The liver is one of the vital organs of the body with mutiple important responsibilities, including metabolism, maintenance of water balance, bile acid production and excretion, detoxification, immune response and so on [1]. In addition, the liver becomes distinguished from other organs mainly by its amazing regenerative ability, which is primarily attributable to the quick reentry of highly differentiated quiescent hepatic cells into the cell cycle in response to liver injury induced by surgical resection (e.g., surgerical resection, pathogenic or chemical factors [2]. Among various liver injury models, partial hepatectomy (PH), resulting in the loss of approximately 70% of the liver volume, is now widely utilized for studying liver regeneration in experimental animals. According to previous reports, many cytokines (e.g., TNF-α, IL-6) are upregulated within 0-6 h after 70% liver resection. DNA synthesis begins at 12-16 h following PH, peaking around at 24-48 h, which results in a steep increase of liver mass at 72 h after PH, followed by the rough restoration of liver mass to normal around at 168 h after PH [3]. In general, rat liver regeneration lasts about 7 days, during which involves many biological events including cell activation, cell differentiation, proliferation and its regulation, redifferentiation, reestablishment of liver structure and function as well, inferring that the mechanism underlying this process is highly complex [4]. Therefore, in order to understand clarify the regenerative process of the liver, it is very essential to clarify the overall and systematically molecular basis of this stringently regulated process.

So far, the high-throughput analysis technologies, such as genechip, serial analysis of gene expression (SAGE), have been extensively used by researchers to identify transcriptome profiles of the regenerating livers, and obtained the valuable data for understanding mechanism of LR. Especially in 2009, Wang et al. analyzed the gene expression profiling of regenerating rat livers at different recovery time points after PH, finding that there are totally 1004 known genes and 857 unknown genes associated with LR [5]. However, the data from microarray analysis is too limited to quantitatively analyze protein levels, or even to reflect the final biological effect of genes. So, despite that the gene expression pattern has come under intense scrutiny, a differential proteomic study could better help to elucidate how the process is triggered and regulated.

Recently, the emergence of comparative proteomics technologies, such as two-dimensional electrophoresis (2-DE), high performance liquid chromatography (HPLC), mass spectrometry (MS) and mass fingerprinting, provides a promising approach for elucidating the mechanism of LR [6]. As a protein separation technique, 2-DE has been widely employed in separating and quantifying proteins that are differentially expressed in proteomics research. Depending on this method, the proteins whose levels were significantly changed under a defined physiological condition can be screened out and identified. For instance, Strey and his coworkers applied two-dimensional gel electrophoresis approach to measure protein expression changes in mouse livers at 6 h and 12 h after PH, and identified twelve up-regulated (at least 2-fold) proteins related to signaling and metabolic pathways [7]. Nevertheless, this technology has some striking limitations such as low load ability, poor separation of hydrophobic, acidic and alkaline proteins. And HPLC or MS techniques overcome the shortcomings of 2-DE, and suit differential proteomics study much better. Also, other techniques such as multi-dimension chromatogram, protein chips have been utilized to complement or substitute this conventional method. Presently, the method of 2-DE separation combined with MS identification is now in common use in this research field. For example, He et al. introduced this method to detect protein expression profiles in rat regenerating livers at 1 hour after PH, and identified a total of 24 differentially expressed proteins. In addition, He’s research team applied 2-DE in combination with MALDI-QTOF-MS compared protein expression patterns between sham-operation group and 7-hour hepatectomized group to identify proteins whose expression may be altered after PH, and discovered 29 differentially expressed proteins [8]. In study by Hsieh et al., they used iTRAQ-coupled LC-MS/MS technique to perform a comparative analysis of protein expression profiling in mouse liver regenerating for 24, 48 and 72 h after PH, and found a total of 270 time-dependently differentially expressed proteins during the regenerative process [9]. Up to date, these technologies are often employed to study the early event occurring in the first 24 hours after surgery, failing to comprehensively display the dynamical changes in protein expressions during LR. For this reason, in this study, we assessed the temporal expression of proteins from the regenerating liver recovering 6, 12, 24, 72, 120 and 168 h following 2/3 PH in rats with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI TOF/TOF MS), thus establishing proteome altas of the normal liver and regenerating liver, which might lay the foundation for further screening out the key factors and cell markers associated with LR.

Studies have shown that, in certain range, the regenerative response is directly proportional to the mass of liver resected ([10] Liver regeneration. Michalopoulos GK, DeFrances MC. Science. 1997;276(5309):60-6]). Usually, the removal of approximately 75% of liver volume was tolerable for normal rodents, but the resection of more than 75% of liver mass often leads to liver failure and death. Accordingly, to induce maximal hepatic regeneration meanwhile maintain sufficient hepatic functionality meeting the essential body metabolic demand, 70% (or 2/3) partial hepatectomy now has been become the most widely used model to study the mechanism of liver regeneration. Using this animal model, a body of evidence has shown that many factors (e.g., HGF, NFκB, TNF and IL6) participate in the initiation of LR [11]. Also, based on this model, many genomic and proteomic studies have been carried out to map the complex dynamic expression changes of genes/proteins at the early phase of LR. However, the dynamical changes of protein expressions during the whole regeneration process have been rarely reported. So, the present study attempted to detect the overall dynamic proteome changes following PH in rats, gaining some insights into the nature of LR.

We categorized all the 220 liver regeneration-associated proteins into six groups. In group 1, 125 proteins were up-regulated post PH in contrast to SO group. In addition, in group 3, 5 proteins were down-regulated only in sham-operated rat livers but obviously unchanged in the hepatectomized liver, thus also sorted in the “up-regulation” pattern. Among these 130 proteins, 25 have been reported to be associated with liver regeneration according to the previous studies (Additional file 1: Table S1), and mainly play the role in lipid metabolism (ACADM, HMGCS2 and PRDX6), carbohydrate and energy metabolism (COX5a, GPD1, PDHA1 and UGP2), amino acid and nucleic acid metabolism (ALDH6a1, ARG1, CPS1, GOT1, HPD and UROD), differentiation and development (EEF1A1, EIF5A, GMFB, KRT6A, NDUFS2 and RGN), signal transmission (ARHGDIA and YWHAE), or inflammatory response (GSTM2 and PDIA3). In terms of the expression pattern, a substantial portion of these proteins began to be dramatically up-regulated at early phase. Particularily, the up-expression of metabolism-related proteins last throughout the entire regenerative process, consistent with Guo’s report that lipid and carbohydrate metabolism were significantly enhanced, which not only compensates metabolic homeostasis, but also can function as driver of the complex cascade of gene activation required for cell proliferation [8]. In addition, some studies suggested that, immediately after PH, inflammatory factors STAT3, TNFα and IL6 were activated to regulate the expression of a large number of genes involved in inflammation, acute phase response and cell proliferation. For example, Taub pointed out that, although an obvious inflammatory response was not seen in liver parenchyma at the initial phase of liver regeneration, there occurred the elevated levels of acute-phase proteins and the release of cytokines responsible for the regulation of inflammation [12]. In general, our findings were in accordance with his statement. However, the up-regulated inflammatory factors in our study were not STAT3, TNFα and IL-6, but GSTM2 and PDIA3 identified as LR-related proteins by other studies [12]. The reason for the discrepancy between ours and others might be attributable to the different experimental conditions, different example batches, different operational methods and so on. It is worthily emphasized that our study detected the an up-regulation (a peak of 2.2-fold at 120 h) of inflammation-related protein GSTM2 which was found to be down-regulated in the Cao’s study using 50% PH mouse model [13]. The difference between results from the two studies may be due to different experimental animals and operational procedures.

Partial hepatic resection in rats induced a number of mediators including inflammatory cytokines, growth factors and various hormones which can activate a complex network of signal transduction that promotes hepatic regeneration. As described previously, the signaling pathway mediated by IL6, TNFα and STAT3 is essential for triggering live regeneration. However, the expressions of these cytokines didn’t show the significant changes in this present study. Instead, signal transmission-related YWHAE, a LR-associated protein identified by Li et al.[14], was dramatically up-regulated by 7.6-fold at 6h post PH. According to the Ingenuity Pathway Analysis conducted by us, the networks from six timepoints after 2/3 hepatectomy all centered around YWHAE protein. Although Li et al. provided the evidence that YWHAE was related to regeneration process [14], before that, nobody realized its importance to hepatic regeneration. YWHAE belongs to the 14-3-3 family of proteins that can signal by binding to phosphoserine-containing proteins and plays a role in a wide variety of cellular functions. Besides functioning in cell division and regulation of insulin sensitivity, its other specific functions need to be further identified. And our study suggested the significance of this protein in LR, which might be the first observation of a new pathway critical to the regenerative process. Besides the newly-identified YWHAE-mediated pathway, cell division control protein 42 (CDC42) was considered as another important factor that directly or indirectly connected to LR termination-involved proteins including APC, CFL1, RUVBL1 and so on. Our data showed that the expression of this protein was 2-fold up-regulated at the termination of LR (including 120 h and 168 h). As a small GTPase of Rho-subfamily, CDC42 plays important roles in diverse cellular functions such as cell morphology, migration and cell cycle. Therefore, it could be concluded that CDC42 was key to the termination of LR. Consistent with this, some protein involved in the end of liver regeneration, such as APC, CFL1, also showed the two-fold up-regulation at the same period, which reinforced the above conclusion drawn by us. Whereas, Yuan and his colleagues used a mouse model with liver-specific knockout of CDC42 and studied the function of CDC42 in hepatic regeneration after PH. Their results showed that the loss of CDC42 led to a significant delay of liver recovery after PH, suggesting the important role of CDC42 in regulating proliferative response during LR [15]. By contrast, our study observed another action of CDC42 in hepatic recovery.

Among the remaining 100 up-regulated proteins, of particular note are GAPDH (up-regulation during the whole LR with a peak of 26.4-fold), the enzyme catalyzing an important energy-yielding step in carbohydrate metabolism. This protein was sorted into the YWHAE-mediated pathway according to our pathway analysis (Additional file 2: Figure S1). The elevated mRNA and protein levels of GAPDH at regenerating rat liver after 70% PH has been reported, which might result directly from the stimulating effect of insulin [16, 17]. As stated previously, the sensitivity of body tissues to insulin is stimulated by YWHAE-mediated pathway. Thus, the significant up-regulation of GAPDH in our study might possibly reflect the up-regulation of YWHAE and the enhancement of YWHAE-mediated pathway. Secondly, another emphasized protein was TUBA1C that was significantly up-expressed during the whole regeneration process, with a peak of 19.3 folds at 24 h following PH. This protein is one of subunits of tubulin which acts as the main component of the microtubule cytoskeleton and play a crucial role in regulating the mitotic spindle during cell division. In this study, alpha-tubulin protein level was detected to have a 1.3-23.0 fold increase in rat regenerating liver, which was agreeable with the Matos’s observations that the level of this cytoskeletal components in liver tissues was strikingly increased in patients with hepatocellular carcinoma [18]. According to above description, to a certain extent, there existed similarity in some respects between liver regeneration and liver cancer.

In group 2, 26 proteins were detected only in partially-hepatectomized rats instead of the sham-operation group, so called “newly-induced proteins”. The failure to detect those proteins in SO group could be due either to true absence of their expressions or to their very low quantity below the detection limit of our detection. Nearly half (12 out of 26) of the “newly-induced proteins” in response to 2/3 hepatectomy were related to cell differentiation and development. In recent decades, there has been an increasing interest in the relationship between liver regeneration and nervous system. Previous researches observed the axonal growth around the portal area in rat liver regeneration model [19]. Similarily, our results showed that, among the 12 cell differentiation and development-involved proteins, neuronal migration protein doublecortin (DCX), neuron-specific calcium-binding protein hippocalcin (HPCA) and heart- and neural crest derivatives-expressed protein 2 (HAND2) have been recognized as the important factors in nervous system development. DCX directed neuronal migration by regulating microtubule organization [20]. HPCA and HAND2 play the important roles in neural regeneration of nervous system [21]. Overall, all of them may be responsible for neuron growth in remnant liver after PH. Our finding was similar to the one reported by Sun et al. [19] where the new synthesis of neuromodulin and GMFB had respectively a 2.6- and 3.1- fold increase at 1h after PH, in despite of the difference between the differential proteins identified by us and that by Sun. In group 4 and group 5, proteins are down-regulated or below detection limit after treated with 2/3 PH, respectively. Among these down-regulated proteins, carbohydrate, lipid and energy metabolism-involved enzymes were particularly noteworthy.

Carbohydrate metabolism-involved enzymes including G6PD, CTSA, PGLS, PYGL, TKT and UQCRC2 showed the down-regulation, ranging from 1.2 to 7.7 fold decrease. Lipid metabolism enzymes or related proteins including ACAA2, ALOX15B, ECI1, FABP7, FABP9, IAH1 and LIPK were also decreased in expression level with a range from 1.1 to 5.8 fold reduction. Two classes of above proteins were mainly responsible for glucose catabolism or fatty acid oxidation. So, their down-regulation might lead to a reduced glucose and fatty acid degradation process. But combining with the expression profiles of the related proteins in group 1, the weakened metabolic process caused by the down-regulation of these metabolism related-proteins was not enough to offset the enhancement of these metabolism activities induced by the up-regulation of a majority of glucose and fatty acid degradation-involved proteins. It has been widely accepted that energy production in the body mainly derives from the catabolism of carbohydrate and fatty acid. Therefore, according to our analysis it could be proposed that, to some extent, cellular ATP production would be increased. However, Cao et al. showed that some key enzymes involved in citric acid cycle and electron transport chain were down-regulated or below detection limit, implying that ATP production would be hampered [13]. Crumm et al. also reported that ATP level in the remaining livers after 30% PH was remarkably decreased and remained significantly lower than SO samples between 24-48 h post PH [22]. Clearly, our conclusion was quite opposite to that of Cao et al. and Crumm et al., and the reason remains unclear and needs to be answered by means of further studies.

According to expression and functional analysis of differentially expressed proteins, the metabolic processes of lipid and carbohydrate were significantly promoted at early phase of LR. Ingenuity Pathway Analysis suggested that YWHAE protein was located mainly in the center of the networks at all the six recovery timepoints after 2/3 hepatectomy, considering this protein as one of the master factors regulating liver regeneration under our experimental conditions. And its exact role in liver regeneration needs to be further explored using other methods, such as gene addition, RNA interference and so on. Additionally, our study also showed the implication of CDC42 in the termination of LR through mediating the specific pathway. All in all, our work may be helpful in expanding our understanding of liver regeneration and hepatic response in acute liver diseases, and it might ultimately lead to the emergence of novel therapeutic strategies to accelerate liver regeneration after hepatic injury.