Liver regeneration is a remarkably rapid and efficient process by which remnant hepatocytes, normally a quiescent population of cells, proliferate and restore the hepatic mass lost after chemical injury or partial hepatectomy.1-3 Studies examining the role of caveolin-1 (CAV1) during liver regeneration after partial hepatectomy in mice have produced contradictory results.4, 5 Using CAV1−/− mice developed in the Kurzchalia Laboratory (KCAV1−/− mice6) our research concluded that CAV1 plays an important role in the modulation of cellular processes during the first hours of liver regeneration.4 KCAV1−/− mice failed to undergo liver regeneration and to accumulate hepatic lipid droplets and progression through the cell cycle was arrested before entering S-phase in KCAV1−/− hepatocytes. As blood glucose and hepatic glycogen levels decrease a few hours after partial hepatectomy, hepatic lipid metabolism becomes essential for hepatocytes to undergo proliferation.7 Therefore, we postulated that CAV1 plays an important role in the modulation of lipid metabolism during liver regeneration in mice. Consistent with this hypothesis, we demonstrated that the wildtype phenotype is rescued by supplementing the diet of KCAV1−/− mice with glucose prior to surgery and during regeneration. In contrast, a separate study in CAV1−/− outbred mice from Jackson Laboratories (JAXCAV1−/− mice) described that JAXCAV1−/− mice showed a higher index of regeneration than wildtype mice after partial hepatectomy and with no significant effects on mouse survival after the operation, suggesting that CAV1 is not involved in liver regeneration.5 Here, by using three different strains of CAV1 null mice, we reassessed and confirmed the requirement of the expression of CAV1 in mice for efficient liver regeneration and lipid storage.
Caveolin-1 (CAV1) is a structural protein of caveolae involved in lipid homeostasis and endocytosis. Using newly generated pure Balb/C CAV1 null (Balb/CCAV1−/−) mice, CAV1−/− mice from Jackson Laboratories (JAXCAV1−/−), and CAV1−/− mice developed in the Kurzchalia Laboratory (KCAV1−/−), we show that under physiological conditions CAV1 expression in mouse tissues is necessary to guarantee an efficient progression of liver regeneration and mouse survival after partial hepatectomy. Absence of CAV1 in mouse tissues is compensated by the development of a carbohydrate-dependent anabolic adaptation. These results were supported by extracellular flux analysis of cellular glycolytic metabolism in CAV1-knockdown AML12 hepatocytes, suggesting cell autonomous effects of CAV1 loss in hepatic glycolysis. Unlike in KCAV1−/− livers, in JAXCAV1−/− livers CAV1 deficiency is compensated by activation of anabolic metabolism (pentose phosphate pathway and lipogenesis) allowing liver regeneration. Administration of 2-deoxy-glucose in JAXCAV1−/− mice indicated that liver regeneration in JAXCAV1−/− mice is strictly dependent on hepatic carbohydrate metabolism. Moreover, with the exception of regenerating JAXCAV1−/− livers, expression of CAV1 in mice is required for efficient hepatic lipid storage during fasting, liver regeneration, and diet-induced steatosis in the three CAV1−/− mouse strains. Furthermore, under these conditions CAV1 accumulates in the lipid droplet fraction in wildtype mouse hepatocytes. Conclusion: Our data demonstrate that lack of CAV1 alters hepatocyte energy metabolism homeostasis under physiological and pathological conditions. (HEPATOLOGY 2011)
Materials and Methods
Generation of Balb/CCAV1−/− and Balb/CCAV1+/+ Mice.
KCAV1−/− mice were backcrossed onto a Balb/C background by mating KCAV1+/− females to wildtype Balb/C males (Animal Resources Centre, WA). Once the Balb/C background was established by 10 generations of backcrossing (N10), heterozygote matings were used to generate CAV1 wildtype (WT) and null Balb/C mice from generation F1 onward. For our experiments we used mice from N10 F2 and F3 generations. Mice were restricted to same-generation pairs, avoiding sibling matings.
Animals, Reagents, and Partial Hepatectomy.
JAXCAV1−/− mice, the only commercial CAV1−/− mouse line available, strain Cav-1tm1Mls/J, and their corresponding controls were obtained from Jackson Laboratories.5, 8 KCAV1+/+ and KCAV1−/− mice were obtained as described.4 Mice were kept under a controlled humidity and lighting schedule with a 12-hour dark period. All animals received care in compliance with institutional guidelines regulated by the Australian Government. When applicable, mice were fed ad libitum with regular mouse chow or a high-fat diet (Research Diets, New Brunswick, NJ; #D12450B and #D12492) for 12 weeks before being sacrificed. For fasting experiments, food withdrawal was initiated at 6 AM when the lights in the animal house were switched on. Mice 10-14 weeks old were fasted for up to 24 hours prior to experimentation. After culling, liver pieces were frozen immediately in liquid nitrogen and stored at −80°C. The larger lobe of the liver was kept for purification of lipid droplets. Partial hepatectomies were carried out as before,4 except that in experiments involving liver regeneration following 2-deoxy-glucose (Sigma-Aldrich, Castle Hill, NSW, Australia) treatment (2-DG, 1 mL of 37 mM 2-DG intraperitoneally after partial hepatectomy), mice were not starved prior to partial hepatectomy. In these experiments we monitored five 2-DG-nontreated JAXCAV1+/+ mice, 15 2-DG-nontreated JAXCAV1−/− mice, seven 2-DG-treated JAXCAV1+/+ mice, and 10 2-DG-treated JAXCAV1+/+ mice during a regeneration time course. For examination of liver regeneration in Balb/Cmice, we subjected 8 Balb/CCAV1+/+ and 10 Balb/CCAV1−/− mice to partial hepatectomy. Mice were monitored during the first 24 or 48 hours of liver regeneration. In order to do a comparative analysis of liver regeneration between Balb/CCAV1+/+ and Balb/CCAV1−/− mice, and because four of the Balb/CCAV1−/− mice did not survive to 24 hours after operation, three Balb/CCAV1+/+ and three Balb/CCAV1−/− mice that survived 24 hours after partial hepatectomy were culled and their livers were collected for examination. The analysis of the progression of the liver regeneration was completed by the examination of three Balb/CCAV1+/+ and three Balb/CCAV1−/− mice at 48 hours after partial hepatectomy.
Lipid droplets were isolated as described.9 Homogenates for cell fractionation were obtained after liver disruption using Ultra Turrax T10 homogenizer (IKA, 47810 Petaling Jaya, Malaysia, #IKA3240000S). Polyclonal antibody against CAV1 was obtained from BD Biosciences (North Ryde, NSW, Australia) (#610060) and adipophilin (ADRP) antibody was from Progen Biotechnik (Heidelberg, Germany; #GP40). Mouse actin antibody Actin was from Chemicon, (North Ryde, NSW, Australia; #MAB1501).
Blood Glucose Measurement and Plasma Biochemical Analysis.
Blood was extracted by cardiac puncture and collected in BD Biosciences microcontainer PST LH (#365987). Blood glucose was measured in 4 μL of blood using the Accu-check blood glucose meter and strips (#03146332186). Plasma was collected after centrifugation of blood samples at 6000 rpm for 10 minutes at room temperature. Total cholesterol, triacylglycerol (TAG), and nonesterified fatty acids (NEFA) were analyzed in plasma samples at the Clinical Pathology Laboratory, School of Veterinary Science (University Of Queensland, Australia).
Liver Lipid Extraction and Thin Layer Chromatography (TLC).
Total hepatic lipid was extracted from 25-30 mg of liver tissue from KCAV1−/− and KCAV1+/+ mice and from 20 mg of liver from Balb/CCAV1−/− and Balb/CCAV1+/+ mice. Livers were homogenized in 200 μL of phosphate-buffered saline (PBS) using the Ultra Turrax T10 homogenizer. Lipid droplets were isolated as described.9 For lipid extraction, 900 μL of chloroform:methanol (1:2) was added and vortexed for 1 minute followed by gentle shaking 4°C for 2 to 3 hours. MilliQ water (300 μL) and chloroform (300 μL) were added, the samples vortexed for 1 minute, and incubated on ice for 1 minute. This procedure was repeated twice. Samples were then centrifuged at 9000 rpm for 2 minutes at 4°C to break phases. Finally, the organic phase was dried under a stream of N2 and stored at −80°C. For TLC, the dried lipid fraction was dissolved in 100 μL of chloroform:methanol (2:1) and 7.5 μL of each sample was run on TLC silica-gel plates (Sigma Aldrich, #Z265292) along with 7.5 μL of TAG standard (4.4 μg/μL) in 100 mL of hexane/diethyl ether/acetic acid (70:30:1). Lipid separation was observed in a UV illuminator after the plates were sprayed with 5% primuline in acetone:water 4:1. Quantification of TAG fractions was done with ImageJ software.
Liver samples were rapidly fixed by immersion in 2.5% glutaraldehyde in PBS and processed for Epon embedding by conventional methods. Stained ultrathin sections were analyzed by moving at random across the electron microscope (EM) grid (two grids per animal) and analyzing digital images taken at a magnification of 4,000× using the iTEM analysis program (Soft Imaging System, Muenster, Germany). A point counting grid was used to measure the volume density of lipid droplets relative to the total hepatocyte volume in random sections.
Quantitative Real-Time Polymerase Chain Reaction (PCR) and Primers.
RNA was extracted using RNAeasy (Qiagen) and 4-5 μg was reverse transcribed. Quantitative RT-PCR was performed in triplicate on three independent RNA preparations. Complementary DNA (cDNA) levels were analyzed in PCR reactions with SYBR Green Technologies (Applied Biosystems) and the relative level of expression was normalized using 18S ribosomal RNA. Statistical analysis was performed on the average of three independent assays using Student's t test. Primer sequences can be provided on request.
Energy expenditure, respiratory exchange ratio (RER), spontaneous physical movement, and food intake were measured simultaneously in each mouse with the Oxymax/CLAMS Comprehensive Lab Animal Monitoring System (Columbus Instruments, Columbus, OH) as described.10, 11 The calculations of carbohydrate and fat oxidation are based on stoichiometric equations reported by Watt et al.10, 11 CHO oxidation = (4.585 × VCO2) − (3.226 × VO2). The CO2 and O2 volume data from the metabolic chamber test were used in the calculation. Energy expenditure was calculated as before.12
Caveolin-1 Knockdown in AML12 Hepatocytes.
Stable CAV1 knockdown AML12 cell lines were developed using SHVRS MISSION short hairpin RNA (shRNA) Lentiviral Particles (Sigma Mission shRNA set) against mouse caveolin-1 following manufacturer protocols. Two of the five different lentiviral particles with shRNA targeting different sequences of mRNA codifying for CAV1 were able to knockdown CAV1 to around 90% of the CAV1 expression shown by the WT AML12 hepatocytes. These stable CAV1-deficient AML12 hepatocytes were termed CAV1-kd#2 and CAV1-kd#4, respectively. Cells were selected using puromycin (1 μg/mL) and pooled populations were used for experiments. WT AML12 hepatocytes were infected with lentiviral particles coding for a scrambled shRNA.
Bioenergetic Metabolism in CAV1-Deficient AML12 Hepatocytes.
Glycolytic rate was measured using the Seahorse XF24 analyzer. Cells were seeded into Seahorse V7 plates at 40,000 cells/well and 24 hours later cells were incubated in either high glucose media (25 mM glucose) or low glucose/oleate media (10 mM glucose, 100 μM oleate) for a further 24 hours. Cells were then washed twice in assay running media (unbuffered Dulbecco's modified Eagle's medium [DMEM] with 5 mM glucose) before being incubated in assay running media in a non-CO2 incubator at 37°C for 60 minutes. Basal extracellular acidification rate (ECAR), a proxy measure of glycolysis was then measured using the Seahorse XF analyzer over three measurement periods, each comprised of a 3-minute mix, 2-minute wait, and 3-minute measure cycles. To determine the effect of impaired oxidative adenosine triphosphate (ATP) production on ECAR, oligomycin was injected into the system at a final concentration of 1 μM. ECAR was then determined, again over three measurement periods, each comprised of a 3-minute mix, 2-minute wait, and 3-minute measure cycles. At the conclusion of the assay the assay media was removed and the Seahorse plate and cells were immediately frozen at −80°C for 24 hours. Plates were then thawed and the cell number in each well was determined using the CyQuant kit (Invitrogen) according to the manufacturer's instructions. ECAR values were then normalized to cell number, expressed as a ratio of 50,000 cells.
Statistical significance was assessed using Student's t test or analysis of variance (ANOVA) II in combination with Bonferroni's multiple comparison test unless otherwise indicated. Significance is indicated as (asterisks or another symbol) *P < 0.05; **P < 0.01; ***P < 0.001.
Genetic Background-Independent Requirement of CAV1 for Successful Liver Regeneration in Mice.
To examine the apparent inconsistency in liver regeneration between genetic backgrounds we generated and analyzed a third CAV1−/− mouse strain on a pure BalbC genetic background (Balb/CCAV1) (see Materials and Methods). As expected, Balb/CCAV1−/− mice did not express CAV1 and they showed lipodystrophy and dyslipidemia (Supporting Fig. S1a-c). In agreement with results obtained in KCAV1−/− mice,4 Balb/CCAV1−/− mice showed impaired liver regeneration. We analyzed the survival ratio of Balb/CCAV1−/− and Balb/CCAV1+/+ and the liver/body regeneration index as indicators of the progression of the liver regeneration. The total postoperation survival rate 48 hours after partial hepatectomy in Balb/CCAV1−/− mice was significantly lower than in Balb/CCAV1+/+ mice (60% in Balb/CCAV1−/− versus 100% in Balb/CCAV1+/+ mice) (Fig. 1A,B). In addition, approximately 80% of the CAV1−/− mice showed significantly delayed liver regeneration, as indicated by the liver/body regeneration index (Fig. 1E). At 24 hours after partial hepatectomy the total liver/body regeneration index (1.85 ± 0.16 versus 2.57 ± 0.11, P = 0.0059, n = 6 Balb/CCAV1−/− and n = 5 Balb/CCAV1+/+ mice, respectively) and the liver/body regeneration index from the deceased (1.51 ± 0.01 versus 2.57 ± 0.11, P = 0.00044, n = 3 Balb/CCAV1−/− and n = 5 Balb/CCAV1+/+ mice, respectively) and from the surviving (2.20 ± 0.03 versus 2.57 ± 0.11, P = 0.05, n = 3 Balb/CCAV1−/− and n = 5 Balb/CCAV1+/+ mice, respectively) Balb/CCAV1−/− mice were significantly lower than in Balb/CCAV1+/+ mice (Fig. 1C,D). Furthermore, analysis of the Balb/CCAV1−/− mice that reached 48 hours of liver regeneration suggested that despite lacking CAV1, some Balb/CCAV1−/− mice might show a compensative mechanism that allows progression of liver regeneration. However, the large variability observed in the values of the liver/body index obtained from the Balb/CCAV1−/− mice at 48 hours of liver regeneration when compared with Balb/CCAV1+/+ mice suggested that, although still progressing, lack of CAV1 perturbs liver regeneration and survival of Balb/CCAV1−/− mice. Taken together, the results clearly demonstrated that loss of CAV1 also impairs liver regeneration in Balb/CCAV1−/− mice.
We next analyzed JAXCAV1−/− mice, the only commercial CAV1−/− mouse line available, that were used by Mayoral et al.5, 8, 13 As shown previously,5 mice demonstrated normal liver regeneration after partial hepatectomy, had similar postoperation survival rates, and after 72 hours of regeneration the liver/body regeneration index was slightly but statistically significantly higher than in the JAXCAV1+/+ mice (3.34 ± 0.175 versus 2.69 ± 0.116, respectively, P = 0.0038) (Fig. 2A,E), suggesting faster regeneration in JAXCAV1−/− mice. Liver regeneration depends on the supply of both glucose and fatty acids to the remnant hepatocytes during the first hours of regeneration. As observed in KCAV1−/− and Balb/CCAV1−/− mice, hepatic oxidative lipid metabolism is disrupted during fasting in JAXCAV1−/− mice (Fernandez-Rojo et al., unpubl. results). In addition, it has been shown that high glucose levels can compensate for inefficient utilization of fatty acids.4, 14 We therefore hypothesized that JAXCAV1−/− hepatocytes primarily utilize glucose, in place of fatty acids, to support liver regeneration after partial hepatectomy. Supporting this hypothesis, JAXCAV1−/− mice showed significantly higher levels of blood glucose than KCAV1−/− mice after 24 hours of fasting (Fig. 2B). In addition, analysis of the respiratory exchange ratio (RER) by indirect calorimetric, a parameter indicating whether mice mainly use carbohydrates (RER = 1) or lipids (RER = 0.7) as a source of energy, showed that the absence of CAV1 increases carbohydrate metabolism in kCAV1 mice (Fig. 2C; Supporting Fig. S2a). However, our data revealed that in JAXCAV1 mice, and independently of the absence of CAV1, the genetic background provides a major preference for higher consumption of carbohydrates when compared with KCAV1−/− mice (Fig. 2C,D; Supporting Fig. S2a). Unlike kCAV1+/+ and kCAV1−/− mice, both JAXCAV1+/+ and JAXCAV1−/− mice showed RER values higher than 1, a well-characterized indicator of “anaerobic glycolysis”15 (also termed “aerobic glycolysis”16, 17) (Fig. 2C).
We next tested the role of carbohydrate metabolism during regeneration in JAXCAV1−/− mice by inhibiting glycolysis in vivo. JAXCAV1+/+ and JAXCAV1−/− mice were treated with 2-DG, a nonmetabolizable, competitive glucose analog, after partial hepatectomy.18 In comparison with untreated JAXCAV1+/+ and JAXCAV1−/− mice and to 2-DG-treated JAXCAV1+/+ mice, 2-DG-treated JAXCAV1−/− mice showed drastically reduced survival rates and were unable to undergo liver regeneration (Fig. 2E). 2-DG administration did not affect the well-being and survival of nonhepatectomized JAXCAV1−/− mice (data not shown), ruling out a systemic lethal effect of 2-DG in regenerating JAXCAV1−/− mice. Thus, these results demonstrate that the ability of JAXCAV1−/− mice to accomplish liver regeneration after partial hepatectomy is dependent on the availability of glucose by the hepatocytes. These results are also consistent with our previous observation that liver regeneration in the KCAV1−/− mice can be rescued by a high glucose diet.4 Furthermore, we obtain insights into the molecular mechanism that might stand behind the ability of JAXCAV1−/− mice to achieve liver regeneration. We analyzed the expression hepatic glucose-6-phosphate dehydrogenase (G6PD) and fatty acid synthase (FASN), whose products catalyze the rate-limiting steps of the pentose phosphate pathway (PPP) and lipogenesis, respectively. Both PPP and lipogenesis have been postulated as crucial metabolic pathways for biosynthesis of new biomass and then proliferation of transformed cells relying on aerobic glycolysis.16, 17 In agreement with the above data, JAXCAV1−/− mice showed higher levels of hepatic G6PD and FASN expression than JAXCAV1+/+ and kCAV1−/− mice (Fig. 2G). G6PD activity provides nicotinamide adenine dinucleotide phosphate, reduced form (NADPH) that is used in reductive anabolic reactions such as the synthesis of fatty acids. Thus, in agreement with indirect calorimetry, JAXCAV1−/− mice have a better hepatic carbohydrate-dependent anabolic metabolism that might provide them with an advantage in comparison with kCAV1−/− mice for liver regeneration.
Moreover, we also demonstrated that JAXCAV1−/− mice have a significantly reduced daily physical activity when compared with KCAV1−/− (Fig. 2F). As JAXCAV1−/− mice produced a similar calorie output to KCAV1−/− mice (Supporting Fig. S2b), an almost complete absence of movement might facilitate the recovery of JAXCAV1−/− mice and liver regeneration after partial hepatectomy. These results confirm that (1) independent of the genetic background, CAV1 plays an important role in the initial phase of liver regeneration by modulating metabolism; (2) JAXCAV1−/− mice undergo liver regeneration due to selective use of carbohydrates as the main source for fuel availability through aerobic glycolysis and the overactivation of the PPP and lipogenesis.
Finally, to study whether lack of CAV1 by hepatocytes directly promotes glycolysis flux in a cell-autonomous manner in hepatocytes, we developed two different stable cell lines of CAV1-deficient AML12 hepatocytes (CAV1-kd in general, CAV1-kd#2 and CAV1-kd#4 AML12 cell lines) (Supporting Fig. S2c). Extracellular flux analysis of cellular metabolism demonstrated that CAV1-kd AML12 hepatocytes showed higher ECAR ratio values (a direct measure of extracellular acidosis in response to glycolysis-dependent efflux of lactate) than WT cells in situations of metabolic stress (Supporting Fig. S2d,e). Although there was a tendency to increase acidosis, CAV1-kd AML12 hepatocytes cultured in high glucose did not show statistically significant higher ECAR ratio values (Supporting Fig. S2c) than the WT AML12 hepatocytes. Moreover, CAV1-kd AML12 hepatocytes showed similar maximal glycolytic capacity to the WT AML12 hepatocytes when we stimulated glycolysis by incubation with the ATP synthase activity inhibitor oligomycin (Supporting Fig. S2d). However, when cells were incubated in a medium low in glucose but supplemented with physiological concentrations of oleic acid (LG/OA medium), CAV1-kd AML12 hepatocytes showed higher ECAR values, even in the presence of oligomycin (Supporting Fig. S2e). These data suggested cell-autonomous effects on carbohydrate metabolism due to the loss of CAV1 in hepatocytes.
Defective Lipid Droplet Accumulation in CAV1−/− Hepatocytes.
We next analyzed lipid droplet (LD) accumulation in regenerating livers from Balb/CCAV1+/+ and Balb/CCAV1−/− mice. Hepatocytes from Balb/CCAV1−/− mice showed a greatly reduced accumulation of LDs during regeneration as compared with those from Balb/CCAV1+/+ mice (Fig. 3). Consistently, liver fractionation by ultracentrifugation showed a significantly decreased level of purified LDs when compared with Balb/CCAV1+/+ mice. Accordingly, the levels of ADRP, the main LD marker in nonadipose cells, in Balb/CCAV1−/− purified LDs were also significantly lower than in the purified LDs from Balb/CCAV1+/+ mice (Fig. 3B), although there were no significant differences in total ADRP protein levels in the liver homogenates (Fig. 3A). Moreover, quantification of hepatic TAG content by TLC demonstrated reduced TAG accumulation in 24-hour regenerating Balb/CCAV1−/− livers (Fig. 3C). Taken together, these data demonstrate decreased TAG content and accumulation of LDs in regenerating Balb/CCAV1−/− hepatocytes, supporting our previous results that the absence of CAV1 reduces hepatocyte ability for storage of TAG. Finally, we analyzed hepatic LD accumulation during liver regeneration in JAXCAV1+/+ and JAXCAV1−/− mice. Liver appearance from JAXCAV1−/− mice did not show high levels of steatosis. However, JAXCAV1+/+ also showed great variability in their steatotic appearance (data not shown). Accordingly, western blot analyses showed very variable expression of ADRP protein levels in both JAXCAV1+/+ homogenates and LD fractions (Fig. 3D,E). Thus, these results support the conclusions of Mayoral et al.5 suggesting that there were no significant differences in hepatic LD accumulation between JAXCAV1+/+ and JAXCAV1−/− mice during liver regeneration.
To further investigate the importance of CAV1 for the ability of hepatocytes to accumulate TAG and generate LDs, we analyzed two independent physiological models of hepatic LD accumulation in CAV1−/− mice: fasting and maintenance on a high-fat diet. First, we studied hepatic LD accumulation in KCAV1, JAXCAV1, and Balb/CCAV1 mice after 24 hours of fasting (Fig. 4; Supporting Fig. S3). When we compared KCAV1+/+ and KCAV1−/− mice, ADRP and GyK transcript levels, both involved in TAG synthesis, were significantly reduced in KCAV1−/− hepatocytes (Fig. 4A), as were ADRP protein levels in the liver (Fig. 4B,C) and in purified LDs (Fig. 4D) during different periods of fasting. Accordingly, hepatic TAG content and the percentage of the cytosolic area occupied by the LDs were significantly reduced in KCAV1−/− mice (Fig. 4E,F). Similar results were obtained in liver samples from 24-hour-fasted JAXCAV1+/+ and JAXCAV1−/− (Supporting Fig. S3a-c) and from Balb/CCAV1+/+ and Balb/CCAV1−/− mice (Supporting Fig. S3d-f).
We next studied the development of steatosis in response to a 12-week high-fat diet (HFD) in mice. Both, KCAV1+/+ and KCAV1−/− mice on the HFD showed increased levels of plasma lipids (TAG, total cholesterol) when compared with mice on a chow diet (Fig. 5A). Moreover, food consumption in KCAV1−/− mice was similar to KCAV1+/+ mice (data not shown). However, KCAV1−/− mice showed a lack of the typical steatotic liver phenotype as judged by several criteria (Fig. 5B). Analysis of ADRP levels in liver homogenates and purified hepatic LD fractions in combination with TLC-liver TAG content quantification and quantitative electron microscopic analysis of liver sections from chow- (data not shown) and HFD-fed mice all showed defective accumulation of TAG in LDs of KCAV1−/− hepatocytes in response to HFD (Fig. 5C-E). Similar results were obtained in liver samples from HFD-fed JAXCAV1+/+ and JAXCAV1−/− (Supporting Fig. S5a-c) and from HFD-fed Balb/CCAV1+/+ and Balb/CCAV1−/− mice (Supporting Fig. S5d-f). JAXCAV1+/+ mice showed some variability in their steatotic phenotype. However, LD quantification by EM again showed that the absence of CAV1 reduces the ability of hepatocytes to accumulate LDs (Supporting Fig. S5c). Finally, Balb/CCAV1+/+ mice fed an HFD increase slightly in weight when compared with chow-fed Balb/CCAV1+/+ mice. They also showed a higher weight gain (Supporting Fig. S4c.1) than Balb/CCAV1−/− mice in response to an HFD. However, and unlike KCAV1+/+ and JAXCAV1+/+ mice,19 and consistent with previously published work,20 they were resistant to obesity when fed an HFD (Supporting Fig. S4a-c) and even lost weight during the last 4 weeks on the HFD. Moreover, analysis of ADRP levels by western blot suggested that in comparison with hepatocytes from chow-fed mice, Balb/C hepatocytes have a refractory response to HFD that it is translated into a significant reduction in the accumulation of LDs in hepatocytes. However, HFD-fed Balb/CCAV1+/+ mice still showed a higher number of hepatic LDs (Supporting Fig. S5c) than HFD-fed Balb/CCAV1−/− mice.
Finally, we directly analyzed the association of CAV1 with LDs. In agreement with previous results obtained in regenerating livers from rats,9 we identified CAV1 in purified hepatic LD fractions from regenerating liver (Fig. 3B,E) from 24-hour-fasted liver (Fig. 4C,D; Supporting Fig. S3bc) and from the liver of mice on an HFD (Fig. 5C; Supporting Fig. S5b,e). These results clearly demonstrate that CAV1 associates with LDs in hepatocytes and that loss of CAV1 dramatically impairs the storage of TAG in LDs of mouse hepatocytes.
In this work we show that, independently of the mouse genetic background, the expression of CAV1 in mouse tissues facilitates the efficient progression of liver regeneration and accumulation of triacylglycerols in hepatocytes in mice. In two different mouse models the total absence of CAV1 in mice reduced hepatocyte ability to restore the liver mass lost after partial hepatectomy. Our data help resolve the controversy created by two different works that published opposite results.4, 5 Scientists in the field suggested that the impure genetic background used in both studies might be behind the origin of these conflicting data. Now, new data from experiments in pure Balb/CCAV1+/+ and Balb/CCAV1−/− mice showed that lack of CAV1 decreases mouse efficient progression of liver regeneration, supporting our previous published mouse model in KCAV1−/− mice.4 However, these results still did not answer why mice used by Mayoral et al.,5 JAXCAV1−/− mice, achieved liver regeneration despite their lack of CAV1. Using similar experimental conditions to those used by Mayoral et al., we confirmed that JAXCAV1−/− mice achieved liver regeneration and mouse survival was only slightly affected by the absence of CAV1. However, treatment of JAXCAV1−/− mice with 2-DG in combination with biochemistry and metabolic analysis demonstrated that in JAXmice lack of CAV1 also reduces the ability to perform liver regeneration when compared with JAXCAV1+/+ mice. Furthermore, this work also provides clues regarding the molecular mechanism that allowed liver regeneration in JAXCAV1−/− mice under Mayoral et al.'s conditions. Our metabolic profiling experiments demonstrated that the genetic background from JAXmice promotes systemic metabolism of carbohydrates including “aerobic glycolysis” instead of lipids as a source of energy during specific phases of the day. However, experiments in nonhepatectomized and hepatectomized mice treated with 2-DG demonstrated that JAXCAV1−/− mice specifically rely on hepatic carbohydrate metabolism during liver regeneration. Interestingly, and unlike in the KCAV1 mice that we used in our initial studies and in JAXCAV1+/+ mice, lack of CAV1 in JAXmice induced a carbohydrate-dependent anabolic adaptation based on increased activity of the PPP and lipogenesis in hepatocytes. Activation of these metabolic pathways is also seen in proliferating transformed cells.16, 17 These metabolic pathways provide NADPH and cell precursors for hepatocyte replication. Therefore, our data suggested that regenerating JAXCAV1−/− hepatocytes reproduced energetic metabolism used by transformed cells during the progression of cancer. Mayoral et al.13 suggested the impairment of transforming growth factor beta (TGF-β) signaling as a possible mechanism explaining accelerated liver regeneration after partial hepatectomy. However, during liver regeneration TGF-β signaling modulates growth arrest at the end of liver regeneration.3 Although the expression of TGF-β receptors and other proteins participating in this pathway are up-regulated after 24 hours of regeneration,21 the TGF-β pathway is not activated until day 4 or 5 after partial hepatectomy.3 Thus, it seems unlikely that the impairment of the TGF-β pathway would be responsible for the progression of liver regeneration during the first hours after partial hepatectomy in JAXCAV1−/− mice. In the absence of comparative data regarding TGF-β signaling in KCAV1−/− mice, impaired TGF-β signaling does not readily explain the controversy created between the original studies on liver regeneration in KCAV1−/− and in JAXCAV1−/− mice.4, 5
The experiments presented here with 2-DG have uncovered a defective metabolic phenotype that, in direct correlation with poor mouse survival, compromised liver regeneration in JAXCAV1−/− mice as compared with JAXCAV1+/+ mice. Moreover, basal analysis of key metabolic genes described metabolic adaptation that allows JAXCAV1−/− mice to regenerate their livers. We do not know yet if the different results obtained by our group and by Mayoral et al. are due to the two different methodologies used for knocking out CAV1, or if the phenotype described is specific to loss of hepatocyte CAV1. Future work should resolve whether different methodologies for knocking out CAV1 in mice can provoke partial adaptation responses, as in JAXCAV1−/− mice, or a lack of adaptation, as seen with KCAV1−/− mice, to metabolic stress situations such as liver regeneration in response to mechanical injury. However, this work clearly shows that, as in both Kmice and in Balb/Cmice, the absence of CAV1 in JAXmouse tissues also reduced the ability of hepatocytes to proliferate and regenerate after partial hepatectomy. Therefore, the expression of CAV1 is important for efficient liver regeneration in mice.
Whether liver regeneration and liver steatosis depends directly on hepatic CAV1 in mice is still unknown. However, our work shows that expression of CAV1 in mice maintains the ability of hepatocytes to store TAG in LD in physiological and pathological conditions of hepatic steatosis. This happens even in situations of high availability of NEFA and external TAG, such as in response to HFD, suggesting that the inability to store TAG may be independent of the lipodystrophy caused by the absence of CAV1 in adipose tissue. Furthermore, we demonstrate that CAV1 associates with a hepatic LD fraction in mice in response to fasting, HFD, and partial hepatectomy. Finally, our data using automated extracellular flux analysis of CAV1-kd AML12 hepatocytes, together with the observed defective liver regeneration in JAXCAV1−/− mice in the presence of 2-DG, supported cell-autonomous effects on carbohydrate metabolism caused by the loss of CAV1 in hepatocytes. Further work should establish the relative contribution of tissue-autonomous effects and general effects of the loss of CAV1 on hepatic physiology in health and disease.
We are grateful to the Australian Cancer Research Foundation (ACRF)/Institute for Molecular Bioscience (IMB) Dynamic Imaging Facility for Cancer Biology, established with funding from the ACRF. The authors acknowledge the use of the Australian Microscopy and Microanalysis Facility at the Center for Microscopy and Microanalysis at The University of Queensland. We thank Lukas Bahati and James Rae for assistant in lipid extraction and TLC performance, and Brian Bynon and Mark Ropper from the Clinical Pathology Laboratory at the University of Queensland for their assistance in the analysis of mouse plasma.