The liver responds to injury with regulated tissue regeneration. During early regeneration, the liver accumulates fat. Neither the mechanisms responsible for nor the functional significance of this transient steatosis have been determined. In this study, we examined patterns of gene expression associated with hepatic fat accumulation in regenerating liver and tested the hypothesis that disruption of hepatic fat accumulation would be associated with impaired hepatic regeneration. First, microarray-based gene expression analysis revealed that several genes typically induced during adipocyte differentiation were specifically upregulated in the regenerating liver prior to peak hepatocellular fat accumulation. These observations suggest that hepatic fat accumulation is specifically regulated during liver regeneration. Next, 2 methods were employed to disrupt hepatocellular fat accumulation in the regenerating liver. Because exogenous leptin supplementation reverses hepatic steatosis in leptin-deficient mice, the effects of leptin supplementation on liver regeneration in wild-type mice were examined. The data showed that leptin supplementation resulted in suppression of hepatocellular fat accumulation and impairment of hepatocellular proliferation during liver regeneration. Second, because glucocorticoids regulate cellular fat accumulation during adipocyte differentiation, the effects of hepatocyte-specific disruption of the glucocorticoid receptor were similarly evaluated. The results showed that hepatic fat accumulation and hepatocellular proliferation were also suppressed in mice with liver specific disruption of glucocorticoid receptor. In conclusion, suppression of hepatocellular fat accumulation is associated with impaired hepatocellular proliferation following partial hepatectomy, indicating that hepatocellular fat accumulation is specifically regulated during and may be essential for normal liver regeneration. (HEPATOLOGY 2004;40:1322–1332.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The liver regenerates in response to a variety of injuries.1–3 The rodent partial hepatectomy model has been a useful tool with which to investigate the signals that regulate this regenerative response. Following partial hepatectomy, most of the remaining quiescent hepatocytes in the remnant liver tissue quickly proliferate leading to rapid restoration of appropriate liver mass.4 Analyses of genetically and pharmacologically manipulated mice using this model have begun to identify the coordinated signaling events that regulate hepatic regeneration. These signals include activation of tumor necrosis factor α (TNFα) interleukin (IL)-6 signaling,5–7 generation of mitochondrial reactive oxygen species8 and prostaglandins9, and activation of stress- and mitogen-activated-protein kinase cascades.10 These signals lead to activation of nuclear factor κB (NFκB) and other transcription factors, which direct an immediate early gene expression program culminating in growth factor–dependent hepatocellular reentry into and progression through the cell cycle11–13. Once this has occurred and normal hepatic mass is restored, the regenerative response is terminated. Despite extensive analyses with this model, a clear understanding of the specific signaling events that are required for initiation, propagation, and termination of the hepatic regenerative response and how such signals are integrated remains elusive.
In parallel with the early signaling events described above, the regenerating liver transiently accumulates large amounts of triglyceride fat.14–16 The specific molecular mechanisms that regulate the development and resolution of this transient hepatic steatosis and its functional significance during the regenerative response remain poorly defined. To more fully characterize the mechanisms involved in regulating the early adipogenic response of the regenerating liver, patterns of gene expression in regenerating liver prior to and during peak hepatic fat accumulation were compared to those seen in nonregenerating liver. In addition, the functional significance of these changes was examined by determining the effect of inhibition of this adipogenic response on hepatic regeneration. The results of these analyses indicate that hepatocellular fat accumulation following partial hepatectomy is specifically regulated and may be essential for normal liver regeneration.
Eight- to 12-week-old male mice were used for all studies. C57Bl/6 mice (Jackson Laboratory, Bar Harbor, ME) were used for the microarray and leptin supplementation studies. Mice with hepatocyte-specific disruption of the glucocorticoid receptor (GR) (AlbCre/GRloxP–loxP) and Cre-negative littermate controls on a mixed C57Bl/6x129 background were used to evaluate the role of GR in the regenerative response. These mice were generated by mating GRloxP–loxP mice, which are homozygous for replacement of the wild-type GR allele with one in which exons 1c and 2 of the GR gene are flanked by loxP sequences,17 to GRloxP–loxP mice transgenic for Cre recombinase under the control of the albumin promoter.18 Mice were kept on 12-hour dark-light cycles and maintained on standard mouse chow and water before and after surgery. Partial hepatectomy or sham surgery was performed as previously described.4, 9 For analysis of hepatocellular proliferation, at the time of partial hepatectomy mice also underwent intraperitoneal implantation of a bromodeoxyuridine (BrdU)-containing sustained-release osmotic pump (Alza, Newark, DE) delivering a flow rate of 20 μg/h of BrdU. For leptin supplementation experiments, mice were treated daily with 1 mg/kg recombinant mouse leptin (Alpha Diagnostics, San Antonio, TX) delivered by subcutaneous injection beginning 3 days prior to partial hepatectomy and continuing until the time of animal sacrifice. This regimen has been shown to influence hepatic gene expression in leptin-deficient ob/ob mice.19 At serial times after surgery animals were killed and liver tissue was harvested into formalin fixative, Optimal Cutting Temperature Compound (O.C.T., Tissue Tek, Torrance, CA), or snap frozen into liquid nitrogen and stored at −80°C. Three to 6 animals were examined at each time point and for each treatment group or genotype. All experiments were approved by the Animal Studies Committee of Washington University and conducted in accordance with institutional guidelines and the criteria outlined in the “Guide for Care and Use of Laboratory Animals” (National Institutes of Health publication 86-23).
Total RNA was prepared from mouse liver tissue using Triazol reagent (Invitrogen, Carlsbad, CA) followed by purification using the RNeasy total RNA cleanup protocol (Quiagen, Valencia, CA). RNA integrity was assessed by formaldehyde/agarose gel electrophoresis and quantified by 260 nm absorbance.
Microarray-Based Gene Expression Analysis.
Ten-microgram aliquots of purified RNA, pooled from replicate animals subjected to partial hepatectomy or sham surgery and harvested at each of the time points under study, were submitted to the Washington University School of Medicine Siteman Cancer Center Core Facility for microarray gene expression analysis. Double-stranded complementary DNA (cDNA) was synthesized from these RNAs, using the SuperScript Choice system (Invitrogen, Carlsbad, CA), and biotinylated complementary RNA probes were generated from these cDNAs, using the Bioarray High Yield RNA transcript labeling kit (Enzo, Farmingdale, NY). The biotin-labeled complementary RNA was purified, fragmented, and hybridized to Mu74A arrays (Affymetrix, Santa Clara, CA) and also to a test chip for quality control assessment of cDNA synthesis, in vitro transcription, and chip handling. After chip hybridization, washing, and staining, performed according to the manufacturer's instructions, the microarrays were scanned by the Affymetrix GeneChip scanner. The resulting image data was captured and converted to numerical output using the Microarray Analysis Suite (Version 4.0) and imported into GeneSpring software (Silicon Genetics, Redwood City, CA) for further analysis. For comparison of gene expression between tissue samples, expression data from each chip was normalized to β2-microglobulin gene expression, a gene whose hepatic expression does not change during the regenerative response that is often used for normalization of gene expression data in analyses of liver regeneration.20 Gene expression data from regenerating liver harvested at each time point was compared to expression data from quiescent liver and from sham-operated liver harvested at the corresponding time point. The data set was filtered by removal of genes whose expression fell below a threshold level of fluorescence intensity (absolute fluorescence intensity units < 300) or was scored as “Not Detected” by the Affymetrix software in all of the samples analyzed. For the remaining genes, expression in regenerating liver was compared to that in quiescent and sham-operated liver, and genes that were induced or suppressed at least 3-fold were identified. These criteria were selected based on the authors' experiences indicating that they have a high probability of identifying differences in gene expression that can be independently validated by other methods of analysis. The resulting sets of induced and suppressed genes were classified into functional gene ontology (GO) groups using GeneSpring software.
Real-Time RT-PCR-Based Gene Expression Analysis.
For specific genes of interest, gene expression was independently reevaluated by real-time reverse-transcriptase polymerase chain reaction (RT-PCR) using both the pooled RNAs analyzed in the microarray analysis and the individual replicates used to generate the pooled samples. For these studies, mouse liver messenger RNA (mRNA) prepared as described above was reverse-transcribed to cDNA by the SuperScript Choice System (Invitrogen). For each gene analyzed, an aliquot of cDNA was added to a reaction mixture containing gene-specific forward and reverse primers (Table 1), deoxy-nucleotides, Taq DNA polymerase, and SYBR Green (Bio-Rad, Hercules, CA). Quantification of cDNA was based on monitoring increased SYBR Green fluorescence during exponential phase amplification in a Real-Time PCR Machine (Bio-Rad), and determination of the PCR cycle number at which the amplified product exceeded a defined threshold (the crossing threshold). These data were also standardized to the expression of β2-microglobulin. The standardized data were used to calculate fold differences in gene expression. Specificity of this assay was verified for each gene under analysis by confirmation of predicted product size and uniformity using melt-curves and agarose-electrophoresis of the PCR products. Specificity was further confirmed by simultaneous analysis of a “reverse-transcribed” reaction mixture containing all components except reverse transcriptase.
Table 1. Primers for Real-Time RT-PCR Analysis of Hepatic Gene Expression
Histology and Immunohistochemistry.
Liver sections were fixed in formalin, paraffin embedded, and stained with hematoxylin and eosin, for nuclear BrdU incorporation (using the BrdU Immunohistochemistry kit, Oncogene, Boston, MA), or for immunohistochemical detection of GR expression. The frequency of nuclear BrdU labeling was determined by examination of at least three random 400× fields and at least 300 cells and nuclei in each tissue section. For these experiments, nuclei were counterstained with hematoxylin. GR immunoreactivity was detected in liver sections with a GR-specific antibody (M20, Santa Cruz Biotechnology, Santa Cruz, CA) and tyramide amplification using the Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) and a Tyramide Signal Amplification Kit (Perkin-Elmer, Boston, MA). For these experiments, nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Sections of liver frozen in Optimal Cutting Temperature Compound were stained with Oil Red O, (Sigma Chemical Co., St. Louis, MO) for analysis of hepatic fat accumulation.
Hepatic Triglyceride Content.
Aliquots of frozen liver were assayed by a modification of the Folch method21 for triglyceride content using a colorimetric triglyceride determination kit according to manufacturer's instructions (Wako, Richmond, VA).
Data were analyzed using Sigma Plot software (SPSS, Chicago, IL). ANOVA for multiple groups was used to determine statistical significance of real-time RT-PCR–based differences in gene expression, triacylglycerol content, hepatocellular BrdU incorporation, and mitotic body frequency. Data are reported as mean ± SE.
Transient Hepatic Steatosis Occurs During Early Liver Regeneration.
The mechanisms that regulate the development and resolution of hepatic steatosis following partial hepatectomy and the functional significance of these events to the regenerative response are unknown. To begin to address this, the timing and magnitude of hepatic steatosis during liver regeneration following partial hepatectomy were examined. This analysis showed that mice exhibit markedly increased hepatocellular fat accumulation 12 to 24 hours after partial hepatectomy (Fig. 1). This accumulation of fat is not detectable after sham surgery. These data demonstrate the specific and transient development of hepatic steatosis during liver regeneration in the partial hepatectomy model.
Induction of an Adipogenic Gene Expression Program during Early Liver Regeneration.
To characterize the molecular signals that regulate histological adipogenic changes during early liver regeneration, cDNA microarray-based gene expression analysis comparing patterns of gene expression in regenerating liver prior to and during peak hepatic fat accumulation (2-12 hours after partial hepatectomy) to those seen in quiescent or corresponding sham-operated liver was performed. A total of 233 genes, including 110 known genes and 123 expressed sequence tags (ESTs) were identified as induced at least 3-fold (Table 2), and another 236 genes, including 71 known genes and 165 ESTs, were identified as suppressed at least 3-fold (Table 3) during the hepatic regenerative response over this period of time. Included among the genes induced during the regenerative response were several that have previously been classified as “adipogenic” (Table 4), i.e., genes whose expression is specifically induced in cellular models of adipogenesis.22–27 These include adipsin (complement factor D), aP2 (adipocyte fatty acid binding protein), S3-12, and FSP-27. These microarray-based gene expression analysis results were independently verified using real-time RT-PCR analysis, which confirmed the increased expression of each of these adipogenic genes in regenerating liver (Fig. 2). These observations suggest that hepatic fat accumulation is specifically regulated during liver regeneration.
Table 2. Genes Induced Greater Than 3-Fold in Regenerating Liver
Fold vs 0
Fold vs Sham
Fold vs 0
Fold vs Sham
Fold vs 0
Fold vs Sham
NOTE. Induced genes were identified by cDNA-microarray based gene expression analysis. Genes were grouped into gene ontology (GO) groups using GeneSpring software. The table contains the GenBank number, gene name, time (in hours after partial hepatectomy) of peak induction, and fold induction in regenerating liver compared to quiescent (vs. 0) and sham-operated (vs. Sham) liver.
Table 3. Genes Suppressed Greater Than 3-Fold in Regenerating Liver
Fold vs 0
Fold vs Sham
Fold vs 0
Fold vs Sham
Fold vs 0
Fold vs Sham
NOTE. Suppressed genes were identified by cDNA-microarray based gene expression analysis. Genes were grouped into gene ontology (GO) groups using GeneSpring software. The table contains the GenBank number, gene name, time (in hours after partial hepatectomy) of peak suppression, and fold suppression in regenerating liver compared to quiescent (vs. 0) and sham-operated (vs. Sham) liver.
Leptin-Supplementation Suppresses Hepatic Adipogenesis and Impairs Hepatocellular Proliferation during Liver Regeneration.
To evaluate the functional significance of hepatic adipogenic changes during the regenerative response, 2 independent strategies were employed in an attempt to suppress adipogenesis and then determine the effect of such suppression on hepatocellular proliferation during liver regeneration. First, a pharmacological strategy was examined, based on the observation that a sustained pattern of hepatic adipogenesis homologous to the one described above is present in leptin deficient ob/ob mouse liver28 and can be reversed by exogenous leptin administration. Recombinant leptin was administered to wild-type mice in an attempt to block hepatic fat accumulation during liver regeneration. The results showed that leptin supplementation did suppress hepatocellular fat accumulation during liver regeneration, based on Oil Red O staining and biochemical determination of hepatocellular triglyceride content in regenerating liver harvested 12 hours after partial hepatectomy (Fig. 3A). Next, the effect of leptin supplementation on the hepatic regenerative response was characterized by quantification of hepatocellular BrdU labeling 48 hours after partial hepatectomy, in leptin- and vehicle-treated mice in which a BrdU-containing sustained-release osmotic pump was implanted at the time of the surgery. The results of this analysis, which reflect an integrated determination of the regenerative response over the 48-hour time course of the experiment, showed that leptin treatment impairs hepatocellular BrdU incorporation following partial hepatectomy (Fig. 3B) (4.3% ± 2.0% in leptin-treated–, 70.3% ± 5.9% in vehicle-treated mice, P < .001) and mitotic body frequency (Fig. 3C) (0.0 ± 0.0 and 8.0 ± 1.1 respectively, P < .001). Despite these effects, there was no evidence of increased hepatic tissue necrosis (Fig. 3C) or increased mortality in the leptin-treated mice. These data show that leptin supplementation can block both hepatic adipogenesis and hepatocellular proliferation during liver regeneration, raising the possibility that the early adipogenic response is essential for appropriate initiation of hepatic regeneration.
Hepatocyte-Specific Disruption of the Glucocorticoid Receptor Gene Suppresses Hepatic Adipogenesis and Impairs Hepatocellular Proliferation during Liver Regeneration.
To further investigate the functional importance of the hepatic adipogenic response for normal liver regeneration, an independent genetic strategy was employed in an attempt to suppress hepatic adipogenesis during liver regeneration. Glucocorticoid signaling has been implicated as an important regulator of adipocyte-specific gene expression and cellular fat accumulation in models of adipocyte differentiation.29, 30 These observations suggest that hepatocellular glucocorticoid signaling could regulate hepatic adipogenic changes during liver regeneration. If this is true and if such changes are important for normal regeneration, as suggested by the results described above, then disruption of hepatocellular glucocorticoid signaling during liver regeneration would be associated with impaired hepatic adipogenesis and suppressed hepatocellular proliferation. Glucocorticoid signaling is mediated by the type II GR, which is expressed in most cell types.31 Ubiquitous inactivation of the GR gene, as a way of disrupting glucocorticoid signaling, results in perinatal lethality in mice.32, 33 Therefore, the effect of hepatocyte-specific disruption of GR on hepatocellular adipogenic changes and proliferation during liver regeneration was assessed. Hepatocyte-specific disruption of GR was accomplished using the Cre/loxP system, by generating mice transgenic for albumin promoter-directed hepatocellular Cre expression (AlbCre)18 and homozygous for a loxP-flanked GR allele (GRloxP–loxP).17 Non-Cre expressing GRloxP–loxP wild-type (WT) mice show typical nuclear expression of GR while AlbCre/GRloxP–loxP mice (liver GR knockout [KO]) exhibit loss of GR immunoreactivity in hepatocytes in adulthood (Fig. 4A). When these animals were subjected to partial hepatectomy, the results showed decreased hepatocellular fat accumulation 12 hours after partial hepatectomy in liver GR knockout mice compared to WT littermate controls (Fig. 4B). Liver-specific disruption of the GR gene was also associated with significantly reduced hepatocellular BrdU labeling (32.4% ± 11.3% in liver GR knockout vs. 73.3% ± 4.2 in WT, P < .01) (Fig. 4C) and mitotic body frequency (2.0 ± 0.8 versus 11.0 ± 2.5 respectively, P < .02) (Fig. 4D) following partial hepatectomy. As was true in the leptin supplementation experiment, there was no detectable increase in mortality in the liver GR KO animals compared to the WT animals in the 48 hours following hepatectomy. These data show that liver-specific disruption of the GR gene is associated with impaired hepatic adipogenesis and decreased hepatocellular proliferation during liver regeneration, which, together with the data described above, indicate that the early hepatic adipogenic response to partial hepatectomy may be necessary for normal hepatic regeneration.
The studies reported here show that prior to transiently accumulating large amounts of fat, the early regenerating liver activates a program of gene expression that is partially homologous to one that is induced during adipocyte differentiation. This suggests that transient hepatic adipogenesis is a specifically regulated component of the regenerative response. Although the molecular mechanisms responsible for such regulation remain unknown, a number of candidates are worthy of consideration. For example, expression of CCAAT enhancer binding protein (C/EBP) transcription factors, including C/EBPβ and C/EBPδ, are increased during and essential for normal liver regeneration34 and are also important regulators of adipocyte differentiation.35 Thus, modulation of C/EBP activity may determine adipogenic changes during early liver regeneration. Alternatively, the recent report demonstrating that hepatic peroxisome proliferator-activated receptor (PPAR)γ overexpression induces hepatocellular adipogenic gene expression and steatosis22 suggests that this nuclear steroid hormone receptor could also direct adipogenic changes in the regenerating liver. PPARγ mRNA expression does not change during liver regeneration (data not shown). However, an undefined endogenous PPARγ activating ligand appears to be increased in a C/EBPβ- and C/EBPδ-dependent way,36 suggesting that C/EBP-dependent regulation of PPARγ activity could direct hepatic adipogenic changes during early liver regeneration.
The data reported here also show that 2 different strategies that suppress hepatic adipogenesis, namely leptin supplementation and hepatocellular GR disruption, are each associated with impaired hepatocellular proliferation during liver regeneration. Although each of these interventions may have other effects, these observations raise the possibility that intrahepatic accumulation of fat may be important for appropriate initiation of hepatocellular proliferation during normal liver regeneration. This hypothesis is further supported by several published observations. For example, proliferation of primary hepatocytes in cell culture is associated with marked intracellular accumulation of fat.37 In addition, fatty acid synthase, the enzyme responsible for de novo synthesis of the most abundant long-chain fatty acid palmitic acid, is frequently up-regulated in human malignancies, including prostate, breast, bladder, lung, liver, melanoma, oral squamous cell, and colorectal cancers.38–44 In fact, fatty acid synthase overexpression is associated with more aggressive neoplastic disease, suggesting that increased fatty acid synthesis or accumulation may provide a selective growth advantage. Indeed, inhibition of fatty acid synthase expression or activity has been shown to impair tumor cell proliferation and survival in a number of cell culture models.38, 40, 42;44;45 The data presented here suggest that fatty acid synthesis and accumulation may also be important for proliferation of nonmalignant cells (e.g., regenerating hepatocytes).
The complexity of the effects of leptin on the hepatic regenerative response is highlighted by contrasting the observations reported herein that leptin supplementation results in impaired liver regeneration in wild-type mice, with recently reported data showing that leptin supplementation of leptin-deficient ob/ob mice rescues the impaired regenerative response seen in those mutant animals.46 Although the specific molecular basis for this observation remains unknown, that the effects of leptin supplementation on ob/ob mice and wild-type mice during liver regeneration are discordant is analogous to the influence of IL-6 supplementation on liver regeneration in wild-type mice (in which liver regeneration is inhibited by IL-6)6 and IL-6 null mice (in which impaired liver regeneration is rescued by IL-6).7 Similar complexity is evident in the role of glucocorticoid signaling during liver regeneration when the data reported here, which show that hepatocellular GR disruption results in impaired regeneration, is compared with published data showing that augmenting GR signaling by systemic glucocorticoid administration suppresses hepatic regeneration.47 In this case, the mechanistic basis for these observations is likely to be explained by cell-specific effects of GR signaling during liver regeneration. For example, in the studies reported here, GR signaling was specifically disrupted only in hepatocytes, whereas in the report on the effect of systemic steroid administration on liver regeneration,47 the activation of Kupffer cell cytokine synthesis, including those important for the initiation of hepatic regeneration, was likely suppressed.48
The specific molecular mechanisms that mediate the effects of leptin supplementation and hepatocellular GR deficiency on hepatocellular fat accumulation and proliferation during liver regeneration are not yet known and will be the subject of future investigation. In contrast to the direct effects of hepatocyte-specific GR disruption on liver regeneration, the effects of leptin are likely to be indirect, based on the recently published observation that leptin supplementation does not influence hepatocyte proliferation in primary culture.49 Nevertheless, understanding each of these mechanisms in greater detail should provide additional insight into the molecular mechanisms that regulate hepatic regeneration and may lead to the development of novel strategies with which to modulate hepatic adipogenesis. This could have clinical therapeutic utility with respect to augmenting hepatic regeneration in the setting of liver diseases or, perhaps, suppressing abnormal hepatocellular proliferation in the setting of liver cancers. Moreover, because similar mechanisms may be involved in the development of transient steatosis during liver regeneration and sustained steatosis in fatty liver disease, elucidation of these signaling events could ultimately lead to improved understanding of the development of and management strategies for fatty liver disease, an increasingly prevalent form of liver injury.
We thank Dr. Brian Dieckgraefe for expertise and assistance with the gene expression analyses, Dr. Jonathan Gitlin for critical review of this manuscript, and the Washington University School of Medicine DDRCC and Siteman Cancer Center for technical support.