Roles of CCAAT/Enhancer-binding Proteins in Regulation of Liver Regenerative Growth*

The expressions and activities of several CCAAT/enhancer-binding proteins (C/EBP) isoforms fluctuate in the regenerating liver. The physiological implications of these variations in C/EBP function remain poorly characterized in the setting of regeneration. However, lessons learned in various hepatocyte cell lines and by studying primary hepatocytes from transgenic C/EBPα-deficient mice suggest that the C/EBP isoforms are likely to influence proliferation, differentiated gene expression, and survival in mature, adult hepatocytes. In addition, these factors are potentially important modulators of liver nonparenchymal cell genes, including those that encode matrix molecules and growth factors that are required for successful liver regeneration. The possibility that members of the C/EBP family of transcription factors actively participate in many aspects of the regenerative response to liver injury is strengthened by growing evidence that many hepatocyte mitogens and co-mitogens regulate C/EBP activity. Furthermore, the C/EBPs themselves appear to regulate the expression of some of these growth regulators.

Liver regeneration is a complex physiological response to liver injury during which the surviving, mature hepatocytes and liver nonparenchymal cells proliferate to reconstitute the pre-morbid mass of the damaged organ. The liver is one of the few tissues in adult vertebrates that retains the capacity for regeneration. However, such compensatory hyperplasia requires careful orchestration to assure that vital liver-specific functions are preserved while hepatocytes, which generally do not proliferate in adult animals, replicate. The study of organisms that are as divergent as urodeles and humans has identified common, fundamental processes that are required for the regeneration of any tissue. This work indicates that regeneration occurs only if injury-related changes in the local extracellular environment motivate differentiated, but proliferationcompetent, cells to reenter the cell cycle while the normal homeostatic mechanisms that couple cell cycle reentry to cell death are suspended (1)(2)(3). Although much has been learned about the mechanisms that regulate hepatic regeneration, many aspects of the process remain poorly understood.

CCAAT/Enhancer-binding Proteins and Hepatocyte Proliferation
CCAAT/enhancer-binding proteins (C/EBPs) 1 are known to play important roles in regulating the expression of multiple hepatocyte-specific genes (4). These transcription factors also help to control hepatocyte progression through the cell cycle (5,6). Thus, the C/EBPs are likely to be important targets for regulation during liver regeneration. Consistent with this concept, variations in the expression of C/EBP mRNAs, proteins, and DNA binding activities have been documented during liver regeneration. C/EBP␣ is the predominant C/EBP isoform that is expressed by adult hepatocytes in healthy livers. During the initial 24 h after 70% (partial) hepatectomy (PH), levels of C/EBP␣ mRNA and protein decline transiently. This is associated with decreased binding activity of C/EBP␣ homodimers in gel mobility shift assays. Decreases in C/EBP␣ are preceded by the induction of other C/EBP isoforms. The level of C/EBP␤ and C/EBP␦ mRNAs, proteins, and DNA binding activities increases in the early pre-replicative period following PH and returns to base line as hepatocytes enter S phase (7)(8)(9)(10).
The importance of these reciprocal variations in C/EBP␣ and other C/EBP isoforms in regulating hepatocyte proliferation during liver regeneration has been debated. Present controversy stems from an apparently poor correlation between the induction of C/EBP␤ DNA binding activity and hepatocyte proliferation in transgenic mice strains with disruption of certain cytokines or cytokine receptors. Mice homozygous for a deletion in the gene for the type 1 TNF receptor have extremely limited hepatocyte proliferation after PH and yet exhibit apparently normal induction of total C/EBP DNA binding activity during the early pre-replicative period (11). In contrast, transgenic mice lacking the gene for type 2 TNF receptors demonstrate normal hepatocyte proliferation after PH but decreased induction of C/EBP␤ DNA binding activity (12). Furthermore, liver regeneration is severely inhibited in IL-6-null mice, although these animals exhibit normal induction of C/EBP␤ mRNAs after PH (13). However, because C/EBP function is regulated, to a large extent, by post-transcriptional modifications of the C/EBPs themselves, as well as by their interactions with specific dimerization partners (14 -21), the former observations do not preclude an important role for the C/EBPs as regulators of hepatocyte proliferation during liver regeneration. Indeed, other evidence supports the biological importance of the C/EBPs as physiologically relevant regulators of hepatocyte proliferation. Transgenic mice with a deleted C/EBP␣ gene have increased hepatocyte proliferation at birth (22,23). Because these animals do not survive long enough to permit evaluation of their response to PH, whether or not C/EBP␣ deletion enhances hepatocyte proliferation during liver regeneration has not been tested. However, hepatocytes isolated from these mice do have increased proliferative activity in culture (24). Conversely, enforced overexpression of C/EBP␣ has been shown to inhibit proliferation in several different hepatocyte cell lines (5,25,26). Thus, it seems reasonable to conclude that changes in the relative amounts (or activities) of C/EBP␣ and other C/EBP isoforms influence the proliferative activity of hepatocytes during liver regeneration. However, because of the complexity of the C/EBP family of transcription factors, the multiple extracellular signals that regulate C/EBP activities, and the fact that hepatocytes co-express several different C/EBP isoforms, the role of each C/EBP family member in the regulation of hepatocyte proliferation during liver regeneration remains uncertain. Now that transgenic mice have been developed with targeted disruption of different C/EBP genes, it will be possible to use these mice to delineate the relative importance of different C/EBP isoforms as regulators of the hepatocyte proliferative response in vivo. Indeed, Greenbaum and colleagues have reported preliminary evidence that C/EBP␤ knock-out mice exhibit impaired hepatocyte proliferation and decreased liver regeneration after PH. 2 There are several mechanisms by which the C/EBPs may regulate hepatocyte proliferation. C/EBP␣ is known to stabilize p21/WAF, a protein that inhibits transition from the pre-replicative (G 1 ) period into S phase (5). Because induction of C/EBP␤ and C/EBP␦ during the early pre-replicative period may inhibit subsequent C/EBP␣ activity by repressing C/EBP␣ gene expression, reciprocal variations in C/EBP isoforms could release hepatocytes from C/EBP␣-mediated cell cycle arrest. There is also some evidence that C/EBP␤ may positively regulate the expression of genes that promote cell cycle progression. For example, C/EBP consensus sequences have been identified in the regulatory regions of cyclin D2 (27) and cyclin A (28) and may also interact with the retinoblastoma protein (RB) (29).
In addition to promoting hepatocyte proliferation, changes in the relative expression of different C/EBP isoforms may affect global changes in hepatocyte gene expression that permit these cells to survive the "stress" of regeneration. For example, C/EBP␤ and -␦ are important in modulating the expression of various acute phase response genes, including those such as inducible nitric oxide synthase (iNOS) (30 -32), that are thought to protect hepatocytes from cytokine-mediated toxicity (33)(34)(35)(36)(37). The activity of iNOS and production of nitric oxide (NO) increase in hepatocytes in the mid-late pre-replicative period following PH (38 -40). NO has diverse biological actions and has recently been shown to inhibit caspase 3 activation by TNF␣ (33,41). Given the pivotal role that caspase 3 plays in apoptosis (42)(43)(44), C/EBP␤-mediated induction of NO production may be one of the mechanisms that prevents injury-related cytokines from causing hepatocyte apoptosis in the regenerating liver. Preliminary experiments in iNOS-deficient transgenic mice support this view. 3 Although liver regeneration imposes a tremendous drain on hepatocyte ATP stores, hepatocyte necrosis is not normally increased by PH. Changes in the expression of phosphoenolpyruvate carboxykinase, glycogen synthase, and many other genes that regulate hepatic metabolism occur quickly after PH (45)(46)(47). C/EBP␤ and -␦ are important transcriptional regulators of many of the genes that are involved in metabolism (48 -50). As such, these transcription factors modify hepatocyte substrate utilization to assure energy homeostasis during the period of intense proliferative activity. Because failure to meet the increased ATP requirements during regenerative growth could lead to hepatocyte ATP depletion and necrosis, successful regulation of C/EBP activity may be important for thwarting hepatocyte necrosis during the regenerative response to injury.

C/EBPs and Liver Nonparenchymal Cells
Successful regeneration of the liver requires much more than the proliferation of hepatocytes. Cells that form supporting structures, such as bile ducts and blood vessels, must also replicate. In addition, new matrix must be deposited to provide a scaffolding for the "rebuilding" effort. Furthermore, evidence that hepatocyte proliferation is driven largely by factors that are produced within the regenerating liver implies that de novo synthesis of autocrine and paracrine growth regulatory factors must follow liver injury (2). The importance of the C/EBPs in regulating these aspects of the hepatic regenerative response has barely been evaluated.

Matrix Production
Workers in the field of hepatic fibrogenesis have long recognized that inflammation initiates fibrosis. The molecular basis for this observation is being discovered. Injury-related cytokines, such as TNF␣ and IL-6, and oxidant stress are known to induce several transcription factors, including C/EBP␤, that activate transcription of the type I collagen gene in stellate cells (2). C/EBP sites have also been identified in the promoters of metalloproteinases (51). Thus, the C/EBPs are likely to regulate matrix production and remodeling by hepatic nonparenchymal cells during liver regeneration.

Production of Paracrine and Autocrine Growth Regulators
Hepatocyte Growth Factor (HGF)-Stellate cells are also the principal source of the hepatocyte mitogen, HGF, in the regenerating liver (2). The regulatory regions of the HGF gene contain C/EBP consensus motifs (52), and recently, C/EBP␤ was shown to trans-activate HGF gene expression. Furthermore, recombinant TNF␣ induces a dose-dependent increase in IL-6 expression in primary stellate cell cultures. This is followed by increased C/EBP␤ DNA binding activity and induction of HGF mRNA expression. 4 These observations suggest that the C/EBPs play a role in regulating the production of paracrine growth factors in the regenerating liver. Definitive proof of this concept awaits analysis of HGF gene induction in the liver remnants of C/EBP␤ knock-out mice that have been subjected to PH.
TNF␣ and TNF-regulated Cytokines-Injury-related cytokines, particularly TNF␣ and IL-6, are now widely acknowledged to play pivotal parts in the initiation of hepatocyte proliferation after PH (11,13,54,55), although the relative importance and precise roles of these factors during liver regeneration remain a topic of considerable debate. Binding sites for C/EBP␤ have been identified in the promoters of a large number of cytokine genes, including TNF␣, IL-1, -6, and -8, and granulocyte colony-stimulating factor (56 -59). Transfection studies with TNF␣ reporter gene constructs have demonstrated that C/EBP␤ is much more active than C/EBP␣ as an inducer of TNF gene transcription. Furthermore, C/EBP␤ and -␦ are the predominate C/EBP isoforms in activated macrophages, which produce large amounts of TNF and TNF-inducible cytokines. Consistent with the concept that C/EBP␤ is important for TNF induction, inflammation-related increases in serum TNF␣ concentrations are blunted in C/EBP␤-deficient transgenic mice. This observation suggests that decreased production of TNF␣ may contribute to the impaired liver regeneration that Greenbaum's group has noted in C/EBP␤ knock-out mice. Decreases in TNF␣, in turn, would be expected to influence the relative abundance of C/EBP family members in the nuclei of hepatocytes and other TNF target cells because TNF or peptide ligands that specifically activate type 1 or type 2 TNF receptors induce an almost instantaneous redistribution of preformed C/EBP␤ and C/EBP␦ from the cytosol into the nucleus (18). Although C/EBP␤ was initially proposed as a potential mediator of the TNF-dependent induction of IL-6, which occurs after PH (11,54), studies using mice homozygous for a deletion in the gene for C/EBP␤ demonstrate that C/EBP␤ is not essential for IL-6 induction because these mice actually develop increased circulating levels of IL-6 as they age (60). Given this, it is unlikely that decreased induction of IL-6 contributes to the impaired liver regeneration that Greenbaum's group has noted in the C/EBP␤ knock-out mice. Rather, these observations demonstrate that organisms have developed many mechanisms to preserve cytokine induction when any given cytokine-induced transcription factor becomes deficient. The latter may help to explain why IL-6 induction is not entirely abolished in transgenic TNF receptor type 1-null mice (11) or in rats that were pretreated with anti-TNF antibodies before PH (54). Similarly, because IL-6 induces C/EBP␤ and C/EBP␤ regulates the transcription of multiple cytokines, including TNF␣, it is difficult to assure that IL-6 is the only cytokine that is deficient in transgenic mice in which the IL-6 gene has been disrupted experimentally.

Regulation of C/EBP Activity during Liver Regeneration
C/EBP function is regulated at many levels, consistent with the growing body of evidence which suggests that these factors regulate critical events that are involved in cellular proliferation, differentiation, and survival. Several hormones (e.g. insulin, glucagon, epinephrine, norepinephrine, and corticosteroids) that increase in the blood after PH (2) are known to regulate either the expression or activity of various C/EBP isoforms. For example, insulin has been shown to alter the turnover of the 40 -42 kDa C/EBP␣ proteins, accelerating their degradation (61). Cyclic AMP is known to increase transcription of C/EBP␤ and -␦ (61-63). Hence, glucagon and epinephrine, which increase adenylyl cyclase activity and promote the accumulation of hepatic cAMP in the early pre-replicative period, are likely to play a role in the induction of these C/EBP isoforms after PH. The cAMP-dependent kinase, protein kinase A, has also been shown to influence the function of C/EBP␤ protein by phosphorylating Ser 299 (21). Norepinephrine and other ␣-adrenergic agents that promote increases in intracellular calcium and activate calmodulin-sensitive kinases in the regenerating liver may also regulate C/EBP␤ because there is evidence that Cam kinase II phosphorylation of C/EBP␤ alters its DNA binding activity (20). Regenerative increases in circulating corticosteroids may contribute to increase the expression of C/EBP␦ mRNAs that have been observed in the regenerating liver (63).
Several hepatocyte mitogens that regulate the hepatocyte proliferative response during regeneration (including epidermal growth factor, transforming growth factor ␣, and HGF) may also influence C/EBP function. These growth factors activate receptor tyrosine kinases that couple to mitogen-activated kinases (64), which are known to phosphorylate C/EBP␤ (65). Injury-related cytokines are also likely to play important roles in regulating C/EBP function in the regenerating liver. For example, IL-6 and IL-1␤ are well established inducers of C/EBP␤ and C/EBP␦ mRNA expression, respectively (53), whereas TNF␣ has been shown to increase the DNA binding activity of C/EBP␤ and C/EBP␦ by affecting the nuclear localization of these proteins (18). The latter appears to have physiological relevance for liver regeneration because pretreatment with anti-TNF antibodies inhibits the nuclear accumulation of these C/EBP isoforms and decreases their participation in DNA complex formation after PH (46).
Thus, many factors that regulate the proliferation and differentiated functions of hepatocytes during the regenerative response are also known to effect the biological activity of the C/EBPs. Given this and the fact that the C/EBPs regulate the proliferation and differentiation of many types of cells, these transcription factors are likely to be important downstream targets for growth factor-initiated signals that drive the regeneration of the liver after it has been injured (Fig. 1). Definitive proof of this concept as well as progress in delineating the relative importance of various C/EBP family members during liver regeneration is now possible by studying the regenerative response to PH in transgenic mice strains with targeted deletion(s) of specific C/EBP genes.