TGF-β is the most well-known hepatocyte proliferation inhibitor and stop signal in the process of hepatic regeneration. There have been investigations on the association between TGF and liver regeneration for >30 years. Nakamura et al reported in 1985 that TGF-β from platelets inhibited the DNA synthesis of adult rat hepatocytes induced by epidermal growth factor (EGF) in primary culture (4). The administration of platelet-derived TGF-β has also been shown to suppress liver regeneration in a two-thirds PH rat model (5). The inhibition of TGF-β1 with monoclonal antibody has been shown to enhance liver regeneration in a porcine model of partial portal vein ligation (6). A previous study also found that inhibiting the TGF-β pathway with the TGF-β type I receptor kinase inhibitor increased hepatocyte proliferation during acute liver damage (7).

Previous investigations demonstrated that TGF-β is predominantly secreted by nonparenchymal cells, including hepatic stellate cells (HSCs), Kupffer cells (KCs) and platelets, in liver regeneration and modulates the proliferation of hepatocytes in a paracrine manner (8,9). Subsequent studies reported that TGF-β can also be expressed in regenerating hepatocytes (10,11). The expression of TGF-β was found to increase at 4 h and reached a peak 72 h following PH when DNA synthesis had stopped, suggesting that TGF-β was involved in the inhibition and termination of liver regeneration (8,12). Further evidence for the anti-proliferation function of TGF-β1 includes the reduction in DNA synthesis in transgenic mice with overexpressed TGF-β1 following PH (13). Although several members are included in the TGF-β superfamily, the β1 form is the most important subtype in regeneration (10,12).

TGF-β modulates the proliferation of hepatocytes through multiple mechanisms. The earliest mechanism to be identified is the inhibition of EGF-induced DNA synthesis by TGF-β following its binding to TGF-β receptors on the surface of hepatocytes (8,11,14). Investigations found that TGF-β receptors were single-pass transmembrane serine/threonine kinases. When the TGF-β ligand binds to the complex of TGF-β receptors, the receptor-regulated cytoplasmic small mothers against decpentaplegic (R-Smad) proteins are phosphorylated and accumulate in the nucleus. The accumulated Smads subsequently inhibit the transcription of several genes by interacting with DNA binding proteins and transcriptional regulators (15–18). Another study showed that the TGF-β signaling may be inhibited by SnoN and Ski, Smad pathway inhibitors, during the proliferative phase of liver regeneration (19). Increased expression of TGF-β and activation of Smad2 have also been found in a T cell-mediated hepatitis model following PH, and were associated with impaired hepatocellular proliferation (20). The activation of Smad2/3, which can be inhibited by Smad7, is required for TGF-β to have an inhibitory role in liver regeneration (21). Only activated TGF-β has the ability to inhibit cell proliferation, as reported by Schrum et al, who observed that activated TGF-β, but not latent TGF-β, affected liver regeneration (22). A previous study found that the activation of TGF-β required the involvement of thrombospondin-1 (TSP-1). Significantly reduced TGF-β/Smad signaling and accelerated hepatocyte proliferation were observed in mice with TSP-1 deficiency following PH (23).

With the exception of EGF, other proliferation-associated factors and signaling pathways can be affected by TGF-β. TGF-β can inhibit the secretion of human hepatocyte growth factor (HGF), which is known to be a complete mitogen in the process of liver regeneration (24). DNA synthesis induced by TNF can also be inhibited by TGF-β in cultured hepatocytes (25). The overexpression of TGF-β1 was shown to inhibit the protein level of cdc25A, a cyclin-dependent kinase-activating tyrosine phosphatase, by enhancing the binding of histone deacetylase 1 to p130 in Alb-TGF-β1 transgenic mice following PH (26). TGF-β1 was found to inhibit DNA synthesis of the G1 stage in cultured hepatocytes treated with HGF and heparin-binding epidermal growth factor-like growth factor by decreasing the expression of cyclin E without affecting the activity of c-Met, epidermal growth factor receptor (EGFR) or mitogen-activated protein kinase (MAPK), or the expression of cyclin D1 (27). This indicated that TGF-β1 has the ability to restrain the growth factor-induced signals between cyclin D1 and cyclin E (27).

Inducing cell apoptosis is important for TGF-β to exert its anti-proliferative effect in liver regeneration. Studies have indicated that liver cell regeneration and atypical bile duct proliferation are associated with TGF-β in fulminant hepatitis (28). The overexpression of activated TGF-β by adenovirus vectors enhanced the apoptosis of hepatocytes, resulting in the death of rats following two-thirds PH (22). Enhanced apoptosis was also found to be accompanied by the overexpression of TGF-β1 in tetracycline-controlled TGF-β1 mice (29), whereas pro-apoptotic signals were suppressed by insulin-like growth factor binding protein-1 (30). In experiments performed by Samson et al, the overexpression of TGF-β1 resulted in increased expression of c-Jun, a potential pro-apoptotic transcription factor, which indicated that TGF-β1 induced hepatocyte apoptosis through a c-Jun-independent mechanism (31). The expression of TGF-β1 increases the generation of ROS, an essential mediator of apoptosis, and induces the apoptosis of hepatocytes in liver regeneration, which is associated with promoting the generation of ROS from mitochondria and inducing cytochrome P450 1A1 (32).

TGF-β is also the target of other proliferation inhibiting factors in liver regeneration. Cation-independent mannose 6-phosphate receptor (CIMPR), an indirect negative regulator of hepatocyte growth, which has progressively overexpressed in liver regeneration 8 h after PH, mediates the activation of latent pro-TGF-β (33). A schematic diagram of the TGF-β signaling pathway is shown in Fig. 1.

Activin, another member of the TGF-β superfamily, which is structurally related to TGF-β, also has an inhibitory effect on hepatocyte proliferation and liver regeneration. Activin A, with a similar ability to that of TGF-β, inhibits EGF-induced DNA synthesis in an autocrine manner without competing with TGF-β (34). A study by Schwal et al indicated that activin induced the apoptosis of hepatocytes in a different manner from that of TGF-β (35). The administration of follistatin, an activin-binding protein that can inhibit the activity of activin, enhances liver regeneration following PH (35–37). The inhibitory effect of activin may also be inhibited by increased expression of SnoN and Ski, which are Smad pathway inhibitors, in early liver regeneration (19). The impairment of activin type I receptors has also been shown to increase DNA synthesis in hepatocytes, providing further evidence of the anti-proliferative function of activin (38).

IL-1 is another significant negative regulator of cell proliferation in the process of liver regeneration. In 1988, Takahashi et al first reported that IL-1 may be involved in the suppression of liver regeneration on examining syngeneic spleen cells treated with poly I:C in a mouse PH model (39). Another study by Nakamura et al showed that IL-1β markedly inhibited DNA synthesis induced by insulin and EGF in cultured rat hepatocytes (40). Similar inhibition of cultured hepatocytes has also been found in other studies (41–43), and enhanced liver regeneration by FK506 can be inhibited by IL-1 (44). The increased expression of IL-1 impairs regeneration via an indirect mechanism in liver injury induced by endotoxin following hepatectomy (45). The mRNA expression of IL-1α in the whole rat liver was found to be downregulated at 10 h, and upregulated 24 and 48 h following PH, suggesting an association between the expression of IL-1 and liver regeneration (46). Further experiments have revealed that sensitivity of hepatocytes isolated from the liver 24 h following PH was increased to the inhibition of IL-1, and the administration of exogenous IL-1β suppressed DNA synthesis at 0 and 12 h post-PH (46). Long-term examination of the expression of IL-1 in the liver in liver regeneration showed that the mRNA levels of IL-1 increased slowly and marginally following 30 and 80% hepatectomy (47). The increased expression of IL-1β was has also been observed in a shrinking liver lobe of a rat portal vein ligation model, indicating that IL-1β was involved in the process of cellular atrophy (48,49). The expression of IL-1β was also increased in the regenerating impaired liver caused by deficiency in the expression of peroxisome proliferator-activated receptor α (50). The suppression of IL-1β by FR167653, a selective inhibitor, was shown to promote liver regeneration in rat models of classic PH (70%) and extended hepatectomy (90%) (51,52). Hepatocyte proliferation was also found to be enhanced by recombinant human interleukin 1 receptor antagonist in a mouse model of acute liver injury induced by CCl₄ (53).

IL-1 is secreted by nonparenchymal cells (NPCs) in the regenerating liver. Goss et al found that IL-1 is produced by KCs and regulated by prostaglandin E2 (PGE2) in the rat liver following PH (54,55). Heparin and PGE1 also suppress the expression of IL-1 in KCs following PH (45). Increased mRNA levels of IL-1α and β, which are associated with KCs and the spleen, were reported in liver NPCs between 30 min and 1 h following PH (56). Medium conditioned by nonparenchymal cells isolated from the regenerating liver inhibit DNA synthesis in primary rat hepatocytes induced by HGF, EGF and TGF-α. This suppression is inhibited by either IL-1 receptor antagonist or by IL-1 α and β antibodies (46).

Although studies on the association between IL-1 and liver regeneration have been performed, the inhibitory mechanism of IL-1 on liver regeneration remains to be fully elucidated. Several studies have shown that IL-1 receptor antagonist (IL-1Ra), a competitive inhibitor of IL-1α and IL-1β, and anti-inflammatory protein, inhibited the inhibition of IL-1 and facilitated liver regeneration, suggesting that IL-1R was required for the inhibitory action of IL-1 (53,57–59). IL-1 was found to contribute to the impairment of liver regeneration by reducing HGF and promoting TGF-β release in reduced-size orthotopic liver transplantation models (59). A previous study showed that IL-1β inhibits the fibroblast growth factor (FGF)19 signaling pathway, which regulates cell growth and metabolism of hepatocytes in liver regeneration (60). IL-1β also inhibits the expression of β-Klotho, a co-receptor of FGF receptor 4 (FGFR4), with involvement of the c-Jun N-terminal kinase (JNK) and nuclear factor (NF)-κB pathways. The activation of extracellular signal-regulated kinase (Erk)1/2 and hepatocyte proliferation induced by FGF19 is also suppressed by IL-1β.

IL-6 is a multifunctional cytokine, which has been investigated extensively. It is a known classical triggering signal of liver regeneration following PH. The essential role of IL-6 and its mechanisms in the process of liver regeneration have been confirmed (61). IL-6 has been shown to act as a positive regulator of liver regeneration in the majority of studies, whereas it has also been shown that IL-6 impairs liver regeneration, which conflicts with the mainstream findings. IL-6 inhibits mature hepatocyte proliferation and DNA synthesis stimulated by TNF and EGF in primary culture (25,42). Beyer and Theologides showed that IL-6 suppressed the DNA synthesis of hepatocytes at 3 days in vitro (43). The inhibitory effect of IL-6 in hepatocyte proliferation may be associated with induction of the expression of p21 in vitro (62).

Upregulated levels of IL-6 have been observed in a number of abnormal liver regeneration models. Increased blood levels of IL-6, which induce the expression of protein inhibitor of activated Stat-3 and the suppressor of cytokine signaling (SOCS)-1, were found to be associated with impairment of liver regeneration in a model of hepatic failure (63). Hyperstimulation with IL-6 also suppressed hepatocyte proliferation by inducing the expression of p21 in transgenic mice overexpressing the human soluble IL-6 receptor/gp80 following PH (64). The overexpression of STAT3, a downstream molecule of IL-6, was found to impair hepatocyte proliferation in fatty liver following PH (65). In addition, the suppression of hepatocyte proliferation was accompanied by increased levels of IL-6 in IL-1Ra-knockout mice following PH (57). The impairment of liver regeneration by hepatitis B virus X protein (HBx) via upregulating the expression of IL-6 was observed in an HBx-overexpressed transgenic mouse model (66). In addition to the data obtained in animal models, elevated levels of IL-6 have also been found in patients with chronic liver disease (67–70).

Although the mechanisms of IL-6 in liver regeneration are controversial, the majority of associated studies support the positive role of IL-6, with IL-6 affecting hepatocyte proliferation in a dose- and time-dependent manner (71,72). The studies mentioned above also indicated that the inhibitory effects of IL-6 in liver regeneration, particulary on a background of liver disease, may be associated with its hyperstimulation of liver cells.

SOCS is a family of proteins, which negatively regulate various cytokine signals (2). SOCS inhibits the activity of STAT3, a downstream signal of cytokines and growth factors in hepatic regeneration. SOCS3, a feedback inhibitor of the IL6/JAK/STAT3 pathway, is an important member of the SOCS family, which has received the most attention in investigations of liver regeneration. SOCS3 is transiently upregulated in the livers of mice following PH and is involved in the inactivation of STAT3 signaling (73–75). SOCS3 can inhibit the phosphorylation of gp130, a component of the IL-6 receptor, JAKs and STATs (76). SOCS3 hepatocyte-specific-knockout mice showed increased capability of cell proliferation and prolonged STAT3 phosphorylation following PH, with enhanced activation of ERK1/2 also observed (77). SOCS3 can also negatively control the proliferative responses mediated by HGF and EGF (78). The overexpression of SOCS3 suppresses IL-22 signaling, another cytokine positively correlated with liver regeneration, via eliminating the IL-22-induced activation of STAT (79). The expression of SOCS3 is primarily induced by IL-6 and is also upregulated by other factors, including TNFα, IL-1, IL-22 and ZIP14, a zinc transporter (73,79–81).

SOCS1, another member of the SOCS family, has a similar but weaker function to that of SOCS3 (74,78). Unlike SOCS3, SOCS1 does not affect the growth signaling mediated by EGFR (78). A previous study indicated that SOCS1 negatively regulated the hepatocyte proliferation induced by HGF by inhibiting c-Met signaling (82).

IFN-γ can downregulate hepatocyte proliferation by activating STAT1 and its downstream genes, including RF-1, p21 and SOCS1, in liver regeneration (83). The inhibitory effect of invariant natural killer T (iNKT) cells, a major lymphocyte in the liver, was also associated with IFN-γ (84). iNKT cells affect liver regeneration to a lesser degree. The deficiency of iNKT cells (CD1d and Jα281 cells) does not alter mouse liver regeneration following PH. However, treatment with α-galactosylceramide, an inducer of iNKT cell activation, impairs hepatocyte proliferation via upregulating the expression of IFN-γ in wild-type mice. Experiments also indicated that IL-4 is the primary mediator in the activation of iNKT cells and induction of IFN-γ (84).

There are other cytokines, which negatively regulate liver regeneration. TNF-α, another pro-proliferation factor in liver regeneration, also has the ability to induce the apoptosis of adenovirus-infected hepatocytes in a coupled TGF-α-IL-1α/β-IL-1ra autocrine cascade manner (85). The administration of α2b-INF inhibits DNA biosynthesis and liver mass restitution by affecting the expression of cell cycle-associated genes, includin c-myc, p53 and c-erbB-2, in rat PH models (86). Further experiments have indicated that the suppression of hepatocyte proliferation depends on the time at which α2b-IFN is administered (87), and the proliferation of hepatocytes was found to be inhibited by granulocyte colony-stimulating factor via upregulating the expression of IL-1β in a liver injury model induced with dimethylnitrosamine (88). Mice with single deficiency of myeloid differentiation factor 88 (MyD88), an adaptor protein for the majority of Toll-like receptors, presented with enhanced hepatocyte proliferation at 32 and 36 h post-PH (89). The decreased activation of STAT3 and induction of SOCS3 have also been detected in MyD88-null mice following PH (89). These results suggest that MyD88 has an antiproliferative effect in liver regeneration by affecting IL-6 signaling pathways. A summary of the cytokines involved in inhibiting liver regeneration is presented in Table I.

The role of p53, a negative regulator of growth and tumor suppressor genes, is complex in the process of liver regeneration and conclusions from different studies are contradictory. The expression of p53 is upregulated, corresponding to the G1/S transition, in the early stage of liver regeneration following PH (9,90). It has been shown that the overexpression of p53 by a transducing adenoviral vector encoding wild-type p53, did not affect the process of liver regeneration (91), however, other studies showed that p53 may positively regulate hepatocyte proliferation in rats. The suppression of the expression of p53 results in the reduction of DNA replication in hepatocytes in vitro (92). A study by Inoue et al showed that the levels of p53 increased in cultured hepatocytes induced with HGF, and inhibiting the expression of p53 by an antisense oligonucleotide resulted in a decrease in the levels of TGF-α and of DNA synthesis (93). It was also found that upregulation of the expression of p53 was synchronous with HGF followed by an increase of TGF-α in a rat PH model, suggesting that p53 is positively correlated with liver regeneration (93).

However, other findings have supported that p53 negatively controls liver regeneration. Previous studies have provided indirect evidence of the anti-proliferation functions of p53. Suppressing p53 with antisense oligonucleotides enabled hepatocytes to recover from growth arrest induced by ultraviolet (94). Increased expression of p53, accompanied by a downregulation in the levels of proliferating cell nuclear antigen (PCNA), was observed in rats treated with a metallothionein antisense oligonucleotide, suggesting an inverse correlation between p53 and liver regeneration (95). The detailed role of p53 in liver regeneration was examined by Arora and Iversen (96). Increased weight gain of remnant-regenerating liver, CYP isoform activities and expression of PCNA were observed in rat PH models, in which the expression of p53 was inhibited with an antisense oligonucleotide, whereas the G₁ cell populations and levels of P21 were also reduced in the p53-inhibited rats, indicating that p53 controls hepatocyte proliferation in a cell cycle checkpoint manner (96). Furthermore, loss of the expression of p53 resulted in early entry into the cell cycle and prolonged the proliferation of hepatocytes in p53 (−/−) mice following PH, and the inhibitory mechanism of p53 was correlated with the expression of mitotic genes, including Forkhead box M1, Aurka, Large tumor suppressor kinase 2, Polo like kinase (Plk)2 and Plk4 (97). Abrogation of the p53 pathways also facilitates liver regeneration in Wild-type p53-induced phosphatase 1-deficient mice following PH (98).

P53 can also suppress cell proliferation in regeneration of damaged liver on a background of chronic disease or drug treatment. An increased expression of p53 followed by the induction of p21 was associated with the inhibition of hepatocyte proliferation in PH models treated with 2-acetaminofluorene (99). Impaired liver regeneration and increased expression of p21 following partial PH was observed in mice lacking c-Jun, a key regulator of hepatocyte proliferation, whereas the activation of p53 eliminates the effect, indicating that p53 is a proliferation inhibitor in liver regeneration (100). The activation of a p53-dependent checkpoint correlates with premitotic arrest and abnormal cell death in transgenic mice overexpressing human Aurora-A following PH (101). The expression levels of p53 and p21 are also upregulated, coinciding with cell cycle arrest and apoptosis in hepatocytes of liver injury models induced by acetaminophen (102–104).

P21, the expression product of the ras gene, also termed wide-type 53-activated factor 1 or cyclin-dependent kinase interacting protein 1, has been identified as a cyclin-dependent kinase inhibitor and downstream gene of p53 (105–107). The negative control role of p21 has been confirmed in various studies. The expression of p21 increased maximally following peak hepatocyte DNA synthesis and cell division, later than that of p53, in rat and mouse PH models, indicating its association with the proliferation of liver cells (90,108). The upregulation in levels of p21 have been examined not only hepatocytes, but also nonparenchymal cells, in the process of liver regeneration (109). In addition to the data mentioned above, overexpressed p21 also coincides with the arrestment of liver regenerating in several disease and drug toxicity models (65,110–115).

Although p21 is a classic downstream signal of p53, the expression of p21 is induced by other factors in addition to p53 in liver regeneration. Increased mRNA levels of p21 induced by p53 were identified and involved in inhibiting hepatocyte proliferation in 2-acetaminofluorene-treated rats (120). Upregulated levels of p53 and p21 were also correlated with impaired liver regeneration in acute liver injury models induced by acetaminophen, and inhibition of the p53/p21 pathway promoted hepatocyte proliferation (102,121). However, no difference in the gene expression of p21 was found between p53-deficient and wild-type mice following PH, although the post-transcriptional regulation may be associated with p53 (116). Furthermore, treatment with TGF-β and activin A upregulated the expression of p21 in cultured primary hepatocytes (116). These data suggested that the induction of p21 is controlled in a p53-dependent and independent manner. Similar expression and localization of p21 was also found in wild-type and p53-null mice treated with carbon tetrachloride, and the increased levels of p21 might be correlated with CCAAT/enhancer-binding protein (C/EBP) (117). Other experiments have shown that levels of p21 were regulated by C/EBPα (122). Increased levels of p21 were observed in c-Jun-knockout mice following PH, and inactivation of the p53/p21 signaling pathway abrogated the suppression of liver regeneration, suggesting that the expression of p21 may be regulated by c-Jun (100).

Important in the proliferation of tumor cells, Bcl-2 also inhibits hepatocyte proliferation in liver regeneration. Bcl-2 transgenic mice show delayed hepatocyte DNA synthesis following PH, and a similar result is observed in cultured liver cells. In addition, the overexpression of Bcl-2 is followed by decreased expression levels of p107 and cyclin E (123). These results indicate that Bcl-2 may be a negative cell cycle modulator in liver regeneration.

P53 upregulated modulator of apoptosis (PUMA), another member of the Bcl-2 family, is also negatively correlated with liver regeneration. PUMA is downregulated, accompanied by increased apoptosis and decreased proliferation, in the early phase post-PH (124). Bcl-2-associated X protein (Bax) inhibitor-1 (BI-1), which has a similar function to that of Bcl-2 family proteins, also regulates liver cell proliferation in vivo. Enhanced liver regeneration, accompanied by the increased expression of cell cycle regulators, including cyclin D1, cyclin D3, Cdk2 and Cdk4, have been observed in BI-1-deficient mice following PH (125).

Several studies have shown that additional oncogenes can inhibit liver regeneration. P27 is also a tumor suppressor gene and cyclin-dependent kinase inhibitor. P27 was shown to be minimally expressed in the mouse liver following PH in a time-dependent manner, and p27 decreased the activity of CDK2 prior to and following peak DNA synthesis, suggesting its negative role in liver regeneration (108).

Fas, also known as APO-1 or CD95, and Fas ligand (FasL) are essential in cell apoptosis and regulate the proliferation of tumor cells (126). The expression of Fas decreased significantly immediately in the rat liver and had recovered gradually to a normal level of quiescent hepatocytes 14 days following PH, suggesting that the Fas/FasL system was also negatively correlated with liver regeneration (127).

N-Myc downstream-regulated gene 2 (NDRG2), is also a known tumor suppressor gene and its expression is reduced or undetectable in various types of tumor (128). The significant downregulation of NDRG2 was examined in regenerating livers, and the overexpression of NDRG2 resulted in a cell cycle arrest via upregulating the p53/p21 pathway and inhibiting cyclin E-Cdk2 activity (129). These data indicate that NDRG2 has a potential negative effect in liver regeneration. A summary of the oncogenes involved in inhibiting liver regeneration is presented in Table II.

Transcription factors important roles in liver regeneration, however, the majority of them positively regulate liver regeneration. Other transcription factors, which may have inhibitory roles in liver regeneration, are less well understood.

C/EBPα, a member of the CCAAT enhancer-binding protein family, is expressed abundantly in the normal liver. However, the expression of C/EBPα was found to decrease markedly in proliferating cells following PH and in cultured hepatocytes stimulated with EGF and insulin, suggesting an inverse correlation between C/EBPα and liver regeneration (130). Positive correlation between p21 and C/EBPα has been observed in C/EBPα-knockout mice. C/EBPα did not affect the mRNA levels of p21 but inhibited the proteolytic degradation of p21, indicating that C/EBPα inhibited hepatocyte proliferation via controlling the expression of p21 in a post-transcriptional manner (122). In aged animal PH models, C/EBPα also suppressed the expression of E2F via forming the C/EBPα-Rb-E2F4 complex, which can bind to E2F-dependent promoters, inhibiting the expression of E2F and inhibiting cell proliferation (131).

Forkhead box (Fox)O1, a transcription factor regulated by Akt, may also be an anti-proliferative factor in liver regeneration. Deleting Akt1 and Akt2 was shown to result in impairment of liver regeneration in mice following PH, whereas liver-specific deletion of FoxO1 reversed this effect, indicating that FoxO1 has a negative effect on liver regeneration (132).