Potential conflict of interest: Nothing to report.
See Editorial on Page 1875
We previously reported that mice subjected to partial hepatectomy exhibit rapid development of hypoglycemia followed by transient accumulation of fat in the early regenerating liver. We also showed that disrupting these metabolic alterations results in impaired liver regeneration. The studies reported here were undertaken to further characterize and investigate the functional importance of changes in systemic adipose metabolism during normal liver regeneration. The results showed that a systemic catabolic response is induced in each of two distinct, commonly used experimental models of liver regeneration (partial hepatectomy and carbon tetrachloride treatment), and that this response occurs in proportion to the degree of induced hepatic insufficiency. Together, these observations suggest that catabolism of systemic adipose stores may be essential for normal liver regeneration. To test this possibility, we investigated the hepatic regenerative response in fatty liver dystrophy (fld) mice, which exhibit partial lipodystrophy and have diminished peripheral adipose stores. The results showed that the development of hypoglycemia and hepatic accumulation of fat was attenuated and liver regeneration was impaired following partial hepatectomy in these animals. The fld mice also exhibited increased hepatic p21 expression and diminished plasma levels of the adipose-derived hormones adiponectin and leptin, which have each been implicated as regulators of liver regeneration. Conclusion: These data suggest that the hypoglycemia that develops after partial hepatectomy induces systemic lipolysis followed by accumulation of fat derived from peripheral stores in the early regenerating liver, and that these events may be essential for initiation of normal liver regeneration. (HEPATOLOGY 2010)
The liver has remarkable regenerative potential, which permits recovery from functional deficits induced following hepatic injury. The rodent partial hepatectomy model has been the most extensively used experimental system for investigating the mechanisms that control this highly regulated response.1 Analyses using this paradigm have identified many signals that are regulated during and necessary for normal liver regeneration.2–6 Nevertheless, an integrated understanding of the mechanisms that regulate liver regeneration does not yet exist, and the signals that initiate and terminate hepatic regeneration remain incompletely defined.
Liver mass is maintained or recovered after injury in proportion to body mass.3, 6 This observation, taken together with the central role of the liver as the principal intermediary between dietary nutrient uptake and extrahepatic energy consumption,7 has led us to investigate the regulation and functional role of systemic metabolic changes in response to partial hepatectomy during the hepatic regenerative response. We previously reported that genetic and pharmacological interventions that suppress the transient hepatic steatosis characteristic of the early regenerative response result in impaired liver regeneration.8 We also characterized the hypoglycemic response to partial hepatic resection and the inhibitory effect of glucose supplementation on liver regeneration.9 Together, these data suggest a model in which the hypoglycemia that follows partial hepatectomy induces systemic lipolysis and accumulation of fat derived from peripheral stores in the early regenerating liver. The studies reported here were undertaken to further characterize the significance of changes in peripheral adipose metabolism during liver regeneration.
BrdU, bromodeoxyuridine; CCl4, carbon tetrachloride; C/EBP, CCAAT/enhancer binding protein; fld, fatty liver dystrophy; MR, magnetic resonance; SOCS3, suppressor of cytokine signaling 3; STAT3, signal transducer and activator of transcription 3.
Materials and Methods
Animal Husbandry, Partial Hepatectomy, and Carbon Tetrachloride Treatment.
Wild-type C57Bl6/J mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Lpin1 null (fatty liver dystrophy (fld)/fld, BALB/cByJ-Lpin1fld/J mice; Jackson Laboratory), heterozygous (fld/+), and wild-type (+/+) mice were obtained by maintaining fld/fld × fld/+ and fld/+ × fld/+ mating pairs. Eight to 12 week-old C57Bl6/J and Lpin1 null, heterozygous, and wild-type mice were maintained on 12-hour dark-light cycles with adlibitum access to standard rodent chow and water throughout the analysis, and subjected to two-thirds partial hepatectomy or sham surgery using standard methodology (see Supporting Materials and Methods).8, 10–12 Some animals were treated with recombinant leptin using a regimen shown to rescue impaired regeneration in ob/ob mice (see Supporting Materials and Methods)13; some were subjected to one-third partial hepatectomy, in which only the median lobes of the liver were resected; and some were treated with carbon tetrachloride (CCl4) (see Supporting Materials and Methods). At serial times after surgery or CCl4 administration, animals were sacrificed and plasma and liver tissue were harvested. Very little morbidity or mortality occurred in experimental animals (summarized in Supporting Materials and Methods). Three or more animals were examined at each time point for each genotype, surgical, and treatment group. 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 (NIH publication 86-23).
Histology, Immunohistochemistry, Messenger RNA (mRNA), and Protein Expression, Body Composition, Hepatic Triglyceride, and Plasma Leptin and Insulin Analyses.
See Supporting Materials and Methods for detailed methods.
Data were analyzed using SigmaPlot and SigmaStat software (SPSS, Chicago, IL). Unpaired Student t test for pairwise comparisons and analysis of variance for multiple groups were used with significance (alpha) set at 0.05. Data are reported as mean ± standard error.
Partial Hepatectomy Induces Systemic Catabolism Prior to Onset of Hepatocellular Proliferation.
To begin to investigate the systemic metabolic response to partial hepatectomy, total, lean, and fat mass were measured at serial times after surgery in wild-type C57Bl/6J mice. The results showed a stereotypical pattern of loss and recovery in each of these parameters after hepatic resection but not sham surgery (Fig. 1A-C). Maximum loss of body weight occurred 24 hours after surgery, with subsequent recovery and return to baseline by ∼2 weeks (Fig. 1A). The amount of weight lost, ∼10% of the initial body mass, was greater than that which could be explained by removal of two-thirds of the liver (∼3% of the initial body weight). Next, changes in lean and fat mass during liver regeneration were determined using magnetic resonance (MR) spectroscopy. The results showed that both lean and fat tissue stores declined and reached their respective nadirs 24 hours after partial hepatectomy, with significantly smaller changes seen after sham surgery (Fig. 1B,C). At 24 hours, lean mass had declined by ∼10% and fat mass by ∼20% of the initial values. These catabolic changes followed the onset of hypoglycemia, detectable 3 hours after partial hepatectomy,9 and preceded the initiation of hepatocellular proliferation, which remains almost undetectable at 24 hours and does not peak until 36 hours after surgery (Fig. 4).9, 10, 12, 14 Recovery of tissue mass followed specific and distinct patterns (Fig. 1B,C), with lean mass increasing more rapidly than fat stores. Individual muscles showed changes in mass comparable to those seen by MR spectroscopy, with quadriceps and gastrocnemius mass decreasing by 5% and 10%, respectively, 24 hours after partial hepatectomy. Mice subjected to partial hepatectomy ate approximately two-thirds as much food over the first 24 hours after surgery as did controls (Supporting Fig. 1A). Thus, the decline in body mass observed after partial hepatectomy (Fig. 1) exceeds that predicted based on decreased caloric intake alone (6% to 7% after 24-hour fast; D.A.R., unpublished observations). These results demonstrate the induction of a reproducible systemic catabolic response after partial hepatectomy in mice.
CCl4 Induces Systemic Catabolism.
To investigate whether the systemic metabolic response described above is specific for partial hepatectomy or a more common response to hepatic insufficiency, changes in body mass were determined in another model of hepatic regeneration, that induced by administration of CCl4.15, 16 As seen after partial hepatectomy, CCl4 treatment induced specific catabolic changes in total, lean, and fat mass prior to onset of hepatocellular proliferation, with earlier recovery of lean versus fat tissue stores (Fig. 2A-F). Mice treated with CCl4 took in approximately one-fourth as much food over the initial 24 hours (Supporting Fig. 1B). Together, these data show that systemic catabolism prior to the onset of hepatocellular proliferation occurs in two different models of liver regeneration.
The Systemic Catabolic Response to Partial Hepatectomy Is Proportional to the Amount of Liver Resected.
The data described above raise the possibility that the systemic catabolic response to a hepatic regenerative stimulus (e.g., partial hepatectomy or CCl4 exposure) might contribute to regulation of liver regeneration. If so, then the extent of this response—like regeneration itself—should occur in proportion to the magnitude of the regenerative stimulus. One-third partial hepatectomy induces significantly less hepatocellular proliferation compared to removal of two-thirds of the liver.17, 18 Therefore, the systemic response to two-thirds partial hepatic resection was compared to that seen after one-third hepatectomy. The development of hypoglycemia and accumulation of hepatic triglycerides, which we previously reported as regulated during and important for normal liver regeneration,8, 9 was examined first. The results showed that the degree of hypoglycemia was significantly less severe (Fig. 3B) and the magnitude of hepatic triglyceride accumulation much lower (Fig. 3C) after one-third versus two-thirds hepatic resection. Further analysis showed that removal of one-third of the liver was also associated with significantly less decline in total and fat mass (Fig. 3A). The decline in lean mass after one-third hepatectomy was not significantly different than that seen after two-thirds hepatectomy (P = 0.3). These data show that catabolism of total body and fat mass after partial hepatectomy occurs in proportion to the degree of induced hepatic insufficiency.
Liver Regeneration Is Impaired in Lipodystrophic fld Mice.
Many interventions that suppress hepatic fat accumulation after partial hepatectomy also result in impairment of liver regeneration8, 19, 20; however, neither de novo hepatic lipogenesis nor dietary fat–dependent hepatic lipogenesis appears to be required for such regeneration.21 Together, these observations suggest that catabolism of existing adipose stores may be essential for normal hepatic regeneration. To address this possibility, regeneration was examined in fld mice.22 The fld mice are homozygous for a mutation in Lpin1, which results in markedly diminished adipose tissue depot size throughout the body.22 Strain-matched wild-type and heterozygous mice appear identical to each other and exhibit comparable amounts of total body fat (18% ± 2% and 18% ± 1%, respectively, versus 13% ± 1% in fld mice; assessed by MR spectroscopy). Wild-type and heterozygous mice also demonstrate equivalent hepatocellular proliferation 36 hours after two-thirds partial hepatectomy, the time of peak proliferation in this model (Fig. 4C). Therefore, heterozygous mice were used as controls for analyses of liver regeneration in fld animals. These experiments showed that the regenerative response to partial hepatectomy was significantly impaired in fld mice, with reduced hepatocellular BrdU incorporation (Fig. 4A-C) and cyclin D1 mRNA and protein expression (Fig. 5A-C) compared to controls. The fld mice also exhibited diminished hepatocellular mitotic frequency (Fig. 2D-F; *P = 0.11 at 48 hours; P = 0.06 at 72 hours) and delayed recovery of liver mass (57% ± 5% versus 62% ± 1% in controls at 72 hours after partial hepatectomy, P = 0.2); however, these differences were not significant. Postoperative mortality was modestly increased in fld mice with 3 of 42 animals dying within 24 hours after partial hepatectomy compared to 0 of 44 heterozygous controls (P = 0.08). There was no increase in hepatic tissue necrosis in surviving null mice compared to controls (Fig. 4D,E). The hepatic regenerative response to CCl4 administration was also investigated, with administration of CCl4 at a dose sufficient to induce robust regeneration in wild-type mice (Fig. 2) resulting in lethality in four of four fld mice versus one of four controls. Together, these data show that liver regeneration is impaired in lipodystrophic fld mice.
fld Mice Contain Less Hepatic Triglyceride After Partial Hepatectomy Compared to Controls.
Next, changes in systemic metabolism after partial hepatectomy were examined in fld mice. The results showed that regenerating liver from fld mice contained significantly less triglyceride than controls (Fig. 6A). Triglyceride content was also reduced in quiescent fld liver, which likely reflects both the systemic adipose deficiency of fld mice and the increased hepatic triglyceride content at baseline in the BALBc genetic background.23fld mice exhibited less severe hypoglycemia 12-24 hours after partial hepatectomy (Fig. 6B) and higher plasma insulin levels 48-72 hours after surgery (Fig. 6C) compared to controls. These findings are consistent with prior characterization of fld mice as insulin resistant,24 and show that systemic metabolic changes characteristic of liver regeneration in wild-type animals are deranged in fld mice.
fld Mice Exhibit Dysregulated Hepatic p21 Expression After Partial Hepatectomy.
The data described above, together with our previous characterization of the inhibitory effect of supplemental glucose on liver regeneration,9 suggest that perturbations in systemic glucose metabolism may contribute to suppressed regeneration in fld mice. The impaired regenerative response associated with dextrose supplementation was characterized by augmented expression of CCAAT/enhancer binding protein alpha (C/EBPα), p21, and p27.9 Therefore, hepatic expression of these factors was compared between fld and control mice. The results showed that C/EBPα and C/EBPβ mRNA and p27 protein expression were comparable in fld and controls (Supporting Fig. 2 and Fig. 5D-F); however, p21 protein was increased in fld liver (Fig. 5D-F). These data raise the possibility that dysregulated p21 expression contributes to impaired regeneration in fld mice.
Plasma Adiponectin and Leptin Levels Are Decreased in fld Mice.
The adipose-derived hormones adiponectin and leptin have each been identified as regulators of liver regeneration.13, 25–29 To investigate whether deficiency of either hormone might contribute to impaired regeneration in fld mice, plasma levels of each were determined before and after partial hepatectomy in fld and control mice. This analysis showed that circulating adiponectin and leptin levels were significantly lower in fld animals at baseline (Fig. 7A-C). Following partial hepatectomy, leptin levels declined in controls and remained low in fld mice (Fig. 7C). Because leptin deficiency is associated with impaired liver regeneration,13, 28, 30 the effect of leptin supplementation on regeneration in fld mice was investigated. This analysis showed that a regimen of leptin supplementation sufficient to rescue impaired regeneration in CCl4-treated ob/ob mice13 did not augment and, in fact, suppressed hepatocellular proliferation 36 hours after partial hepatectomy in fld mice compared to untreated fld mice and leptin-treated controls (Supporting Fig. 3). Adiponectin levels increased after partial hepatectomy in controls (Fig. 7B), but remained almost undetectable in fld mice (Fig. 7A,B). These data suggest that impaired liver regeneration in fld mice may be mechanistically related to that recently described in adiponectin-null mice.26 Diminished activation of signal transducer and activator of transcription 3 (STAT3) and augmented induction of expression of suppressor of cytokine signaling 3 (SOCS3) were observed in liver after partial hepatectomy in those animals.27 Therefore, STAT3 activation and SOCS3 expression, each of which modulate liver regeneration,31, 32 were quantified after partial hepatectomy in fld and control animals. The results showed comparable STAT3 phosphorylation in both groups; however, the ratio of phosphorylated:total STAT3 was reduced. Moreover, in contrast to the analysis of adiponectin-null mice,27 hepatic SOCS3 expression after partial hepatectomy was significantly lower in fld mice than in controls (Fig. 8A-C) These data suggest that decreased circulating levels of adiponectin and disruption of STAT3 activation (which induces SOCS3 expression) may contribute to impaired liver regeneration in fld mice.
The studies reported here were undertaken to further characterize the regulation and functional significance of changes in systemic metabolism during normal liver regeneration. The results show that a systemic catabolic response is induced in each of two distinct models of liver regeneration. These experiments also show that catabolism of total and systemic fat mass—like regeneration itself—occurs in proportion to the degree of induced hepatic insufficiency. Surprisingly, catabolism of lean mass was not significantly different after one-third versus two-thirds partial hepatectomy. These data raise an intriguing question about liver:body mass regulation: To what body mass compartment is liver mass proportionately regulated? The answer to this question may offer insight into mechanisms of liver regeneration, and will be the subject of future investigations. Our data provide the first detailed characterization of systemic metabolic changes in these classic models of liver regeneration, and the findings reported here also offer insight into previously published analyses of liver regeneration. For example, tumor necrosis factor α and interleukin-6 have each been identified as essential regulators of normal liver regeneration after partial hepatectomy and CCl4 administration,15, 16, 33–36 and also induce cachexia.37 Thus, the catabolic response to liver injury in these models of liver regeneration may be induced by signals that are required for normal regeneration. Together, these considerations raise the possibility that the requirement of such signals for recovery of liver mass following hepatic injury may be mediated by the catabolic response they induce.
We and others have reported that interventions associated with disruption of transient hepatic accumulation of fat during early liver regeneration, including genetic disruption of caveolin-1 or hepatic glucocorticoid receptor expression, and leptin or propranolol supplementation, are associated with impaired regeneration.8, 19, 20 Subsequent analyses by Newberry et al.21 showed that regeneration is not impaired in mice in which dietary uptake of fat or de novo hepatic synthesis of fatty acid is disrupted (intestine-specific microsomal triglyceride transfer protein-null and liver-specific fatty acid synthase-null mice, respectively. That study also showed that regeneration is unaffected in liver fatty acid binding protein-null mice. In each case, the regenerating livers accumulated triglyceride fat but to a lesser extent than controls,21 leading those investigators to speculate on the existence of a critical “threshold of adaptive lipogenesis” that was not crossed in those animal models. Whether the genetic interventions evaluated in the Newberry study affected systemic adipose stores was not reported. Together with our previously published data, these considerations suggest that catabolism of existing peripheral adipose stores followed by hepatic accumulation of systemically derived fat may promote liver regeneration. To test this possibility, we investigated liver regeneration in fld mice, which have diminished peripheral adipose stores.22 The results showed that early hepatic fat content was reduced and liver regeneration impaired following partial hepatectomy in these animals. The increased insulin levels in fld mice 48-72 hours after partial hepatectomy is consistent with prior characterization of insulin resistance in these animals.24 Furthermore, the increased blood glucose levels 12-24 hours after surgery in fld mice, together with our previous characterization of the hypoglycemic response to partial hepatectomy and the inhibitory effect of glucose supplementation on early hepatic fat accumulation and liver regeneration in wild-type mice,9 suggest that perturbations in systemic glucose metabolism may contribute to impaired regeneration in fld mice. Indeed, hepatic p21 expression, which is increased by dextrose supplementation,9 was also augmented in regenerating fld mouse liver. Collectively, these data suggest a model in which the hypoglycemia that follows partial hepatectomy induces systemic lipolysis and accumulation of fat derived from peripheral stores in the early regenerating liver, and that these events provide or regulate essential signals for normal liver regeneration.
The specific mechanisms responsible for impaired liver regeneration in lipodystrophic fld mice require further elucidation. Future analyses should address whether the requirement for systemic adipose stores during normal liver regeneration is based on adipose as a source of metabolic fuel to support regeneration,38 lipid precursor for new membrane synthesis, a specific signal that initiates the regenerative response itself, or perhaps all of these. Our data showing that circulating levels of adiponectin are markedly reduced in fld mice together with published data demonstrating that adiponectin-null mice exhibit impaired liver regeneration26, 27 raise the possibility that this hormone may be such an essential adipose-derived signal. Because the gene that is mutated in fld mice, Lpin1, is also expressed in liver,22 another important consideration is that absence of hepatic Lpin1 expression might contribute to impaired regeneration in fld mice. In this regard, it is intriguing to consider that the Lpin1 gene product (lipin 1) is bifunctional in liver: It catalyzes an essential step in glycerolipid biosynthesis,39 which may be critical for synthesis of new cell membranes, and also coactivates peroxisome proliferator-activated receptor alpha (PPARα) activity, which is required for normal liver regeneration40, 41 and may be regulated by binding phospholipid.42 Nevertheless, the Lpin1 homolog Lpin2, whose protein product exhibits similar enzymatic and transcriptional coactivator activity, is also highly expressed in liver, and hepatic Lpin2 expression is increased in fld mice.39, 43 Ultimately, analyses of liver regeneration in other adipose-deficient lipodystrophic models and in adipose-specific and liver-specific Lpin1-null mice will be necessary to define the relative importance of each of these activities of Lipin1 during normal regeneration and the precise mechanisms responsible for deranged regeneration in fld mice.
We thank Trey Coleman for assistance with triglyceride and EchoMRI (Echo Medical Systems, Houston, TX) analyses.