the appearance of autoantibodies and dysfunction [53]. The first AIH animal models involved the induction of transient hepatitis by immunizing rabbits or mice with complete Freund's adjuvant [54][55][56]. Other models included endotoxin and plant lectin-induced hepatitis. Tiegs and colleagues first demonstrated that 20 mg/kg concanavalin A (ConA) induces T cell-mediated liver damage that is mainly dependent on CD4 + T cells in mice [57]. More recent techniques involve the application of genetic engineering technology [58]. Knocking out specific genes allows ConA models to mimic specific AIH subtypes [53,59]. The most important considerations about these models are (1) the ConA model is a very severe model of liver injury that can lead to high mortality due to severe hemorrhage; and (2) it is difficult to find the antigen of autoantibodies, which is a limitation of the gene-engineered AIH model.
Cell-based therapies are a promising tool for the treatment of AIH. It has recently been highlighted that peptide-major histocompatibility complex class II (pMHCII)-based nanomedicines displaying different cellular epitopes ameliorate AIH conditions without the suppression of host immunity [60].

Routes and Strategies Adopted to Repopulate the Liver and Biliary Tree Through Cell Therapy
Other factors could influence transplanted cell effectiveness in treating diseased livers, including cell delivery routes and repopulation strategies. These factors alter cellular engraftment and functionality. Therefore, several cell therapy protocols in experimental models were studied to maximize the therapeutic effect of transplanted cells by reducing potential complications [61][62][63][64].
In general, cell delivery to the liver could be performed in multiple ways due to its good anatomical accessibility. The vascular route is the most widely used, and cell delivery has been performed mainly by injecting cells into the portal vein, hepatic artery, or splenic artery [65][66][67]. Less relevant approaches involve cell delivery to the liver parenchyma via injection into the intraperitoneal cavity, or through percutaneous tissue [65,68]. Although intravenous injection is relatively safe for cell delivery, it is not considered a perfect route because it carries the risk of emboli formation, which could lead to liver infarction, cell damage, or poor cell engraftment due to vascular shear stress [69,70]. However, the study by [71••] showed that perfusing cells in a 3D-spheroid form via the intraportal vein could improve cell localization in treated livers, with superior therapeutic benefits in a mouse model. The hepatic artery could provide a better alternative cell delivery route in the presence of portal hypertension-related chronic liver disease [69,72]. Intrahepatic/parenchymal cell injection is a promising cell delivery strategy, but some complications possibly related to this approach include the risk of injuring parenchymal tissue and blocking the hepatic vascular system or pulmonary capillaries [65].
Developing new modified cell delivery protocols is necessary to achieve more desirable liver cell therapy effects. Intrasplenic injection of hyaluronan-coated hBTSCs in severe combined immunodeficient (SCID) mice showed better cellular engraftment and differentiation into mature hepatocytes [73••]. A recent study by Hwang et al. (2019) coated human adipose-derived stem cells with lipid-conjugated heparin before injecting them via the tail vein into the APAP-induced ALF mouse model and found that this coating method could enhance cell therapy effect on liver damage [74]. The study by Laing et al. found that multipotent adult progenitor cells could be delivered to the donor's liver in vitro via normothermic machine perfusion before liver transplantation to provide the liver with anti-inflammatory and superior immunomodulatory properties [75].
Other strategies to obtain cell therapy effects in liver disease models that do not involve cell transplantation have been tested. These strategies could be performed using cell-free products or derived extracellular nanovesicles, either alone [76] or in combination with transplantable material. A study by Mardpour et al. (2019) tested the intraperitoneal injection of a mixture of polyethylene glycol (PEG) macromeres and MSC-derived extravesicles (MSC-EVs) as a delivery strategy to extend the beneficial effects of cell-free products for a longer time [77]. More research is still needed regarding the effects of different coating materials in the delivery system.
Ongoing studies are needed to compare the different cell delivery approaches in liver disease models with cells of different properties and sizes. Based on the fact that cells can easily be coated or capsulated with natural or synthetic constructs, studies are also needed to evaluate the combination of transplanted cells with different promising materials, scaffolds, or microbeads in order to improve liver cell therapy outcome.

Bioengineered Liver Scaffolds and Their Transplantation
Bioengineered liver scaffolds aim to generate a highly functional cell delivery system and/or a whole transplantable liver. End-stage liver disease (ESLD) requires radical treatment via liver transplantation in order to save patient lives. However, due to donor shortages and long ESLD patient waiting lists, transplantation surgery may not be available to most patients and it is challenging to find an alternative way to provide transplantable liver scaffolds [78,79].
Different sources and types of liver scaffolds have been studied in the literature [80][81][82][83][84][85][86] (Fig. 2). The most promising and interesting source of liver scaffolds is native hepatic tissue [87][88][89]. Decellularization of the native livers of lab animals, pigs, and humans has been tested to determine the most suitable and applicable way to generate transplantable bioengineered liver scaffolds [85,[90][91][92]. Lab animal liver decellularization was studied as a preliminary step to evaluate the gentlest decellularization protocol by characterizing their biochemical components and structural properties as compared with native tissues before testing in larger animals or human tissue [90] (Fig. 2). Pig livers are considered the most readily available source of transplantable scaffolds for preclinical studies and further clinical applications [93][94][95]. It is worth mentioning that either one liver lobe of an adult pig or the whole liver of a piglet could be suitable for further transplantation [94,95]. Human liver decellularization is also being investigated as a way to use human livers unable to be transplanted due to the presence of ischemia or other liver diseases. Therefore, decellularization is a good alternative option to increase the number of suitable liver scaffolds that could be recellularized and transplanted [82,96].
Various decellularization protocols have been described for generating a cellular hepatic scaffold. Washing out of cells is mainly performed using chemical detergents such as sodium dodecyl sulfate (SDS) and Triton-X-100, endonuclease enzymes such as DNases, or different combinations of chemicals and enzymes with agitation for liver slices, or by either continuous or pressure-controlled whole liver perfusions [82,85,90,92,[97][98][99]. Decellularization under oscillating pressure has shown better results in terms of perfusion homogeneity and acellular tissue integrity [85,92,98,100]. The main aim of optimizing different decellularization methods is to obtain a balance between the micro/ ultrastructure and remaining DNA level in decellularized tissue [90].
However, many complications are related to the decellularization process, including the loss of some native critical components and bioactive molecules. Many studies have investigated how to improve the structural properties of acellular scaffolds through conjugation with crosslinking agents or extracellular matrix (ECM) particles or by improving their functionality with modifying agents that could increase cellular attachment and hemocompatibility, such as CD31, REDV peptide, fibronectin, heparin, and gelatin [80,84,93,[101][102][103][104][105]. One study used mice and pig livers that were partially hepatectomized and maintained a few days in vivo before being harvested and decellularized to generate hepatic acellular scaffolds in an active regenerative state. These active scaffolds could be more functional and hemocompatible, with fewer limitations than scaffolds obtained from intact native livers [106].
The native liver-derived ECM can be used as an intact scaffold for further orthotropic or heterotopic transplantation [102,107,108]. There are different forms of acellular ECM, including hydrogel, powder, and sheet tissue papers. These forms are fabricated using digested ECM, homogenized ECM, or ECM ink, respectively. These ECM forms could be used for cell delivery, but they have relevant limitations, mainly due to the additional loss of biochemical components during the digestion and fabrication processes [97,[109][110][111][112][113][114][115][116].
Conversely, natural polymers such as collagen and synthetic biodegradable materials have also been studied alone or in combination to generate cell-supporting and functional hepatic scaffolds or cell delivery systems [117][118][119][120][121][122]. However, these scaffolds could not provide cells with all the bioactive molecules required for their growth as compared with decellularized ECM, which is why synthetic materials were mixed with ECM powder or solubilized ECM to make them more biocompatible and functional. However, many factors regarding the physical and mechanical properties of the synthetic scaffolds should be optimized in order to be approved for clinical application [110,123].
Cell seeding of acellular hepatic scaffolds was performed mainly via cell perfusion into vascular systems and/or the common bile ducts when using whole decellularized hepatic scaffolds that had intact inlets and outlets, or by direct multiple parenchymal injections, while cells could be laid onto the top surface of hepatic decellularized slices [82,84,[124][125][126][127].
Different cell types could be used for repopulating decellularized livers, including primary hepatocytes, iPSCderived hepatocyte-like cells, stem cells, or mesenchymal stromal cells for recellularization of the parenchymal areas, and endothelial progenitor cells or human umbilical vein endothelial cells (HUVEC) for reendothelization [85,[128][129][130][131][132]. However, optimization of the recellularization process is still a challenge when more than one cell type per scaffold is used.
Recellularized scaffolds could be transplanted after culturing in a bioreactor in vitro for cell differentiation and maturation [131,133,134]. The transplantation of recellularized hepatic scaffolds was performed experimentally and preclinically to evaluate biocompatibility and in vivo functionality. Different ways to transplant the recellularized liver constructs were studied to optimize the most effective minimally invasive technique to restore recipient liver function or compensate for the loss of function [61,82,87,103,135].
To conclude, the most important considerations in generating highly functional cell delivery systems and transplantable hepatic scaffolds include minimizing the deleterious effects of the decellularization process, modifying the acellular scaffolds to overcome existing limitations, and appropriately selecting cells.