By mimicking the development signaling, stepwise chemical treatment induces differentiation of human embryonic stem cells (hESCs), induced PSCs (iPSCs), and human PSCs into hepatocytes. The target and mechanism of small molecules are summarized in Table 1. GSK3β inhibitor CHIR99021 (CHIR) activated Wnt/β-catenin to induce DE from hESCs combined with Nodal signaling activator activin A in RPMI + B27 medium (Raggi et al., 2022). Another GSK3β inhibitor BIO was also used to activate Wnt/β-catenin pathways (Tasnim et al., 2015; Onai, 2019). Activin A is able to be substituted with IDE1, an activator of Nodal signaling (Luo et al., 2021). The combination of IDE1, CHIR, and LY294002 (a PI3K inhibitor) induced DE specification from human umbilical cord-derived mesenchymal stem cells (Luo et al., 2021). However, this combination induced mesoderm formation in human PSCs probably due to the excessive level of Wnt signaling favoring mesoderm formation (Pan et al., 2022). To cope with, PD0332991, a CDK4/6 inhibitor was added to the combination of IDE1 and CHIR, resulting in more specific DE differentiation (Pan et al., 2022). To promote DE differentiation to PFG, Touboul T et al. treated hESCs with Wnt/β-catenin inhibitor IWR-1 together with activin A and noggin, resulting in induction of PFG gene expression (Touboul et al., 2016), although noggin was later found to impede with the promoting effect of IWR-1 on PFG induction. For further differentiation into HBs, a combined protocol of BMP4, IWR-1, and TGFβ inhibitor SB431542 eliciting hepatic markers, including fetoprotein (AFP), albumin (ALB), and prospero homeobox 1 (PROX1) was established (Touboul et al., 2016). Du, C et al. combined A83-01, a TGFβ signaling inhibitor, with two epigenetic modulators sodium butyrate (NaB) and dimethyl sulfoxide (DMSO) to induce hepatic differentiation from DE (Du et al., 2018). Finally, Touboul T et al. showed that CHIR treatment generated proliferative HBs with co-expression of HNF6 and PROX1, although activation of Wnt/β-catenin pathways theoretically promote biliary differentiation (Touboul et al., 2016).

Hepatocyte growth factor (HGF) and its receptor c-Met regulates hepatic maturation during liver development. In developing livers, the expression of HGF and c-Met upregulated between 4 and 21 days after birth (Hu et al., 1993). In fetal hepatocytes, treatment of HGF together with glucocorticoids promoted the expression of ALB (Kamiya et al., 2001), suggesting the effect of HGF on hepatic maturation. N-hexanoic-Tyr-Ile-(6) aminohexanoic amide (Dihexa), a small-molecule angiotensin IV analog, was found to efficiently activate HGF/c-Met pathway (Benoist et al., 2014). Treatment of Dihexa and glucocorticoids successfully induced functional HLCs from hepatic progenitor cells (HPCs)(Siller et al., 2015; Asumda et al., 2018). In addition, HGF can be replaced by other small molecules. Shan, J et al. identified two chemical compounds functional hits 1 (FH1) and functional proliferation hit 1 (FPH1) through high-throughput screen (Shan et al., 2013). Later, Du, C et al. established a protocol including FH1, FPH1, A83-01, and glucocorticoids, which yielded functional HLCs more efficiently compared to growth factor groups (Du et al., 2018). Luo S et al. combined FH1 with growth factors to induce hepatocyte differentiation from HPCs. FH1 induced hepatocytes with higher efficiency and similar maturity compared to HGF induced ones (Luo et al., 2021).

Recently, the hepatic differentiation of current good manufacturing practice (cGMP) compliant human iPSC (hiPSC) and hESC lines have been validated. The protocol successfully generated hepatocytes from cGMP cell lines using small molecules combined with growth factors in 21 days (Blackford et al., 2019). The ALB expression was detected but insignificant in the induced hepatocytes, suggesting that the hepatocytes was immature (Blackford et al., 2019). In comparison, the pure chemical protocol from Du et al. produced hepatocytes from hESCs within 13 days. The ALB positive hepatocytes account for 67.7% versus 37.1% in the growth factor group (Du et al., 2018). The hepatic differentiation using growth factor-free small molecule cocktails should also be validated in cGMP compliant hiPSC and hESC lines for future clinical application. Moreover, the combination and dosage of small molecules vary between research groups. Comparison studies are needed to evaluate the differentiation efficiency of protocols and to conclude a cGMP compliant protocol.

Recently, studies have shown that somatic cells can be directly reprogrammed into specific cell lineage, including hepatocytes. Direct reprogramming without entering pluripotent cell state shortens the time for cell differentiation and lower the risks of tumorigenesis, which is deemed as a promising method to yield hepatocytes for HTx.

Early in 2011, Sekiya et al. established three combinations comprising two transcription factors: HNF4α with FOXA1, FOXA2 or FOXA3, which successfully converted mouse embryonic fibroblasts (MEFs) into induced hepatocyte-like cells (iHeps) (Sekiya and Suzuki, 2011). However, the conversion efficiency of the protocol was insufficient (only 0.3%) (Sekiya and Suzuki, 2011). Hui’s group discovered a specific combination of three transcription factors consisting of GATA4, HNF1α and FOXA3, which converted p19^Arf−/− mouse tail-tip fibroblasts to iHeps with a higher conversion efficiency (Huang et al., 2011; Ji et al., 2013). The researchers suggested that Hnf1a and Foxa3 transduction was already sufficient to induce iHeps and addition of Gata4 enhanced the conversion efficiency (Ji et al., 2013). Later, Hui et al. showed that transduction of Hnf1β and Foxa3 directly induced MEFs into hepatic stem cells which had the bipotential of differentiating into hepatocytes and cholangiocytes (Yu et al., 2013).

The clinical application of iHep generated by viral transduction in HTx have been limited due to the safe uncertainty, genetic instability and tumorigenesis risk. Therefore, studies about chemical induction of hepatocytes from fibroblasts have thrived in the field (Tables 1, 2).

Guo, R et al. showed that combination of a single transcription factor FOXA1, FOXA2, or FOXA3 with six compound cocktails, including CHIR, RepSox, valproic acid (VPA), Parnate, TTNPB, Dznep (termed CRVPTD), directly induced MEFs into iHeps (Guo et al., 2017). Researchers postulated that FOXAs upregulated liver-specific genes independent of any specific signaling pathway and thus was difficult to replace. Horisawa, K et al. demonstrated that liver-specific gene expression was promoted in MEFs by sequential and cooperative binding of FOXAs and HNF-4α to chromatin. FOXA3 exert unique regulations on the gene expressions among FOXA proteins, that is, by binding to and co-transversing the target genes with RNA polymerase II, which is indispensable for reprogramming MEFs to hepatocytes (Horisawa et al., 2020).

Later studies developed protocols independent of ectopic expression for direct reprogramming. Li, X et al. established a seven-compound cocktail (VPA [V], TD114-2 [T], E−616452 [6], tranylcypromine/Parnate [P], forskolin [F], AM580 [A], and EPZ004777 [E]) to regulate reprogramming-related signaling pathways (T6FA) and to modulate the epigenetic profile (VPE) (Li et al., 2017a). The small-molecule cocktail induced extra-embryonic endoderm (XEN)-like cells from fibroblasts bypassing the pluripotent state, which could differentiate towards hepatocytes (Li et al., 2017b). A later study elucidated that compounds CHIR, E−616452, and forskolin firstly worked in a cooperative manner to activate Sox7; CHIR/forskolin and E−616452 then activated Gata4 and Sall4 expression, respectively. The consecutive activation of the crucial transcription factors attributed to the conversion of MEF to XEN-like cells (Yang et al., 2020). Bai, Y et al. established a two-stage chemical cocktail to directly reprogram MEFs to iHeps (Bai et al., 2023), modified from the cocktails (C6FAE and C6F5UE) presented in Yang’s work (Yang et al., 2020). The researchers found that addition of vitamin C in the second stage was the optimal protocol to induce iHeps, with the iHep subpopulation of 15% in total cells (Bai et al., 2023). Principal component analysis and RNA-seq showed that the iHeps resembled PHHs (Bai et al., 2023). Finally, the iHeps was capable of engrafting the liver in Fah^−/− mice and improving liver functions in vivo in a similar manner to PHHs (Bai et al., 2023). Noteworthy, Zhong Z et al. directly reprogramed MEFs into iHeps with one step chemical induction consisting of SB431542, CHIR99021, BIX01294 (G9a KMT inhibitor), LDN193189, and DAPT (Zhong et al., 2024). During 12-day induction, no pluripotency gene expression (Oct4, Sox2, and Nanog), immature HB markers, and HPC markers were detected, suggesting that the conversion of MEFs to iHeps bypassed the intermediate stem cell stage (Zhong et al., 2024). Mechanistically, the chemical cocktail suppressed SNAI1 expression, which induced mesenchymal to epithelial transition and HNF4α expression, thereby promoting iHep generation (Zhong et al., 2024).

Hui’s group reported that the induction of iHep generated by transcription factor cost 14 days and the induction efficiency was around 23% (Huang et al., 2011). In comparison, in studies using chemical approaches to produce iHep, the induction time was 20, 12, 12 days and the efficiency was over 20%, 15%, and 80%, respectively (Li et al., 2017a; Bai et al., 2023; Zhong et al., 2024). Collectively, the chemical approach inducing iHep from fibroblasts has a similar and even more robust induction efficiency to genetic approach, making it a promising method to produce hepatocytes from MEFs.

Epigenetic modulations are important mechanisms of transcriptional regulation during normal development. During differentiation from PFG towards HBs, histone acetylation significantly increased at the liver regulatory elements mediated by histone acetyltransferase P300 (Xu et al., 2011). Moreover, H3K27 KMTs enriched at Pdx1 gene (a pancreatic gene) upstream which suppressed the pancreatic development while favored hepatic development (Xu et al., 2011). The methylation of CpG islands is regulated by DNA methyltransferases (DNMTs) and mediate gene expressions by mediating promoters (Wilkinson et al., 2023). Based on the mechanisms, histone deacetylase (HDAC) inhibitors and DNA/histone methylation mediators are used to promote hepatic differentiation. It was reported that combination of NaB with DMSO elicited hepatic differentiation from DE with comparable hepatic marker expression (AFP and HNF4α) to the growth factor-treated group (Tasnim et al., 2015). Demethylation of CpG by DNMT inhibitors allowed liver-specific gene expression, including AFP and ALB (Snykers et al., 2009). It was reported that combination of nanaomycin A, a selective DNMT3B inhibitor, with FGF4 and BMP4 promoted the differentiation from DE to HPCs (Nakamae et al., 2018).

Epigenetic modulation is also required in cell reprogramming. In differentiated cells, high levels of DNA methylation occurred in the CpG-rich promoters of pluripotency-related genes (Baral et al., 2022; Ghazimoradi and Farivar, 2020). Therefore, inhibition of DNMTs promotes demethylation of CpG islands and thereby removes the epigenetic barrier of cell reprogramming. It was demonstrated that the combination of RG108 (a DNMT inhibitor) and with BIX01294 promoted the reprogramming of MEFs to iPSC upon Oct4 and Klf4 transduction (Shi et al., 2008).

Histone acetylation/deacetylation is mediated by histone acetyltransferases and HDACs, respectively. Histone acetylation is related to open chromatin structure and transcriptional activation (Baral et al., 2022). HDAC inhibitors, including VPA and NaB, were shown to facilitate cell reprogramming in several studies (Aguirre-Vázquez et al., 2021; Duan et al., 2019; Mali et al., 2010). Histone lysine methylation is tuned by KMTs and histone lysine demethylases. The effect of histone lysine methylation on gene transcription varies, either activating or inhibitory. LSD1 specifically catalyzes the demethylation of H3K4/9me1/2 (Dai et al., 2021). Enhancer of zeste homolog 2 (EZH2) and DOT1L catalyzes methylation of H3K27 and histone three lysine 79 (H3K79), respectively (Wu et al., 2023; Lee et al., 2022). Small molecule inhibitor of LSD1 (Parnate), EZH2 (DZNep), and DOT1L (EPZ004777) promoted somatic reprogramming efficiency (Martinez-Gamero et al., 2021; Liuyang et al., 2023).

Based on the aforementioned mechanisms, researchers have applied small molecules to modulate epigenetic profiles in somatic cells and thus to promote somatic reprogramming into hepatocytes bypassing the iPSC stage (Table 1 and 2). Zhu, S et al. demonstrated that combination of NaB, Parnate, and RG108 with a Wnt signaling activator significantly reprogrammed MEFs into multipotent progenitor cells with transient expression of three transcription factors (OCT4, SOX2 and KLF4) (Zhu et al., 2014). Multipotent progenitor cells differentiated towards EPCs and hepatocytes sequentially in the hepatocyte induction medium (Zhu et al., 2014). Guo R et al. used VPA, Parnate, TTNPB, and Dznep combining with other small molecules to induce direct reprogramming from MEFs to iHeps with only a single factor FOXA3 (Guo et al., 2017). Li, X et al. developed a 7-compound protocol consisting of VPA, EPZ004777, Parnate and other small molecules which directly induced the XEN-like cells from MEFs (Li et al., 2017b). Zhong Z et al. combined BIX01294 with other signaling modulators to directly converted MEFs to iHep (Zhong et al., 2024).

HPCs can be yielded either from livers or iPSCs and thus becomes one of accessible cell sources for HTx. However, the limited ability of ex vivo proliferation impedes with its application. During liver development, Wnt signaling promotes the proliferation, expansion, and differentiation of HPCs (Perugorria et al., 2019). Hedgehog signaling and BMP4 facilitates the proliferation of fetal HPCs in vitro (Hirose et al., 2009; Wang et al., 2015), while TGFβ signaling remarkably inhibits HPC colony formation in vitiro (Clark et al., 2005). Notch signaling activation promotes HPC differentiation towards cholangiocytes (Duan et al., 2018). These data suggested that fine modulation of these critical signalings might contribute to HPC expansion.

Our lab established a protocol consisting of EGF, CHIR, E−616452 and two bioactive lipids lysophosphatidic acid and sphingosine 1-phosphate, hereafter termed as ECELS. Mouse HBs maintained self-renewal in the Matrigel-coated media containing ECELS and bovine serum albumin (Lv et al., 2015). These expandable HBs could differentiated to mature hepatocytes by chemical induction (Lv et al., 2015). Pan, T et al. established a chemical cocktail of CHIR, A8301, and SAG (a Hedgehog signaling activator) combined with EGF, HGF and BMP4, which significantly promoted the expansion and stemness maintenance of HPCs (Pan et al., 2019). Later, researchers replaced EGF, HGF, and BMP4 with forskolin, Dihexa, and vitamin C, respectively. This chemical cocktail enabled the long-term self-renewal with efficient proliferation rate for at least 20 passages and the capability of differentiating into hepatocyte and cholangiocytes (Pan et al., 2021).

Mature hepatocytes converted to liver progenitor-like cells and re-differentiated into hepatocytes in response to chronic periportal liver injuries in vivo (Li et al., 2020). The liver progenitor-like cells generated duct-like cells when liver damage persisted (Li et al., 2020). Yan et al. established a transition and expansion medium which consisted of EGF, HGF, CHIR, lipids lysophosphatidic acid, sphingosine 1-phosphate, A83-01, and Y-27632. This protocol elicited significant transition of hepatocytes to duct-like cells and promoted the proliferation of duct-like cells (Wu et al., 2017). Miyoshi T et al. converted PHHs harvested from cirrhotic livers to liver progenitor cells by a protocol including Y-27632, A83-01, and CHIR (YAC). The induced liver progenitor cells differentiated into mature hepatocyte under induction of YAC, oncostatin M, and dexamethasone (Miyoshi et al., 2022).

The quick loss of functions in isolated hepatocytes is one of the main hurdles in hepatocyte-based cell therapy. Therefore, it is urgent to investigate on the underlying mechanism and discover corresponding strategies. Our lab demonstrated that mechanical tension represented by actin remodulation elicited hepatocyte dedifferentiation towards HPCs through Yap activation. Yap deletion in isolated hepatocytes resulted in maintenance of ALB and HNF4α expression (Sun et al., 2019). Based on the iterative chemical screening, we found that combination of an actin polymerization inhibitor Latrunculin B and actomyosin contraction inhibitor Blebbistatin led to increased levels of ALB and HNF4α. Addition of Dasatinib, XAV939, and LY294002 (Yap mediators) further facilitated hepatic genes expression (Sun et al., 2019). This chemical cocktail (termed LBDXL) was able to maintain the hepatic gene expression and metabolic functions of hepatocytes cultured in Matrigel-coated medium for up to 3 weeks (Sun et al., 2019).

The activation of epithelial to mesenchymal transition in hepatocytes results in loss of normal hepatic functions and regenerative capacity, where TGF-β signaling pathway play an important role (Xue et al., 2013). Xiang C et al. confirmed that the expression of components of TGF-β signaling pathway significantly upregulated in cultured PHHs. The researchers established a 5-compound (5C) protocol, consisting of SB431542, forskolin, DAPT (Notch inhibitor), IWP2 (Wnt inhibitor), and LDN193189 (BMP inhibitor) to suppress the expression of epithelial to mesenchymal transition marker genes (Xiang et al., 2019). 5C-cultured PHHs still supported hepatitis B virus infection after 4-week culture (Xiang et al., 2019). Additionally, Chen Y et al. developed a small-molecule cocktail consisting of SB431542, acetylcysteine (ROS inhibitor), CHIR, and Y-27632 (SACY)(Chen et al., 2021). Compared to 5C culture system, SACY exhibited higher level of ALB and α1 antitrypsin production and urea synthesis (Chen et al., 2021).

Loss of functions in hepatocytes was also observed in vivo, specifically in chronically injured liver. Lin P et al. showed that damaged hepatocytes harvested from CCl4-induced mice liver regained hepatocyte phenotypes and functions with treatment of five compounds (forskolin, CID755673, GSK429286A, ETC-1002, and phenylpropanoid glycoside) (Lin et al., 2023). Injection of revitalized hepatocytes promoted liver regeneration in CCl4-induced liver injury (Lin et al., 2023).

Gupta et al. observed disrupted sinusoidal endothelium structure at 24 h post-HTx, suggesting endothelial disruption permits transplanted cell entrance into liver parenchyma (Gupta et al., 1999b; Gupta et al., 1999a). Therefore, facilitating sinusoidal endothelium disruption can enhance engraftment efficiency by promoting the entrance of hepatocytes into liver plates (Figure 2).

Small-molecule drugs including monocrotaline, doxorubicin, rifampicin, and phenytoin increased cell engraftment versus controls by causing hepatic endothelial damage (Joseph et al., 2006; Kim et al., 2005; Wu et al., 2008). Thalidomide has the anti-inflammatory properties via selective inhibitory activity of tumour necrosis factor-α (TNFα) and the anti-angiogenesis effect via downregulating VEGF levels. (Amare et al., 2021; Barbarossa et al., 2022). Viswanathan et al. found that Thalidomide was able to induce transient upregulation of serum hyaluronic acid levels versus control, suggesting the rapid endothelium damage (Viswanathan et al., 2016). Thalidomide pretreatment markedly upregulated transplanted cell numbers in liver versus non-treatment groups (Viswanathan et al., 2016), also improving engraftment via immune regulation discussed later. It is speculated that the effect of Thalidomide on hepatic endothelial damage is related to its anti-angiogenesis property. Thalidomide inhibited proliferation of human umbilical vein endothelial cells in vitro probably via downregulation of transcription factor SP1 (Moreira et al., 1999). Another study demonstrated that Thalidomide suppressed VEGF production by human peritoneal mesothelial cells and thereby inhibited endothelial tube formation via STAT3/SP4 signaling (Zhu et al., 2021). These data offer a mechanistic insight of how Thalidomide disrupt hepatic endothelial integrity, although further studies are needed.

In conclusion, using small molecules to induce endothelial disruption benefits cell engraftment. Mechanistically, endothelial disruption increases sinusoid permeability, allowing hepatocytes to pass the endothelial barrier and enter liver parenchyma.

Slehria et al. observed sinusoidal blood flow cessation in liver after HTx, persisting over 3 h and ameliorating by 24 h (Slehria et al., 2002). Researchers suggested that the overall sinusoidal perfusion declined from 94% to 84% on day 1% and to 76% on day 7 after HTx (Wilhelm et al., 2004), suggesting transplanted hepatocytes cause immediate and long-term disruption to sinusoidal blood flow and microcirculation.

Hepatic microcirculatory obstruction affects transplanted cell distribution. Cells entering sinusoids and integrating into liver plates avoid immune clearance, while those remaining in portal areas confront clearance (Gupta et al., 1999b; Koenig et al., 2005). Rajvanshi P et al. suggested hepatocyte location relates to portal vein radicles size. The intrasplenically-transplanted hepatocytes were largely located in periportal area (zone 1) in mice, while in downstream midlobular (zone 2) or perivenous (zone 3) areas in larger animals (Rajvanshi et al., 1996). Indeed, occlusion is more likely in smaller vessels versus larger ones, hindering sinusoidal distribution while increasing portal areas exposure, leading to immune clearance of transplanted hepatocytes.

Hepatic microcirculatory obstruction causes ischemia-related cell injury in the liver. After HTx, transient ischemia in portal areas leads to an immediate sinusoidal blood flow cessation (Wilhelm et al., 2004; Slehria et al., 2002). Wilhelm A et al. observed markedly increased Kupffer cells (KCs) on day 1 post-HTx (Wilhelm et al., 2004), which release ROS and proinflammatory cytokines including TNF-α, IL-1, IFN-γ (Peralta et al., 2013). ROS induces transplanted cell injury by inducing mitochondria dysfunction, protein synthesis disruption and cell membrane damage (Bardallo et al., 2022). The proinflammatory cytokines promote neutrophil and peripheral macrophage infiltration, which further enhance the immune reaction. Indeed, Gupta S et al. observed the infiltration of OX-43 antibody-positive granulocytes, phagocytes, and activated macrophages in portal areas (Gupta et al., 1999b). Overall, ischemia activates the immune response, leading to transplanted cell injury.

This evidence suggests that hepatic microcirculatory obstruction interferes with cell distribution and triggers immune responses. Thus, ameliorating the hepatic microcirculation with vascular dilators is a promising method to improve transplanted hepatocyte engraftment.

Nitroglycerine relaxes vascular smooth muscle cells and cause vascular dilation by increasing NO levels and activating guanylate cyclase. Slehria S et al. first reported that nitroglycerine treatment in rats significantly reduced periportal blood flow perturbation after HTx and increased transplanted hepatocytes in the liver and specifically in zone 2 at 2h after HTx but not long-term (Slehria et al., 2002), showing sinusoid dilatation contributed to hepatocyte distribution but not translocation. Other mechanisms may play a role in hepatocyte entry into liver plates. In another study, nitroglycerine significantly decreased KC activation and endothelial injury (Bahde et al., 2013), both of which are crucial in preventing ischemia-reperfusion injury. Bosentan (BOS), an unselective endothelin receptor blocker, increased desmin + hepatic stellate cells (HSCs) numbers and VEGF release of HSCs in rat (Bahde et al., 2013), potentially aiding cell entry via increasing hepatic sinusoid permeability. Moreover, conditioned medium of BOS-treated HSCs containing BOS and VEGF prevented TNF-α- or H₂O₂-induced cytotoxicity against hepatocytes (Bahde et al., 2013). BOS-incubated cells engrafted 2-fold higher than controls while systemic treatment of BOS in recipient rats did not (Bahde et al., 2013). A selective endothelin-1 receptor A blocker, Darusentan (DAR), was effective in dilating hepatic sinusoids, which allowed greater entry of hepatocytes and thus enhanced cell engraftment. Moreover, DAR reduced endothelial injury, KC activation and hepatic ischemia (Bahde et al., 2014). Contrary to BOS, DAR failed to prevent cytotoxicity and negatively affect hepatocyte proliferation in vitro (Bahde et al., 2014), suggesting that DAR favors systemic use. Noteworthy, the different effect of BOS and DAR on hepatocytes and host indicates the endothelin receptor A or B signaling pathway exert distinct biological functions in the liver, whereas the underlying mechanism warrants further investigations.

IBMIR was first described in islet transplantation which cause instant islet loss (van der Windt et al., 2007; Mou et al., 2022; Yan et al., 2022), involving interactions between platelet aggregation, coagulation system, complement system, and neutrophil/macrophage infiltration (Kale and Rogers, 2023). Hepatocyte-induced IBMIR occurs both in vitro and patients receiving HTx and is a dominant factor contributing to early cell loss of HTx (Lee et al., 2016).

Tissue factor (TF) is the initiating factor of coagulation extrinsic pathway, which expressed on cell membrane of mice and human hepatocytes (Kopec and Luyendyk, 2014). In the tubing loop system model, infusion of 5 × 10⁵ hepatocytes into the whole-blood significantly increased D-dimer levels and decreased platelet counts (Stéphenne et al., 2007). The increase of D-dimer levels was also observed in a Crigler-Najjar patient after the first infusion of hepatocytes (Stéphenne et al., 2007). The changes were prevented by anti-TF monoclonal antibody, suggesting the procoagulant activity of hepatocytes depended on TF (Stéphenne et al., 2007).

Complement system can be directly activated by hepatocytes through antibody recognition, lack of membrane-bound regulators, and extracellular matrix proteins exposure (Nilsson et al., 2014), and indirectly by the crosstalk with coagulation system. For example, fXa and thrombin can convert C3 into C3a (Dzik, 2019). TF-induced thrombin cleaves C5 at a novel R947 site and forms new fragments, C5_T and C5b_T (Krisinger et al., 2012). Reciprocally, complement proteins interact with coagulation system. C5a can elicit significant increase in TF expression of endothelial cells (Dzik, 2019). Activation of coagulation and complement contributes to the infiltration of neutrophils, monocytes and macrophages, resulting in hepatocyte loss.

Collectively, TF-initiated coagulation activates the IBMIR with interactions between coagulation, complement system, and immune response promote IBMIR process. Interventions with each step with small molecules can attenuate IBMIR and enhance cell engraftment (Figure 2).

In Fah^−/− mice, donor hepatocytes have a proliferative advantage for liver repopulation since host cell proliferation is impaired. (Overturf et al., 1996). Our lab demonstrated that host hepatocytes in Fah^−/− mice liver experienced senescence characterized by cell cycle arrest, which favored the proliferation of donor hepatocytes (Chen et al., 2019). Senescence disrupts cell connections and degrades extracellular matrix, providing engraftment space (Chen et al., 2019). Among patients with Crigler–Najjar disease where the proliferation of host hepatocytes is impaired, transplanted hepatocytes successfully alleviated the liver functions (Jorns et al., 2012). Moreover, in α1-antitrypsin deficiency mice models with inhibited host hepatocytes proliferation, wild-type donor hepatocytes replaced 20%–98% of host hepatocytes (Ding et al., 2011). However, in most inherited metabolic disorders host hepatocytes viability is not significantly compromised (Barahman et al., 2019a). The proliferation advantages can be achieved by: (1) inhibit host cells proliferation and (2) promote transplanted cells proliferation. For the former one, several precondition strategies including hepatic irradiation, portal embolization, and partial hepatectomy are used in clinical trials (Nguyen et al., 2020). However, these methods are invasive to patients’ body. Using small molecules to intervene with host and donor hepatocytes viability is an ideal, noninvasive, and controllable method.

The Hippo pathway, critical for liver regeneration, comprised of MST1/2 phosphorylating and activating the LATS1/2, which then phosphorylates and inactivates YAP and TAZ. Deletion of MST1/2 or LATS1/2 genes increased nuclear level of YAP/TAZ and expression of downstream genes (Driskill and Pan, 2021). Fan F et al. identified a MST1/2 inhibitor, XMU-MP-1, effectively suppressed MST1/2 kinase and thereby activated downstream YAP in various cells. In Fah^−/−Rag2^−/−IL2rg^−/−(FRG) mice treated with XMU-MP-1, transplanted hepatocyte proliferation was remarkably promoted, improving liver repopulation and ALB levels (Fan et al., 2016).

HGF/c-Met signaling pathway plays an important role in liver regeneration. HGF binds to its specific receptor c-Met, activating downstream pathways and leading to cell proliferation (Zhao et al., 2022). It was reported that the c-Met receptor agonist antibody 5D5 induced significant proliferation in hiPSC-derived hepatocyte-like cells (hiPSC-HLCs) in vitro (Yuan et al., 2019). In FRG-SCID mice, 5D5 remarkably promoted liver repopulation rate of hiPSC-HLCs(Yuan et al., 2019). Small molecule agonists targeting c-Met receptor warrant further development.

Jiang M et al. observed that the HPC and cell cycle genes significantly increased from day 4 after HTx in Fah^−/− mice, peaked at day 30 and then decreased to normal. The labeled hepatocytes expanded at day 30 and almost repopulated the liver at D120 (Jiang et al., 2023). These results suggested that the transplanted hepatocytes dedifferentiate to an HPC stage to proliferate and then re-differentiate to hepatocyte after repopulation. Combination of ROCK inhibitor Y27632 and CHIR99021 (termed YC) markedly increased the expression of HPC and cell cycle genes in vivo and thereby promoted hepatocytes proliferation in treated mice (Jiang et al., 2023). Clinically used small molecule drugs of ROCK signaling Netarsudil (N) and Wnt signaling LY2090314 (L) were also tested. Similar to YC, NL treatment also promoted the dediffereantiation and redifferentiation of hepatocytes and favored the liver repopulation in vivo (Jiang et al., 2023).

Chemical induction of hepatic fate from pluripotent stem cells (PSCs) can provide an unlimited hepatocyte source. The insights into the signaling pathway and transcription factor network of liver development provide cues for yielding hepatocyte-like cells (HLCs) from PSCs by chemical modulation. During embryonic liver development, bone morphogenic protein (BMP), Wnt/β-catenin and Notching signaling promote definitive endoderm (DE) formation; fibroblast growth factors (FGFs) and BMPs promote hepatoblasts (HBs) specification from DE; Notch and TGFβ signaling inhibition and Hippo signaling activation drive HBs into hepatocyte fate. Downstream of signaling pathways, transcription factors including FOXAs, GATA4/6, HNF4α, and HNF1 interact with each other and activate liver-specific gene expression, accompanied by epigenetic changes. For instance, pioneer factors recruit the histone acetyltransferase p300 to deposit H3K9ac/14ac at liver-specific gene regulatory elements, directing hepatic programs (Xu et al., 2011). To recapitulate liver development, the differentiation of PCSs comprises three main steps: ‘PSCs to DE’, ‘DE to hepatic progenitor cells (HPCs)’, and ‘HPCs to HLCs’. Stepwise chemical approaches derive HLCs by regulating these developmental paradigms and epigenetic landscapes. Additionally, reopening liver genes with small molecules enables direct fibroblast reprogramming (Rombaut et al., 2021; Park et al., 2019; Bai et al., 2023).

A comparative study showed that the chemically induced hepatocytes resemble growth-factor derived counterparts (Gao et al., 2020). The expression of liver drug-metabolizing enzymes, transporters, and nuclear receptors in primary hepatocytes are significantly higher than that in HLCs (Gao et al., 2020).

Despite the progress made in chemical induction of hepatocyte, it is noteworthy to mention that consensus on optimal small molecule combinations, doses and hepatocyte quality is lacking, hindering protocol standardization and clinical translation. Future work establishing current good manufacturing practice (cGMP)-grade protocols and improving functional stability is needed for transplantation applications. Comparing differential gene expression between primary and induced hepatocytes may identify targets to enhance cell competence for transplantation.

In general, there are two strategies to improve low cell engraftment: prime the host hepatic microenvironment for HTx (hepatic irradiation and partial hepatectomy) and prime the donor hepatocytes (alginate encapsulation) (Nguyen et al., 2020). Hepatic irradiation (HIR) has manifested its efficiency in animal preclinical studies and clinical trials (Soltys et al., 2017; Barahman et al., 2019b). However, the efficiency depends on irradiation parameters and host radiosensitivity. It is noteworthy that HIR has hepatoxicity and may cause side effects in patients. Therefore, combination of HIR and systemic small molecules including anti-IBMIR and chemotherapeutic drugs may reduce dosage of both radiation and small molecules, lowering the risks of adverse drug reactions. Recent studies have shown that hepatocyte microbeads encapsuled by alginate are protected from immune rejection and improve cell engraftment (Jitraruch et al., 2014). A recent clinical trial has validated its efficiency and safety in pediatric patients with ALF (Dhawan et al., 2020). However, other problems involved in HTx such as microcirculatory obstruction cannot be tackled with alginate encapsulation. It is interesting to envisage that load alginate encapsulation with small molecules can improve cell engraftment.

Challenges to chemical approaches still remains to be solved before wide applications, including drug specificity, target accuracy, pharmacokinetic properties, and safety. For instance, CHIR99021, a GSK3β inhibitor that activates Wnt/β-catenin signaling, is widely used in regulating cell differentiation, reprogramming, and proliferation. However, Wnt/β-catenin signaling activation promotes the development of liver cancer and the resistance to immunotherapy (Zhou et al., 2022; He and Tang, 2020). Hippo pathway play critical role in liver physiology and tumorigenesis. Depletion of MST1/2 drives the formation of hepatocellular carcinoma and/or cholangiocarcinoma (Driskill and Pan, 2021). Whether in vivo administration of Wnt/β-catenin (CHIR990221) and Hippo signaling activators (XMU-MP-1) will lead to tumorigenesis or affect tumor development in chronic liver failure recipients with hepatocellular carcinoma needs to be studied. TGFβ signaling, ROCK signaling, and others are ubiquitous and regulates numerous critical physiologic processes in healthy cells. How to minimize the systemic effect while maximize the hepatic effect of in vivo administration of small molecules is an important problem before the large-scale clinical use. Developing liver-specific or hepatocyte-specific drug delivery system is one of the feasible strategies, including nanoparticle-based and small molecule-based drug delivery systems (Unagolla et al., 2024; Sharma et al., 2021). Off-target effects are another challenge of clinical application of small molecules. In vitro screening is insufficient to mimic the in vivo complicated microenvironment. Therefore, small molecules may cause off-target effects in vivo despite its in vitro efficacy. Target prediction models based on deep learning are prosperous strategies to mitigate off-target effects.

Despite the obstacles, some small molecules mentioned in this review have already applied clinically, whose application in HTx can be quickly translated due to its confirmative safety and efficiency in human (Table 3). For instance, ripasudil has already been licensed for treating ocular hypertension and open-angle glaucoma, which has potential for rapid translation as a therapeutic to promote hepatocyte engraftment (Table 3). It is noteworthy that systemic adverse drug reactions need further investigation if ripasudil is used systemically, as it rarely enters systemic circulation through the blood-ocular barrier when administered as eye drops in current clinical trials (Kusuhara and Nakamura, 2020). The indication and clinical trials of other small molecules in this review was accessible in Table 1 and 2. In the future, studies comparing various chemical protocols of hepatocyte differentiation and reprogramming are urgently needed in order to standardize a paradigm protocol for potential clinical trials. Additionally, studies on simultaneous interventions of each process during hepatocyte engraftment are warranted, because the reason of low cell engraftment is multifactorial. We believe that the in-depth probes into stem cell biology and progress in chemical screenings and deliver systems will promote the application of HTx by ingenious chemical formula in the future.