PTEN Down-Regulation Promotes β-Oxidation to Fuel Hypertrophic Liver Growth After Hepatectomy in Mice

ANIMALS

Male wild-type (wt) mice (C57BL/6) were from Envigo (Horst, NL). Male hepatocyte-specific inducible Pten knockout (PtenKO) animals (AlbCre-ERT2^Tg/+Pten^fl/fl) and corresponding controls (AlbCre-ERT2^+/+Pten^fl/fl, PtenC) were kindly provided by M. Foti (University of Geneva, Geneva, Switzerland). All animal experiments were in accord with Swiss Federal Animal Regulations and approved by the Veterinary Office of Zürich (KEK-ZH-Nr. 2012-203).

ANIMAL SURGERY AND SUBSTANCES

PH (68%) and 91% hepatectomy (a model of lethal liver failure) were performed as reported.18 Substances (Supporting Table S1) were intraperitoneally injected in 100-μL volume at doses and times indicated in the results. See the Supporting Information for further details.

HISTOLOGICAL STAINING

Hematoxylin and eosin (H&E), periodic acid Schiff, and immunochemical stainings (see Supporting Table S2 for antibodies) were performed on 3-μm archived liver sections and Oil Red O on cryosections.

HEPATOCYTE SIZE

Hepatocyte area was assessed on histology under exclusion of lipid droplets and vessels. Size was measured through forward scatter by flow cytometry.

WESTERN BLOTTING

The procedure was performed as reported.3, 18 Antibodies are listed in the Supporting Information.

QUANTITATIVE REAL-TIME POLYMERASE CHAIN REACTION

RNA extraction, complementary DNA synthesis and qPCR were performed as described.3, 18 Taqman gene expression assays (Applied Biosytems, Rotkreuz, Switzerland) are listed in the Supporting Information.

INDIRECT CALORIMETRY AND Echo-MAGNETIC RESONANCE IMAGING

Indirect calorimetry experiments were conducted at the Small Animal Metabolic Phenotyping core facility (University of Geneva) and approved by the Geneva Health Head Office. Energy expenditure and the respiratory exchange ratio (RER) were derived from O₂ consumption and CO₂ production; RER differences between PtenKO and PtenC were assessed through AUC (area under the curve) analysis.

LIPID AND GLYCOGEN MEASUREMENTS

Quantitation kits were used to measure triglycerides (TGs; ab65336; Abcam, Cambridge, UK). High-density lipoprotein (HDL; MAK045-1KT; Sigma-Aldrich, Buchs, Switzerland), and glycogen (MAK016-1KT; Sigma-Aldrich).

STATISTICAL ANALYSIS

Data are presented as mean ± SD unless stated otherwise. Differences between groups were assessed by a two-tailed t test assuming unequal variance. Five to 8 mice/group were analyzed. For survival after 91% hepatectomy, 10 animals/group were included. Differences were considered significant at P < 0.05 and indicated by an asterisk (*). Statistical analyses were performed using Prism software (version 6.0; GraphPad Software Inc., La Jolla, CA).

PTEN IS ASSOCIATED WITH TRAS TURNOVER, WEIGHT GAIN, AND HYPERTROPHY IN REGENERATING LIVER

To explore the roles of PTEN and TRAS, we assessed lipid-associated parameters and PTEN levels following standard (68%) hepatectomy (PH) in wt mice. Histology confirmed the TRAS peak at 16 hours and its gradual disappearance around the times of major parenchymal growth.3, 6 The steatotic peak was preceded by the induction of perilipin 2 (Plin2), which promotes hepatocyte-adipocyte transdifferentiation and is required for lipid droplet formation (Fig. 1A).19 Around the steatotic peak, Cd36 (fatty acid translocase) was up-regulated, consistent with peripheral import of fat.4 Furthermore, β-oxidation genes (carnitine palmitoyltransferase 1a [Cpt1a], hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase (trifunctional protein), alpha/beta subunit [Hadha/b]) were elevated at the expense of lipogenic genes (stearoyl-CoA desaturase 1a [Scd1a], acetyl-CoA carboxylase (Acc), fatty acid synthase [Fasn]; Fig. 1B), suggesting that fat is being accumulated for energetic needs.

PTEN protein was down-regulated at the TRAS peak to levels (≤80%) known to induce carcinogenesis in mice.8 PTEN down-regulation occurred after Plin2 induction and persisted during lipid disappearance (Fig. 1C). Reduction in PTEN may hence be associated with TRAS turnover, but not with its accumulation.

To estimate PTEN's function during LR, we simulated elevated PTEN activity through inhibition of PI3K. Injection of wortmannin (0.75 mg/kg) at 13 hours post-PH led to reduced mitoses, a reduced hepatocyte size (however, only if lipid vesicles were excluded from hepatocyte area), but an increased liver-to-body-weight-ratio (Lw/Bw) at 48 hours. The latter was likely attributed to lipid accumulation, which was strongly elevated compared to vehicle controls (Fig. 1D).

When PTEN was inhibited by bisperoxovanadium (bpV; 3.3 mg/kg) at 13 hours post-PH, Lw/Bw and hepatocyte size were increased, whereas TRAS was diminished (Fig. 1E). Notably, mitotic counts were also reduced through bpV, suggesting that PTEN inhibition during LR specifically affects the phosphoinositide-dependent AKT-mTOR axis, which promotes hypertrophy at the expense of hyperplasia.12 In contrast, PI3K inhibition additionally affects phosphoinositide-independent STAT3 activation, hence impacting on both hyperplasia and hypertrophy.12 Taken together, these findings indicate that PTEN down-regulation is associated with TRAS turnover, weight regain, and hypertrophy in regenerating liver.

HEPATOCYTE-SPECIFIC Pten DEFICIENCY ACCELERATES FUNCTIONAL LIVER RECOVERY THROUGH HYPERTROPHY

PTEN inhibition after hepatectomy promotes liver weight recovery, however bpV acts systemically and may exert unspecific effects. We therefore used inducible hepatocyte-specific Pten knockout mice to define the impact of PTEN deficiency on LR. Knockout was induced by Tyro3, Axl, and Mertk (TAM) in AlbCreERT2^tg/+-Pten^fl/fl (PtenKO) 4 days before hepatectomy to avoid pre-existing fatty liver that may impair regenerative capacity.20 Pten expression was not significantly altered post-PH in Cre-lacking control mice (PtenC; Fig. 2A), suggesting that PTEN down-regulation (Fig. 1C) is regulated posttranscriptionally following resection. Regenerating livers in PtenKO mice remained Pten deficient, reflecting liver reconstitution from differentiated (albumin-positive) hepatocytes (Fig. 2A).

At PH, the starting weight of the liver remnant was slightly increased in PtenKO relative to PtenC. Post-PH, the difference in liver weight became more pronounced over time, leading to hepatomegaly in PtenKO after a week (Fig. 2A). Notably, accelerated weight gain was not associated with proliferation, but with enhanced hepatocellular hypertrophy typified through deficient S/M phase progression and p27 up-regulation (Fig. 2B,C and Supporting Figs. S1 and S2).

To determine whether accelerated weight gain leads to improved liver recovery, we performed 91% hepatectomy, which causes lethal liver failure in wt mice.18 The best measure to assess recovery of liver function is 7-day survival, the critical period after liver loss. Remarkably, survival was raised to 40% after 91% resection in PtenKO mice (Fig. 2D), indicating that the surplus hypertrophic liver mass generated by PTEN deficiency is functional.

Given the significant hypertrophy at 72 hours post-PH in PtenKO, we investigated AKT-mTOR-S6K (S6 kinase) signaling, known to promote a hyperplasia-to-hypertrophy switch.12, 21 Activating AKT phosphorylation was markedly increased by PTEN deficiency (Fig. 2E). Likewise, enhanced S6K phosphorylation indicated elevated mammalian target of rapamycin complex 1 (mTORC1) activity. Increased phosphorylation of AKT and S6K was also evident at 32 hours (Supporting Fig. S3), consistent with AKT-mTORC1-S6K signaling as a hypertrophic driver in regenerating liver from PtenKO.

Pten DEFICIENCY PROMOTES GLUCOSE STORAGE AND LIPID METABOLISM IN REGENERATING LIVER

Acceleration of LR in PtenKO is expected to rely on additional energy supply. We hence assessed hepatic energy stores, mainly consisting of TGs and glycogen.

Following hepatectomy, both PtenKO and PtenC displayed early drops in liver glycogen and serum glucose (Supporting Fig. S4A), a reported signal required for initiation of regeneration.4 Notably, glycogen stores recovered faster in PtenKO relative to PtenC, accompanied by down-regulation of gluconeogenic expression (peroxisome proliferative activated receptor gamma coactivator 1-alpha, phosphoenolpyruvate carboxykinase 1, glucose 6 phosphatase; Supporting Fig. S4A). Given the similar serum glucose levels, these findings indicate that PTEN deficiency counteracts glucose usage in regenerating liver.

Hepatic TG content was elevated in PtenKO at hepatectomy and peaked at 16 hours akin to PtenC (Fig. 3B). The subsequent decline in TG levels was delayed in PtenKO relative to PtenC. Therefore, loss of PTEN before hepatectomy leads to an expanded TRAS period. In contrast, inhibition of PTEN at the TRAS peak shortened this period (Fig. 1E), suggesting that PTEN deficiency does not directly enhance TRAS formation during LR. Indeed, PTEN loss did not affect Plin2 and Cd36 expression post-PH and even down-regulated fatty acid binding protein 4 (Supporting Fig. S4B), which is required for hepatocyte-adipocyte transdifferentiation and is enhanced in resting PtenKO liver.10

Resting PtenKO liver is known to amass fat partially through increased lipid synthesis.9, 11 Accordingly, lipogenic molecules (sterol regulatory element-binding protein-1 [SREBP1] and its transcriptional targets, Scd1 and Fasn) were up-regulated at hepatectomy in PtenKO liver remnants (Fig. 3B). After hepatectomy, however, the lipogenic program was suppressed (Fig. 3B), indicating little contribution of hepatic lipogenesis to TRAS. An expanded TRAS period may hence be secondary to the increased hepatoperipheral lipid shuttle pre-existing in PtenKO before PH.9 Indeed, the elevations in fibroblast growth factor 21 (FGF21), a hepatokine that promotes the import of peripheral lipids for their hepatic use,11, 22 were extended in PtenKO relative to PtenC and correlated with respective TRAS periods (Fig. 3C). In agreement, serum TGs were elevated in PtenKO at PH, but dropped during the TRAS peak similar to PtenC (Fig. 3C). Likewise, serum levels of HDL—transporting TGs into liver—were increased in PtenKO versus PtenC during LR (Fig. 3C), along with an elevated hepatic expression of hormone-sensitive lipase (Lipe) and lipoprotein lipase (Lpl; Fig. 3C)—lipases that free fatty acids from TGs.23, 24

Altogether, these results suggest that the expanded TRAS period in PtenKO results neither from increased lipogenesis nor from elevations in active lipid import. Rather, the systemic redistribution of lipids into liver after tissue loss4, 25 may be enhanced because of an elevated mobilization of peripheral fats pre-existing in PtenKO liver before PH.9, 10 Given that FGF21 can increase hepatic energy expenditure,22 the expanded TRAS period might further relate to a prolonged catabolic phase in regenerating PtenKO liver.

Pten DEFICIENCY PROMOTES LIPID OXIDATION AS AN ENERGY SOURCE AFTER HEPATECTOMY

To gain insight into concrete metabolic outputs during LR and the effects of PTEN loss thereupon, we measured RER (CO₂/O₂) post-PH by indirect calorimetry. The typical diurnal shift toward lipid oxidation was recorded for both PtenKO and PtenC (Fig. 4A; control data in Supporting Fig. S5). Upon resection, RER markedly dropped during the first 32 hours, identifying lipid oxidation as the favored energy source during the TRAS period of regenerating liver. Unlike PtenC, RER around the TRAS peak remained close to 0.7 in PtenKO, indicating increasing fat catabolism with lowering PTEN levels. From 32 hours onward, RER began to rise toward prehepatectomy levels. Another reduction in RER was recorded in mutant animals toward 72 hours, thus before TGs were disappearing in PtenKO liver. At 16 hours, peroxisome proliferator-activated receptor alpha (PPARα; a key transcription factor promoting β-oxidation) and its targets, CPT1A and carnitine octanoyltransferase (CROT), were up-regulated in PtenKO liver (Fig. 4B), along with increased expression of several β-oxidation genes (Supporting Fig. S6). Therefore, LR leads to a profound increase in lipid oxidation, which is enhanced through hepatic PTEN down-regulation and correlates with the disappearance of liver fat.

Pten DEFICIENCY FUELS HYPERTROPHIC LR THROUGH β-OXIDATION OF TRAS-DERIVED LIPIDS

The increases in lipid oxidation observed in PtenKO animals after hepatectomy may originate from the metabolism of TRAS lipids to fuel the regenerative process. If TRAS lipids are oxidized to provide energy for liver growth, inhibition of β-oxidation should result in persisting steatosis and a diminished liver weight after hepatectomy. We targeted the rate-limiting β-oxidation enzyme, CPT1A, using etomoxir. Low doses (10 mg/kg) were applied every 12 hours starting 16 hours post-PH, so as to inhibit lipid oxidation without compromising survival. At 72 hours post-PH—when liver is lean and PtenKO display hypertrophy with elevated liver weight—etomoxir did not significantly alter Lw/Bw or hepatocyte size in PtenC relative to vehicle controls, despite a mild increase in hepatic lipids (Fig. 4C). The lack of distinct effects may relate to the low dose, but also to the inefficacy of etomoxir in inhibiting the oxidation of medium and short-chain fatty acids. In PtenKO, however, etomoxir markedly increased hepatic TG content while reducing liver weight. Moreover, hepatocyte size of PtenKO was diminished by etomoxir and no more different from that of PtenC hepatocytes (Fig. 4C). Therefore, mild inhibition of β-oxidation does counteract hypertrophic parenchymal growth driven by PTEN loss; however, it appears insufficient to impact on the general regeneration process. We conclude that TRAS provides lipids for β-oxidation to fuel hypertrophic LR, a process promoted by PTEN down-regulation.

POTENTIAL CATABOLIC AND HYPERTROPHIC MEDIATORS DOWNSTREAM OF AKT

Overactive AKT signaling (Fig. 2E and Supporting Fig. S3) is likely responsible for the metabolic and hypertrophic changes in regenerating PtenKO liver. AKT can induce β-catenin26 (directly or through glycogen synthase kinase 3 beta [GSK3b] inhibition), which is an important promoter of lipid catabolism in the liver.27 On immunofluorescence, we observed increased nuclear expression of β-catenin at 16 hours post PH in PtenKO, along with elevated inhibitory GSK3b phosphorylation (Fig. 5A). In support of its catabolic role, hepatic targets of β-catenin (acyl-CoA dehydrogenases27) were up-regulated in PtenKO at 16 hours (Supporting Fig. S6). However, chronic overactivity of the hypertrophic driver, mTORC1,4 may likewise promote β-oxidation.17 We treated mice with moderate doses of the mTORC1 inhibitor rapamycin (1 mg/kg at 13 hours post-PH and subsequently every 24 hours). In PtenC at 72 hours postresection, rapamycin had little effect on liver weight, hepatocyte size, and TG content (Fig. 6B). In PtenKO, rapamycin reduced liver weight and hepatocyte size, but did not significantly impact on TG content. These findings portray mTOR as a major promoter of hypertrophy in Pten loss-driven regeneration, whereas associated β-oxidation may preferably be regulated by other AKT-dependent molecules such as β-catenin (Fig. 6C).