PPARβ Regulates Liver Regeneration by Modulating Akt and E2f Signaling

The current study tests the hypothesis that peroxisome proliferator-activated receptor β (PPARβ) has a role in liver regeneration due to its effect in regulating energy homeostasis and cell proliferation. The role of PPARβ in liver regeneration was studied using two-third partial hepatectomy (PH) in Wild-type (WT) and PPARβ-null (KO) mice. In KO mice, liver regeneration was delayed and the number of Ki-67 positive cells reached the peak at 60 hr rather than at 36–48 hr after PH shown in WT mice. RNA-sequencing uncovered 1344 transcriptomes that were differentially expressed in regenerating WT and KO livers. About 70% of those differentially expressed genes involved in glycolysis and fatty acid synthesis pathways failed to induce during liver regeneration due to PPARβ deficiency. The delayed liver regeneration in KO mice was accompanied by lack of activation of phosphoinositide-dependent kinase 1 (PDK1)/Akt. In addition, cell proliferation-associated increase of genes encoding E2f transcription factor (E2f) 1–2 and E2f7–8 as well as their downstream target genes were not noted in KO livers 36–48 hr after PH. E2fs have dual roles in regulating metabolism and proliferation. Moreover, transient steatosis was only found in WT, but not in KO mice 36 hr after PH. These data suggested that PPARβ-regulated PDK1/Akt and E2f signaling that controls metabolism and proliferation is involved in the normal progression of liver regeneration.


Introduction
Proliferating cells require metabolic activity to generate energy and intermediates for the biosynthesis of macromolecules used for producing cell or tissue mass [1]. Increased aerobic glycolysis and fatty acid (FA) synthesis are both considered as major metabolic alterations during cell proliferation. Cell proliferation is controlled by the cell cycle, which is regulated by Cyclins, Cyclin-dependent kinases (Cdks), or E2fs [2].
Peroxisome proliferator-activated receptors are nuclear receptor ligand-depended transcription factors that regulate gene expression. Three types of PPARs (a, b/d and c) were identified. Even though they share sequence similarity, they all have unique physiological functions involved in control of metabolism [3]. PPARa regulates FA transport and metabolism and regulates energy homeostasis while PPARc is involved in adipocyte differentiation and lipid storage in adipose tissue [4]. PPARb participates in the regulation of lipid and glucose metabolism, wound healing and inflammation [5].
The roles of PPARa and PPARc in liver regeneration were studied and different results were obtained [6][7][8][9][10]. Some studies showed that PPARa deficiency delayed liver regeneration after partial hepatectomy (PH) in mice through inhibition of genes involved in cell cycle control, cytokine signaling, fat metabolism, and impaired Ras signaling [6,8]. Metabolomic studies also revealed activation of PPAR signaling and increased lipid metabolism in regenerating rat livers after PH [11]. Pharmacological activation of PPARa in rodent causes hepatomegaly and leads to the development of liver cancer indicating its role in proliferation [12]. However, other studies concluded that PPARa is not essential for the increased expression of Cdks and Cyclins as well as cell proliferation after PH [7]. Similarly, contradicting data have been reported for PPARc, activation of PPARc by pioglitazone inhibits PH-induced hepatocyte proliferation in rat [13], while others identified that transactivation of the PPARc signaling pathway by fatty acids is essential for rat liver regeneration [11]. Moreover, liver-specific PPARc-null mice with diet-induced hepatic steatosis have reduced hepatic regeneration after PH [10]. Thus, a clear reappraisal for the role of PPARa and PPARc in liver regeneration might be necessary.
Unlike PPARa and PPARc, the role of PPARb in liver regeneration has not been studied. In addition to regulating glucose and lipid metabolism, PPARb displays an anti-inflammatory activity, which could be important in the modulation of liver regeneration [14]. Ligand activation of PPARb protects against CCl 4 -induced hepatotoxicity by repression of pro-inflammatory genes [15]. Compared with its anti-inflammatory role, the function of PPARb in regulating cell fate is more complicated. GW501516activated PPARb promotes liver repair by stimulating hepatic stellate cell proliferation via the p38 and JNK MAPK pathways in CCl 4 treated mice [16]. Regarding dermatological wound healing, PPARb increases mouse keratinocyte survival via activation of PDK1/Akt signaling or ceramide kinase after injury [17,18]. However, ligand activation of PPARb inhibits cell proliferation in human HaCat keratinocytes [19]. In tumor cell lines, PPARb activation promotes the proliferation of human breast and prostate cancer cell, but had no effect on the proliferation of A549 and H1838 human lung cancer cells [20,21]. Hence, the proliferative effect of PPARb seems to be cell type-specific and the role of PPARb in liver regeneration remains to be determined.
The current study tests the hypothesis that PPARb has a role in regulating liver regeneration. The data showed that liver regeneration was delayed in PPARb-null (KO) mice. Differential gene expression profiling revealed the inhibition of expression in genes and pathways that are involved in metabolism and proliferation in regenerating KO livers. Specifically, PPARb deficiency affected the activation of Akt and the expression of E2fs. Pathways that control glycolysis, FA synthesis as well as cell proliferation were de-regulated in regenerating PPARb KO livers. The data suggest a role for PPARb in regulating liver regeneration that is mediated at least in part through Akt and E2f-regulated pathways.

Mice, Partial Hepatectomy, and Sample Preparation
Wild-type (WT) and PPARb-null mice (KO) male mice (3-5 month, C57BL/6) were kept in steel microisolator cages at 22uC with a 14-hr/10-hr light/dark cycle. Food and water were provided ad libitum throughout the entire feeding period. Standard PH was performed using the procedure described previously [1]. Mice were killed at the indicated time-points. The liver and body weights at the time of death were used to calculate the liver-tobody-weight ratios. The results obtained were the mean of three to five mice per time point. Part of the livers were fixed in 10% formalin, embedded in paraffin, and stained for histological analysis. All the animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals under protocols approved by the University of California Davis Animal Care and Use Committee.

Ki-67 Immunostaining
Immunostaining was performed with primary Ki-67 antibody (NeoMarkers, Fremont, CA) to monitor hepatocyte proliferation. The number of Ki-67-labeled nuclei was counted in at least 10 low-magnification (20X) microscope fields for each section.

Western Blot
Liver protein (40 mg) was electrophoresed on SDS-polyacrylamide gels under reducing conditions. Proteins from the gels were transferred to the polyvinylidene fluoride membrane. Antibodies specific for PDK1, Akt, p-Akt (Thr308), Cyclin D, Cyclin E (Cell Signaling Technology, Danvers, MA), and b-Actin (Santa Cruz Biotechnology, Santa Cruz, CA) were used for detection of proteins.

RNA-sequencing Library Construction, Sequencing, and Bioinformatics Analysis
Mouse liver RNA was prepared using TRIzol (Invitrogen, Carlsbad, CA). RNA concentration and integrity were determined by the Agilent 2100 Bioanalyzer using a RNA Nano Bioanalysis Chip. RNA-sequencing library preparation and sequencing was carried out by the Genome Sequencing Facility at University of Kansas Medical Center (Kansas City, KS). cDNA libraries were prepared with 2 mg of total RNA using the TruSeq RNA Sample Preparation Kit (Illumina). The libraries were clustered and sequenced on an Illumina HiSeq 2000 instrument with 100 bp single end reads.
Total reads of RNA-sequencing experiments were analyzed using the combination of TopHat (2.0.0) and Cufflinks (1.3.0) [22]. TopHat was used to align reads to the mouse reference genome (NCBI37, mm9) and discover transcript splice sites. These alignments from TopHat were then assembled into transcripts using Cufflinks. Cuffdiff, a component of the Cufflinks package, was used to estimate FPKM (fragments per kilobase of exon model per million mapped fragments) and differentially express transcripts. Statistics on the libraries and the number of tags mapped are presented in Table 1. The differential expression analysis was done by using the Baggerley's test. The genes that have differential expression levels were extracted with a Benjamini-Hochberg corrected p-value ,0.005 for pathway analysis. The RNA-seq data discussed in this paper have been deposited in NCBI Gene Expression Omnibus and are accessible through Series accession number GSE47062 (http://www.ncbi.nlm.nih.gov/geo/query/ acc.cgi?acc = GSE47062). All biological function and pathway analyses were performed using the Functional Annotation Tool in the Database for Annotation, Visualization and Integrated Discovery (DAVID, david.niaid.nih.gov). Functional pathways or processes with p,0.05 and Bonferroni value ,0.1 were accepted.

Real-time Quantitative Polymerase Chain Reaction (qPCR)
Hepatic RNA isolated by TRIzol (Invitrogen, Carlsbad, CA) was reverse transcribed to generate cDNA followed by amplification using the ABI Prism 7900HT sequence detection system (Applied Biosystems, Foster City, CA). The hepatic mRNA levels were normalized to GAPDH mRNA level.

Statistical Analysis
Data are given as mean 6 SD. Statistical analysis was performed using Student's t test or one-way analysis of variance. Significance was defined by p,0.05.

Growth Suppression after PH in PPARb-null Mice
PH was done in WT and KO mice, and livers were collected 1-3 days after the surgery. Liver-to-body weight ratios were significantly reduced in regenerating KO livers in comparison with the regenerating WT livers at 36-48 hr (Fig. 1A). Ki-67 immunohistochemistry showed that the number of proliferating hepatocytes rapidly increased and peaked at 36-48 hr after PH in WT livers (Fig. 1B). In contrast, the number of Ki-67-positive hepatocytes in KO livers was significantly less than that in WT livers at 36-48 hr (Fig. 1C). However, the significantly higher number of Ki-67-positive hepatocytes in KO livers than that in WT livers indicated the compensatory proliferation in KO livers after delayed regeneration. Since Ki-67 protein is present during all active phases of the cell cycle (G1, S, G2, and mitosis), but is absent from resting cells (G0), the data indicated that there was about one day delay in cell proliferation due to lack of PPARb.

Differential Gene Expression Profiling after PH
RNA-sequencing was performed using hepatic RNA derived from WT and KO mice 48 hr after PH to study the mechanism underlying delayed cell proliferation due to PPARb deficiency. The data showed that 88% of the reads were mapped to mm9. Figure 2A showed that 1344 genes (1126 down-and 218 upregulated) had more than a 2-fold change at the mRNA level due to PPARb deficiency. Biological function analysis of these genes showed that a majority (68%) are involved in cell cycle control, DNA replication, and lipid homeostasis. Moreover, 437 out of 1344 genes could be assigned to 34 KEGG pathways that include 21 metabolism and 13 cell proliferation pathways (Fig. 2B).

Suppression of PDK1/Akt and Abolishment of E2f Activation by PPARb Deficiency
Since PDK1/Akt is downstream to PPARb [23], we determined the expression of PDK1/Akt in regenerating WT and KO livers by western blotting. The results showed that the PDK1/Akt pathway was activated 36-48 hr after PH in WT livers and such activation was absent in KO livers (Fig. 3A). It was shown that Akt can regulate E2f activity [24], so the expression of E2fs was studied by qPCR. The data showed that expression levels of E2f1-2, and E2f7-8 peaked at 36-48 hr when hepatocytes are actively proliferating in WT mice ( Fig. 4B-E). However, the cell proliferation-associated induction of E2f1-2, and E2f7-8 mRNA was either reduced or not found in regenerating KO livers. Transcript isoforms analysis was performed for the E2fs. Changes in E2f1-2 and E2f7-8 expressions were mainly due to the increased expression levels of isoforms ENSMUST00000103145 (E2f1), ENSMUST00000061721 (E2f2), ENSMUST00000073781 (E2f7), and ENSMUST00000058745 (E2f8) 48 hr in WT mice (Fig. 3F). The levels of E2f3-4 and E2f5-6 mRNAs did not change in both WT and KO regenerating livers (data not shown). Consistently, qPCR data showed that the expressions of 38 E2fs target genes involved in cell cycle control, DNA repair and replication, as well as G2/M checkpoints were suppressed due to PPARb deficiency (48 hr after PH) (Fig. 4A). We further studied the expression of some of the differentially expressed E2fs targets during liver regeneration in both WT and KO mice. The expression of a few genes involved in cell cycle control was also studied due to their importance for proliferation. The data showed that the expression of Cyclin D was higher in WT than KO at 24-36 hr (Fig. 4B). In addition, western blots indicated that Cyclin D level was induced at 36-48 hr in WT mice and the increase was absent in regenerating KO mice (Fig. 4C). Similarly, quantification of hepatic Cyclin E/Cdk2 and Cyclin A, B/Cdk1 mRNAs revealed that they were increased in regenerating WT, but not in KO livers (36-48 hr) (Fig. D-H). The level of Cyclin E protein was increased at 36-48 hr in WT, but not in KO, mouse livers (Fig. 4I).

Inhibition the Expression of Genes Involved in Glycolysis in Regenerating PPARb-null Livers
Nineteen genes involved in the glycolysis pathway showed differential expression in regenerating WT and KO livers (.1.5 fold). All the shaded genes shown in figure 5A had elevated expression levels in regenerating WT, but not in KO, mouse livers (48 hr). The findings were confirmed by qPCR. The expression of genes involved in glycolysis (Hexokinase 2, Hk2; Glucose phosphate isomerase 1, Gpi1; Phosphofructokinase, Pfkl; Aldolase C, Aldoc; Glyceraldehyde-3-phosphate dehydrogenase, Gapdhs; Phosphoglycerate kinase 1, Pgk1; Phosphoglucomutase 1, Pgm1; Enolase 1, Eno1; Pyruvate kinase liver and red blood cell, Pklr) were analyzed at time-points during liver regeneration. All except Hk2 at 24 hr, showed higher expression levels in WT than KO at 24, 36, and 48 hr after PH (Fig. 5B-J). Four genes, Pfkl, Aldoc, Gapdhs, and Pgm1, showed higher expression levels in KO than WT at 60 or 72 hr after PH thus suggesting a compensatory effect at a later time.
Inhibition the Expression of Genes Associated with Cholesterol, Triglyceride, and Fatty Acid Biosynthesis in Regenerating PPARb-null Livers Ninety-seven genes that regulate lipid homeostasis showed differential expression in regenerating WT and KO mouse livers. Among them, 14 and 12 genes are involved in cholesterol metabolism and FA synthesis, respectively (Fig. 6A-B). qPCR has confirmed that all of them had higher expression levels in WT mice. The expression of four genes (Sterol regulatory elementbinding protein, Srebp; ATP citrate lyase, Acly; FA synthase, Fasn; Acetyl-Coenzyme A carboxylase alpha, Acc) were also examined at all time-points during liver regeneration. The induction of Srebp peaked at 24 hr and was sustained until 48 hr in WT mice. However, the induction of Srebp in KO did not occur until 60 hr after PH (Fig. 6C). Consistent with the expression profile of Srebp in WT and KO mice, all the studied downstream targets of Srebp (Acly, Fasn, and Acc) showed higher expression in WT than KO 36-48 hr after PH. (Fig. 6D-F).
Morphological data showed that WT mice accumulated lipid droplets in hepatocytes 36 hr after PH (Fig. 7A) whereas this transient steatosis was not noted in KO mice at the same timepoint. There were no signs of steatosis in sham-operated WT and KO mice. Moreover, the expression levels of PPARa and PPARc were also higher in WT than KO mice 36 hr after PH (Fig. 7B-C).

Discussion
The role of PPARb in regulating lipid and carbohydrate homeostasis as well as proliferation has been shown in adipose tissue, muscle, skin, lung, colon using in vivo and in vitro models [25]. However, the results are inconsistent across the different models. For example, synthetic PPARb agonists promote cholesterol accumulation in human macrophages [26]. However, fat mass is not reduced in adipose-specific PPARb-null mice [27]. In terms of cell proliferation, human keratinocytes treated with PPARb ligands (L165041 or GW501516) had increased cell proliferation [28], while decreased cell proliferation was found in human HaCaT keratinocytes treated with GW501516 [29]. Controversial findings also exist in with PPARb function in various cancer cells. For instance, PPARb activation has been shown to stimulate proliferation of breast cancer cells (MCF-7, T47D, LNCaP), but not human colon carcinoma cells (HT29, HCA-7, SW480) [20]. Despite efforts to study PPARb in different tissues, the role of PPARb in regulating hepatocyte proliferation has never been studied. Using the PH-induced liver regeneration model, we report for the first time that PPARb is involved in liver regeneration. Gene profiling after PH of WT and KO mice provided insight into the cellular mechanisms relevant to the role of PPARb in liver regeneration. Our data showed that PPARb has   a role in liver regeneration by regulating metabolism and cell proliferation. It is clear that proliferation requires an adapted metabolic response of the cells; hence, PPARb-mediated metabolism is likely linked to cell proliferation.
Akt, which is a downstream target of PPARb [24], has a dual and integrated role in regulating metabolism and proliferation. In terms of metabolism, Akt regulates glucose homeostasis as well as FA synthesis [30]. Activation of Akt in 4-hydroxytamoxifentreated human retinoic pigment epithelial cells induces the expression of genes that are involved in cholesterol and FA biosynthesis [31]. In addition, Akt can activate Srebp to upregulate the expression of Fasn and Acc, which are the key enzymes involved in FA synthesis [32]. Our data showed the activation of PDK1/Akt, which was accompanied by upregulation of Srebp, Acly, Fasn, and Acc 24-36 hrs after PH in WT mice. However, such coordinated up-regulation was not found in regenerating PPARb-null mice. Of particular importance is that Acly is a key enzyme integrating glucose and lipid metabolism pathways [33]. For proliferation to occur, glycolytic flux needs to be induced to convert glucose into pyruvate, which leads to the production of lactate and acetyl-CoA to facilitate FA synthesis [34]. The up-regulation of Acly in regenerating liver can result in increased acetyl-CoA production, leading to enhanced cholesterogenesis and lipogenesis [35], which may cause fat accumulation found in regenerating liver to stimulate cell proliferation. In contrast, the lack of induction of Acly in regenerating PPARb livers failed to induce glycolysis and FA synthesis that resulted in no fat accumulation and delayed hepatocyte proliferation in PPARb mice.
E2fs regulate the expression of both proliferative and metabolic genes [35]. They are essential to regulate genes that are involved in DNA replication and cell cycle progression by exerting cellcycle-specific expression pattern and by binding directly to the E2f-binding sites [2]. The classical E2fs, which include E2f1-6, regulate the transcription of target genes when bound to the promoters as heterodimers with a ''differentiation regulated transcription factor protein'' (DP) while the atypical E2f7-8 bind to promoters as homodimers or heterodimers without a DP. The E2f family is split into two groups by function: transcription activators and repressors. Activators such as E2f1-3 promote cell cycle progression, while repressors (E2f4-8) inhibit cell cycle [36]. E2f7-8 can inhibit the action of E2f1 via a negative feedback loop [35]. Since some E2f family members may have overlapping functions, loss of one family member may be compensated by  another [35]. This may explain why E2f1 deficiency has no effect on liver regeneration [37]. Our data showed that among the eight E2fs studied, four (E2f1-2 and E2f7-8) increased their expression levels during liver regeneration. The induction E2f1 mRNA level was up more than 25 fold at 36 hr after PH. However, such inductions were either not noted or drastically reduced in KO mice. Moreover, thirty-eight downstream genes of E2fs involved in cell cycle regulation, DNA replication and repair, as well as checkpoint control, failed to be up-regulated in regenerating KO livers. These findings not only implied that PPARb has a role in coordinating the regulation of multiple E2fs, but also suggested that PPARb-mediated E2f actions might be essential for the normal progression of liver regeneration.
E2f1 is the best characterized member of the E2f family members. E2f1 orchestrates a complex control of oxidative and glycolytic metabolisms that are essential for cell proliferation and adaptation to energy demands [35]. Activation of E2f1 increases glucose-stimulated insulin secretion and favors the process of glycolysis in pancreatic b-Min6 cells [38]. Decreased insulin secretion was identified in E2f1-null mice [38]. Our data showed that about 70% of the genes involved in glycolysis and FA synthesis pathways failed to be induced during liver regeneration due to PPARb deficiency (48 hr after PH). These findings suggested that the role of PPARb in regulating liver metabolism and cell proliferation during liver regeneration might be through E2fs. Because of the significant role of E2fs in cell cycle and metabolism, it is important to understand the mechanism by which E2fs are regulated. Since Akt, the downstream of PPARb, activates E2f in NIH3T3 fibroblast cells and Akt phosphorylation correlates with increased E2F mRNA level [24,39], the regulation of E2f by PPARb might be Akt dependent. Furthermore, at the transcriptional level, by performing motif analysis of our published ChIP-sequencing data, we found that hepatic retinoid6receptor a (RXRa) bound to E2f1 and E2f2 (Chr2:154388649-154388661, Chr4:135737726-135737738, respectively) in mouse liver [40]. It was shown that the PPARc agonist rosiglitazone increases the binding of PPARc to DR1 sites in the E2f1 and E2f2 gene promoters in 3T3-L1 cells [41]. The RXRa binding site found in mouse liver coincides with the RXRa/PPARc binding sites found in 3T3-L1 cells. E2fs regulate adipocyte differentiation through modulating the expression of PPARc in 3T3-L1 preadipocytes [39]. Moreover, the binding of RXRa to the E2f1 and E2f2 in mouse livers could also be enriched by treating mice with all-trans retinoic acid (unpublished). These data indicate that E2f1-2 can be transcriptionally regulated by RXRa/PPARc. Whether E2f1-2 can be directly regulated by PPARb remains to be determined. It is worth noting that the basal level of E2f1-2 is not different between the WT and KO mice, the differential expressions of E2fs are only found when hepatocytes are actively proliferating, thus reflecting the cell cycle-specific expression pattern of E2fs. It is interesting to note that there is substantially hepatic lipid accumulation 36 hr after PH. The role of the accumulated lipid may function as an energy source for liver regeneration [42]. It is possible that the accumulated lipid in regenerating liver serves as a ligand for PPARc or b, which in turn activates E2f1-2 and lead to metabolism and cell proliferation. Thus, the transcriptional induction of E2fs could be due to activation of Akt or the FAactivated PPARc pathway. This scenario is further supported by our findings that PPARc is induced in regenerating WT mouse livers.
Taken together, our data showed the potential interaction between PPARb and E2fs in regulating metabolism and cell proliferation in regenerating mouse liver. The lack of activation of PDK1/Akt and E2fs-mediated pathways due to PPARb deficiency resulted in delayed regeneration (Fig. 8). It would be interesting to test whether activation of PPARb can facilitate liver regeneration and this work is currently on going.