The liver has the unique capacity to regenerate after surgical resection. However, the regulation of liver regeneration is not completely understood. Recent reports indicate an essential role for small noncoding microRNAs (miRNAs) in the regulation of hepatic development, carcinogenesis, and early regeneration. We hypothesized that miRNAs are critically involved in all phases of liver regeneration after partial hepatectomy. We performed miRNA microarray analyses after 70% partial hepatectomy in rats under isoflurane anesthesia at different time points (0 h to 5 days) and after sham laparotomy. Putative targets of differentially expressed miRNAs were determined using a bioinformatic approach. Two-dimensional (2D)-PAGE proteomic analyses and protein identification were performed on specimens at 0 and 24 h after resection. The temporal dynamics of liver regeneration were characterized by 5-bromo- 2-deoxyuridine, proliferating cell nuclear antigen, IL-6, and hepatocyte growth factor. We demonstrate that miRNA expression patterns changed during liver regeneration and that these changes were most evident during the peak of DNA replication at 24 h after resection. Expression of 13 miRNAs was significantly reduced 12–48 h after resection (>25% change), out of which downreguation was confirmed in isolated hepatocytes for 6 miRNAs at 24 h, whereas three miRNAs were significantly upregulated. Proteomic analysis revealed 65 upregulated proteins; among them, 23 represent putative targets of the differentially expressed miRNAs. We provide a temporal miRNA expression and proteomic dataset of the regenerating rat liver, which indicates a primary function for miRNA during the peak of DNA replication. These data will assist further functional studies on the role of miRNAs during liver regeneration.
- liver regeneration
- proteomic analysis
partial hepatectomy (PH) is a commonly performed procedure to treat primary and secondary hepatic malignancies (13, 16). After PH, the lost functional mass is replaced in a process of compensatory growth in which quiescent hepatocytes reenter the cell cycle (21). Liver regeneration is a very complex and well-orchestrated phenomenon. It can schematically be represented as consisting of three main steps: initiation, proliferation, and metabolic adaption (7). All of these processes are triggered and modulated by specific growth factors [e.g., hepatocyte growth factor (HGF)], cytokines (e.g., IL-6), matrix remodeling, and several stimulation and inhibition feedback loops. However, regulation of the initiation, progression, and termination stages of liver regeneration is not completely understood, and inducing this process, therapeutically remains difficult (14, 22).
Within the last decade, it has become evident that posttranscriptional regulation of gene expression by small ribonucleic acids (RNAs) is a central component of the cellular gene regulatory network. MicroRNAs (miRNAs) are the most abundant class of small, endogenous noncoding RNAs. miRNAs are single-stranded RNAs of ∼22 nucleotides in length that act as posttranscriptional regulators. miRNAs inhibit protein synthesis by blocking translation via complementary binding of messenger RNA (mRNA) or by suppressing translation and the subsequent degradation of target mRNAs (3). miRNAs act in a variety of cellular processes such as development, organ homeostasis, and cancer, and a single miRNA is potentially able to target up to several hundred mRNAs (18). Previous studies have shown that specific signatures of differentially expressed miRNAs can be assigned to either hepatic development (30) or carcinogenesis (23). In both cases, specific miRNAs affect target genes that promote cell proliferation and replication.
We hypothesized that miRNAs may also mediate cell cycle arrest in the adult liver. Subsequently, a temporal activation of those pathways known to be involved in both liver development and carcinogenesis may contribute to liver regeneration after PH. In this regard, Marquez et al. (20) reported on the course of miR-21 during murine liver regeneration, which had been shown to target specific tumor suppressor genes. However, data on miRNA expression signatures and their putative roles during liver regeneration are still sparse. In fact, temporal changes in miRNA expression profiles after PH have thus far been assessed in a recent murine model by Song et al., (28) who showed that mice exhibited differential expression of a subset of miRNAs in the early phase of regeneration after two-thirds PH. Nevertheless, the complex regulatory networks operated by miRNAs during all phases of hepatic regeneration are still poorly understood.
In the present study, we performed microarray analyses of miRNA expression in rats after 70% PH to elucidate the temporal pattern of miRNA changes during all phases of liver regeneration. Additionally, we identified putative targets of the differentially expressed miRNAs using a bioinformatic approach. Subsequently, we collected proteomic data by performing two-dimensional (2D)-PAGE and protein identification. We identified a novel set of miRNAs that play a role during hepatic regeneration. Our study provides a comprehensive dataset for further functional and therapeutic studies geared toward the elucidation of the complex molecular networks involved in miRNA-based control of liver regeneration.
MATERIALS AND METHODS
Animal Studies and Evaluation of Liver Regeneration
All experiments were performed in accordance with federal law and were approved by the federal authorities for animal research (license no. G 0154/09). A 70% PH was performed on male Wistar rats (270–310 g) under isoflorane inhalative anesthesia (2.5%) according to the procedure by Higgins (11). Liver tissue specimens and blood samples (n = 3 animals) were taken from untreated animals (0 h) and 2 h, 6 h, 12 h, 24 h, 48 h, and 5 days after PH. Sham laparotomy controls were performed at 12, 24, and 48 h. Surgery (PH, sham laparotomy, and harvesting of normal liver) was performed between 10 AM and 2 PM to avoid differences related to circadian rhythms in miRNA expression. All animals received 50 mg/kg body wt 5-bromo-2-deoxyuridine (BrdU; Sigma-Aldrich, St. Louis, MO) intraperitoneally 1 h before tissue harvesting.
Primary rat hepatocytes.
Primary rat hepatocytes were isolated from untreated animals and 24 h after PH (n = 3 animals) with a modification of the two-step perfusion protocol by Seglen (26). Briefly, livers were flushed with Leffert's buffer containing EGTA via the portal vein prior to digestion with collagenase in CaCl buffer. Hepatocytes were purified via density gradient centrifugation (50% percoll; Biochrom, Berlin, Germany).
Sections (2 μm) of formalin-fixed liver tissue were deparaffinized and stained with mouse monoclonal anti-BrdU antibody or mouse monoclonal anti-proliferating cell nuclear antigen (PCNA) antibody (Dako, Hamburg, Germany). Polyclonal goat anti-mouse IgG (Dako) and Fuchsin or 3,39-diaminobenzidine were used for visualization, and sections were counterstained with hematoxylin. Slides were analyzed with a combined light and fluorescence microscope (Axio Imager M1; Zeiss, Jena, Germany), and images were processed using AxioVision LE software (Zeiss).
EDTA plasma was used for quantification of IL-6 and HGF concentrations using commercially available ELISA kits in accordance with the manufacturers' instructions (Pierce Biotechnology, Rockford, IL; Institute of Immunology, Tokyo, Japan).
miRNA Screening and Validation
Total RNA, including the miRNA fraction, was isolated using the miRNeasy Mini Kit (QIAGEN, Valencia, CA). RNA quality was assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA).
miRNA microarray screening.
Samples were analyzed with a Geniom Realtime Analyzer (febit, Heidelberg, Germany) using the Geniom Biochip MPEA Rattus norvegicus (Sanger miRBase version 14.0 September 2009) from febit. In brief, hybridization was performed automatically for 16 h at 42°C and a microfluidic-based primer extension assay that utilizes bound miRNAs as primers for enzymatic elongation with biotinylated nucleotides was applied. Streptavidin-phycoerythrin was used in combination with a Consecutive Signal Enhancement procedure. The microarray data have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession no. GSE23696 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE23696).
A representative heat map containing 85 of the 323 investigated miRNA with the highest overall variability was generated (see Fig. 2). Red and green indicate up- and downregulation, respectively. Principal component analysis of all investigated samples was performed to investigate clustering of all three samples at each time point.
Real-time quantitative PCR analysis for miRNA.
cDNA was synthesized from 10 ng of RNA sample using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). PCR was performed with TaqMan microRNA Assays (Applied Biosystems, Life Technologies). Rno-U6 snRNA served as the endogenous control.
Samples (n = 3) from untreated animals (0 h) and 24 h after PH were used for proteomic analysis. Sample preparation and 2D-PAGE were performed according to Proteome Factory's (Berlin, Germany) 2D-PAGE sample preparation protocol for tissues. Briefly, 70 μg of protein was applied to vertical rod gels for isoelectric focusing at 8,820 Vh in the first dimension. The isoelectric focusing gels were then incubated in an equilibration buffer containing 125 mM trisphosphate (pH 6.8), 40% glycerol, 65 mM DTT, and 3% SDS for 10 min and subsequently frozen at −80°C. After being thawed, the equilibrated isoelectric focusing gels were immediately applied to SDS-PAGE gels. After 2D-PAGE separation, gels were stained with FireSilver (PS-2001, Proteome Factory).
For comparative analysis, the gels were digitized, and 2D image analysis was performed using the Proteomweaver software 3.1 (Definiens, Munich, Germany). Gels were normalized using the standard Proteomweaver algorithm, spots were automatically detected, and data were obtained by matched analysis. All gels were then compared with each other to create a group of superspots. During this step, mismatches were carefully considered and manually edited. Spot intensities were normalized using the pair-match-based normalization algorithm of the Proteomweaver software. Statistical evaluation was performed according to the following criteria: statistical significance for P < 0.05 (Student's t-test), and the minimal significant factor is 1.95.
Protein identification using nanoLC-ESI-MS/MS was performed by Proteome Factory. The mass spectrometry (MS) system consists of an Agilent 1100 NanoLC system (Agilent, Germany), a PicoTip emitter (New Objective, Woburn, MA), and an Esquire 3000 Plus ion-trap MS (Bruker, Bremen, Germany). Protein spots were in-gel digested by trypsin (Promega, Mannheim, Germany) and applied to nanoLC-ESI-MS/MS. After trapping and desalting the peptides on enrichment columns (Zorbax SB C18, 0.3 × 5 mm, Agilent) using 1% acetonitrile/0.1% formic acid solution for 5 min, peptides were separated on Zorbax 300 SB C18, 75 μm × 150 mm columns (Agilent) using an acetonitrile/0.1% formic acid gradient from 5% to 40% acetonitrile within 40 min. MS spectra were automatically taken by the Esquire 3000 Plus according to the manufacturer's instrument settings for nanoLS-ESI-MSMS analyses. Proteins were identified by an MS/MS ion search of the Mascot search engine (Matrix Science, London, UK) and the nr protein database (National Centre for Biotechnology Information, Bethesda, MD).
All data are expressed as means ± SD. Statistical analyses (Empirical Bayes Statistics adjusted for multiple testing and Student's t-test) were performed using Bioconductor open source software. A P value of < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001).
We performed miRNA profiling in a rat model of 70% PH that is known to be highly reproducible and commonly applied in studies on liver regeneration (21). We investigated the period between 0 h and 5 days after liver resection to correlate miRNA expression changes with all phases of liver regeneration. We investigated BrdU incorporation and PCNA expression to quantify the temporal pattern of liver regeneration. In our model, DNA synthesis peaked 24 h after resection. The increase in the numbers of BrdU positive cells started at 12 h (5.9 ± 2.3 positive cells/mm2), peaked at 24 h (558.9 ± 176.9 positive cells/mm2), and declined thereafter (Fig. 1A), whereas PCNA expression remained elevated until the 48-h time point (Fig. 1B). To determine the levels of cytokine production during liver regeneration, we investigated HGF and IL-6 levels in blood samples throughout the entire study period. HGF levels peaked as early as 2 h after resection (Fig. 1C) and the peak of IL-6 concentration followed at 6 h (Fig. 1D).
miRNA Expression Changes Correlated with Liver Regeneration in a Distinct Temporal Pattern
To analyze global miRNA expression during rat liver regeneration, we investigated the expression pattern of 323 miRNAs at different time points after PH using miRNA microarrays. Figure 2A displays an overview of these data in the form of a representative heat map and a principal component analysis of all investigated samples (Fig. 2B), which shows that the three samples at each time point clustered well together. We confirmed the reliability of the microarray platform by measuring the expression of five miRNAs at two time points using quantitative real-time PCR (qRT-PCR). There was a strong correlation between the microarray and qRT-PCR measurements in the investigated samples (Fig. 2C).
In our microarray analyses, we found 29 miRNAs differentially expressed during rat liver regeneration (P < 0.05; >20% change compared with normal liver), which suggests that multiple miRNAs are involved in liver regeneration in the rat model. These miRNA were differentially expressed at 12, 24, and 48 h compared with sham controls (P < 0.05). Among them, 22 miRNAs were significantly downregulated (Table 1), whereas seven miRNAs were upregulated (Table 2). Out of the 22 miRNAs that were downregulated, 13 miRNAs showed a temporal decrease of > 25% compared with controls. These miRNAs included five members of the let-7 family, which is known to be strongly involved in cell cycle control (31). Potential targets of the regulated miRNAs are summarized in Tables 1 and 2. Interestingly, the temporal pattern of miRNA expression changes correlated with the progress of liver regeneration and peaked 24 h after resection (Fig. 3). Among all downregulated miRNAs, only miR-352 showed a significant downregulation as early as 2 h after resection. At 12 h, 23% of the downregulated miRNAs reached significance compared with normal liver, whereas all downregulated miRNA showed significant downregulation at 24 h. Only four miRNAs remained downregulated at the 48-h mark, whereas none remained downregulated at day 5. The upregulated miRNAs showed similar temporal expression patterns (Table 2). Whereas none of the changes reached statistical significance in the early period of the regeneration, all significantly upregulated miRNAs peaked at 24 h. Interestingly, only three miRNAs (miR-33, miR-153, and miR-743b) were > 25 % upregulated at this time point.
To test whether downregulation at 24 h was attributed to changes in miRNA expression in hepatocytes rather than other hepatic cells types, nine miRNAs, which were downregulated by > 25 % and miR-21 were investigated in isolated hepatocytes. Among them, six miRNAs were significantly downregulated in isolated hepatocytes from regenerating livers (Fig. 2D).
The differentially expressed miRNAs (>25% change) were assigned to three different groups: 1) the early-downregulated group consisting of miR-352, which was significantly downregulated in the early phase of liver regeneration; 2) the late-downregulated group consisting of let-7b, let-7c, let-7e, let-7f, let-7i, miR-26a, miR-125b-5p, miR-126, miR-207, miR-223, miR-352, miR-375, and miR-872, which were downregulated during the peak of liver regeneration; and 3) the upregulated group contained miRNAs that were expressed at higher levels during the peak of DNA replication (miR-33, miR-153, and miR-743b). These results demonstrate a unique miRNA expression profile during liver regeneration in the rat model.
Upregulated Proteins Match with Target Prediction of Differentially Expressed miRNAs During the Peak of DNA Replication
Based on our finding that miRNA expression changes peaked during the proliferative phase of rat liver regeneration, we performed global proteomic analysis of normal (Control group) and proliferating liver tissue 24 h after resection. High throughput image analysis revealed 27 spots that were significantly upregulated 24 h after resection, whereas 12 spots were downregulated (Fig. 4). We identified proteins in all identified spots at 24 h by mass spectrometry and an MS/MS ion database search. Within the upregulated spots, we identified 65 proteins that were mainly associated with cell metabolism (Table 3). Out of these, 23 proteins were identified as putative targets of 17 miRNAs that were significantly altered during liver regeneration (Table 1). These data indicate that miRNAs could possibly contribute to the regulation of 23 of 65 of the upregulated proteins. At the same time, only three proteins [transthyretin (Ttr), glutathione S-transferase mu 1 (Gstm1), and chaperone, ABC1 activity of bc1 complex homolog (S. pombe) (Cabc1)] were found in the 12 downregulated spots. None of them could be correlated to putative targets of the upregulated miRNAs.
In this work, we report extensive genome-wide profiling of miRNA expression changes through all phases of rat liver regeneration after PH. The expression of 16 miRNAs changed significantly (i.e., >25%) in a specific temporal pattern that correlated with the progression of liver regeneration. Our data suggest a regulatory network based on an orchestra of multiple miRNAs that likely contributes to the regulation of liver regeneration. This may be of particular importance during the peak of regeneration, 24 h after PH. In addition, our proteomic analysis revealed 65 proteins that were found in upregulated protein spots at this time point. Interestingly, 25 of these proteins are associated with cell and energy metabolism. Therefore, our findings are in line with a previous report that showed extensive upregulation of proteins involved in cellular energy metabolism in mice 24 h after 50% partial hepatectomy (2). Although the proteins identified in our study are involved in the same metabolic pathways described by Cao et al. (2), none of them [except for Cps1 (2), Acadl (6) and Mat1a (12), which has been shown to be critical for the hepatic response to mitogenic signals (4)] previously have been specifically identified as differentially expressed following PH.
As a next step, we were interested in matching the identified proteins with the differentially expressed miRNAs to assess the potential role of miRNAs in the fine-tuning of liver regeneration. Therefore, we performed putative target analysis focusing on all miRNAs that were significantly altered by at least 20% at 24 h (Table 3). Interestingly, our analysis showed that about one-third of the upregulated proteins identified during the peak of regeneration were potential targets of the downregulated miRNAs. This is in line with the general assumption that as many as 30% of protein-coding genes are controlled by miRNAs (17). We are aware of the fact that although our data indicates well-orchestrated regulation by miRNAs, miRNA regulation of each target should be proven in further individual in vitro and in vivo studies, since bioinformatic target prediction does not necessarily predict biologically relevant regulation. Moreover, mRNA changes during liver regeneration should be investigated and correlated to changes in miRNA expression and protein changes. However, our dataset may assist these studies and our grouping of miRNAs that were found to be regulated during liver regeneration into early (<12 h) and late regulated miRNAs (12–48 h), as well as upregulated miRNAs may give indications for direction of further functional studies.
In general, the role of miRNAs during liver regeneration is incompletely understood. Hand et al. (10) showed that miRNA deficiency in adult mice resulted in an increased turnover and regeneration of hepatocytes. Thereafter, Marquez et al. (20) investigated the course of miR-21 in a mouse liver regeneration model. miR-21 had previously been shown to target specific tumor suppressor genes, such as Pdcd4 (1) and Tpm1 (32), and promote cell cycle progression. Interestingly, miR-21 is upregulated during the early stages of liver regeneration, and it targets Pellino 1, an ubiquitin ligase involved in activation of NF-kB signaling (20). In addition, Song et al. (28) recently showed that miRNAs are critical regulators of hepatocyte proliferation during the early phase of liver regeneration. They performed microarray analysis up to 18 h after liver resection in mice. Of note in mice, the regenerative peak after PH usually occurs 36 h after resection (21). The study by Song et al. (28) revealed differential expression of a subset of miRNAs, notably the induction of miR-21 and repression of miR-378. For both miRNAs, they identified Btg2 and ODC1 as novel targets, which are relevant for DNA synthesis in hepatocytes after liver resection (24, 29). These previous studies all show that specific miRNAs are crucial to induce liver regeneration and promote cell proliferation, but they did not consider the regenerative peak after PH. In our rat model, the regenerative peak occurs 24 h after resection. This may explain why we found a specific subset of miRNAs that has not been previously identified in models of liver regeneration. In contrast to Marquez et al. (20) we did not find upregulation of miR-21 (Fig. 3). Indeed, we found downregulation of miR-21 24 h after PH, but the expression change of miR-21 observed in our study was not below our 20% cut-off (Fig. 3), not even in isolated hepatocytes (Fig. 2D). One factor that might account for these differences, as we mentioned, is the use of different time points for the analysis. Although miRNAs and their regulatory networks are often conserved among species, we cannot exclude the possibility that species or even strain specificity are somehow responsible for these contrary findings (19). In this regard, further studies of human surgical specimens after liver regeneration, although difficult to obtain, are highly warranted to verify the cumulative data provided by the experimental studies. Moreover, comparative studies with rat and mouse samples at multiple time points during liver regeneration may be required to clarify the discrepancy between our findings (especially regarding miR-21) and the findings of Marguez et al. (20) and Song et al. (28). Since only a relatively mild effect of complete miRNA deficiency on regeneration after PH was observed by Song et al., the role of miRNA during liver regeneration is possibly described more accurately as being a moderator or conductor rather than a regulator or initiator. This concept corresponds to our observation that not a single miRNA but an orchestra of miRNAs is regulated during the peak of liver regeneration.
Of note, we performed our microarray analyses on whole liver specimens but not on purified hepatocytes. It is therefore likely that the relatively low degree of change (down to 46%) observed in miRNA expression after hepatectomy is caused by a significant contribution from nonparenchymal cells, which constitute ∼30% of all liver cells (27). However, our analysis of miRNA changes in isolated hepatocytes confirmed downregulation of six miRNA that were found to be downregulated in the microarray analysis (Fig. 3). Of note, we investigated all miRNAs shown in Fig. 3, except for let-7b, let-7c, let-7f, and let-7i. We could not confirm downregulation of miR-233, miR-352, and miR-872 at 24 h. This may be due to regulation of these miRNAs in other cell types than hepatocytes and remains to be investigated. Interestingly, as expected, downregulation in isolated hepatocytes was stronger than in whole liver specimens [e.g., miR-126: whole liver 75% (microarray) and hepatocytes 31% (qRT-PCR)]. Nevertheless, as mentioned by Hand et al. (10), some miRNAs might still be missed in our study since we did not perform microarrays with isolated hepatocytes.
In contrast to liver regeneration, the role of miRNAs during liver development and carcinogenesis has been more intensely studied (8, 9, 25). In both hepatic development and carcinogenesis, miRNAs have been shown to play a specific role in cell cycle regulation and progression. Interestingly, these regulatory circuits are also likely to be involved in adult liver regeneration when normally quiescent hepatocytes reenter the cell cycle (3). Specifically, let-7 family members, which repress cell cycle progression (31), are upregulated in adult (as compared with fetal) livers, which indicates a higher proliferative activity in fetal livers (30). Furthermore, miR-26a exhibits tumor-suppressive activity by targeting cyclin-D2/E2, and its loss contributes to hepatic carcinogenesis (15). In our study, we found downregulation of seven let-7 family members and miR-26a. Therefore, these miRNAs are very likely to contribute to cell cycle control during liver regeneration. Of note, downregulation of these miRNAs follows a specific temporal pattern, which indicates a well-defined temporal regulation of miRNA expression changes during liver regeneration. Further investigations on the physiological mechanisms that control miRNA expression after PH and on the role of single miRNAs in the process of liver regeneration are necessary.
Perspectives and Significances
Our study provides a dataset of the temporal miRNA expression in the regenerating rat liver after PH. We demonstrate that the expression of several miRNAs changes during liver regeneration and that these changes are most evident during the regenerative peak after liver resection. Altered miRNAs display specific temporal patterns that suggest a primary physiological function of a subset of miRNAs in the fine-tuning during each period of liver regeneration. Thus, our dataset might assist further studies on the contribution of miRNAs to novel miRNA-based therapeutic strategies aimed at modulating the miRNA-based control of liver regeneration (5).
This study was funded by the European Regional Development Fund (EFRE 10144342).
No conflicts of interest, financial or otherwise, are declared by the author(s).
The authors thank Annekatrin Leder, Steffen Lippert, Kerstin Nehls, Natalie Schlüter, and Barbara Mitko for expert technical assistance, Nora N. Kammer for performing surgical procedures, and Dr. Mehmet H. Morgul (Visceral, Transplantation, Thorax, and Vascular Surgery, Universitätsklinikum Leipzig, Germany) and Andreas Berger (Department of Internal Medicine, Martin-Luther-Universität Halle-Wittenberg, Germany) for critical reading of the article.
- Copyright © 2011 the American Physiological Society