Brought to you by Portland Press Ltd.
Published on behalf of the International Federation for Cell Biology
Cancer Cell death Cell cycle Cytoskeleton Exo/endocytosis Differentiation Division Organelles Signalling Stem cells Trafficking
Cell Biology International (2012) 36, 721–732 (Printed in Great Britain)
Characterization of transcriptional profiling of Kupffer cells during liver regeneration in rats
Cunshuan Xu*†1, Xiaoguang Chen†, Cuifang Chang†, Gaiping Wang*, Wenbo Wang*, Lianxing Zhang*, Qiushi Zhu* and Lei Wang*
*College of Life Science, Henan Normal University, Xinxiang 453007, Peoples Republic of China, and †Key Laboratory for Cell Differentiation Regulation, Henan Normal University, Xinxiang 453007, Peoples Republic of China


KCs (Kupffer cells), as an important hepatic immunoregulatory cells, play a key role in LR (liver regeneration). Uncovering the transcriptional profiling of KCs after PH (partial hepatectomy) would likely clarify its implication in LR. Here, we isolated KCs by methods of Percoll density gradient centrifugation and immunomagnetic beads. Transcriptional profiles of KCs were monitored up to 168 h post-PH using microarray. By comparing the expression profile of KCs at 2–168 h post-PH with that of the control and applying the statistical and bioinformatics criteria, we found 1407 known and 927 unknown genes related to LR. K-means clustering analysis grouped these 1407 genes into robust 14 time-course clusters representing distinct patterns of regulation. Based on gene-set enrichment analysis, genes encoding products involved in cytokine signalling, inflammatory response and cell chaemotaxis were highly enriched in the cluster characterized by gradual up-regulation and then return; genes in defence response and immune response were enriched in clusters ‘the general down-regulation during LR’; genes in fatty acid synthesis and sterol metabolism were preferentially distributed in the cluster ‘gradual increase’; whereas genes in the categories ‘lipid catabolism’ and ‘glycolysis’ were enriched in cluster ‘decrease at two intervals’. According to the above analysis, KCs were seemingly sensitive to operative stimulus; immune defence and detoxification function of KCs obviously dropped post-operatively; fatty acid synthesis were enhanced, whereas lipid catabolism and glycolysis were reduced after PH. This study provides a detailed in vivo gene expression profile of KCs, providing a framework to better understand the molecular mechanisms underlying the regeneration process at cellular level.


Key words: Kupffer cells, liver regeneration, partial hepatectomy, rats, transcriptional profiling

Abbreviations: DTT, dithiothreitol, IEF, isoelectric focusing, IL, interleukin, IPG, immobilized pH gradient, KC, Kupffer cell, LR, liver regeneration, NPC, non-parenchymal cell, PH, partial hepatectomy, RT–PCR, reverse transcription–PCR, SEC, sinusoidal endothelial cell, SO, sham-operated, TC, temporal cluster, TGFβ, transforming growth factor β, TNFα, tumour necrosis factor α

1To whom correspondence should be addressed (email cellkeylab@126.com).


1. Introduction

LR (liver regeneration) is an orchestrated, multi-step process in response to the loss of hepatic mass, during which the residual lobes enlarge to compensate for the missing liver mass (Michalopoulos, 2010). The regeneration process can be divided into three phases including initiation, growth promotion and growth inhibition, involving the sequential changes in gene expression, multiple cellular processes, complex interaction with cytokines and growth factors, and morphological rebuilding (Dong et al., 2007). PHx can stimulate the initiation of regeneration within a short time (Takeishi et al., 1999). Several cytokines play an important role in this event (Taub, 2004). More specifically, in initial phase, the priming factors [e.g. IL-6 (interleukin-6) and TNFα (tumour necrosis factor α)] sensibilize hepatocytes to respond to stimulations of growth factors (Su et al., 2002). In growth-promoting phase growth factors [e.g. HGF (human growth factor) and EGF (epidermal growth factor)] stimulate hepatocytes to enter the cell cycle and proliferate (Desbois-Mouthon et al., 2006). In inhibition phase, the signals terminating cell growth, such as TGFβ (transforming growth factor β), are responsible for the decrease of DNA synthesis, further repressing hepatocyte proliferation (Stoick-Cooper et al., 2007). The evidences from many studies have demonstrated that the abnormalities of these factors would lead to the impaired or delayed regeneration (Böhm et al., 2010). Apart from hepatocytes (making up approximately two-third of the total liver cells), the liver comprise many other cell types, such as SECs (sinusoidal endothelial cells), HSCs (hepatic stellate cells) and KCs (Kupffer cells). Among these cells, KCs have been well established to be of importance to the regeneration process (Meijer et al., 2000; Kiiasov and Gumerova, 2002). In contrast with the comprehensive understanding on the impacts of growth factors or inhibitors on LR, the investigations about the relevance of KCs with LR were very limited in recent years. Based on the data currently available, KCs are considered to be crucial in LR because of the strong ability of this immune cell to produce growth factors and cytokines contributing to LR (Amemiya et al., 2011). KCs have the potential to exert either stimulatory or inhibitory influences on LR depending on these cytokines generated by itself (Malik et al., 2002). Roberts et al. (2007) pointed that KCs are the most predominant cell synthesizing TNFα, as one of the master regulators in the priming of LR. KCs also had negative effect on liver repair by releasing inhibitory factors IL-1α/β, TGFβ and so on (Boulton et al., 1997). The relationship between KC and LR appears to be very complex. For this purpose, as early as in 1999, Takeishi et al. (1999) used macrophage depletion method to investigate KC repopulation and function during LR, attempting to elucidate the role of this hepatic macrophage in LR. However, they just could observe the influence of KCs on LR through tracking the kinetics of this cell, but could not elucidate the relevance of KC with LR at molecular level. In the present study, we performed a PH (partial hepatectomy) in rats, separated KCs from the regenerating liver after PH using Percoll density centrifugation and immunomagnetic beads sorting (Wang and Xu, 2010), and examined the transcriptional profiles of KC post-PH using Rat Genome 230 2.0 Array (Wang et al., 2009).

2. Materials and methods

2.1. Preparation of rat PH model

Healthy SD (Sprague–Dawley) rats, aged 10–12 weeks and weighing 190±20 g, were bred in the Experimental Animal Center of Henan Normal University under normal conditions. A total of 114 rats were randomly divided into 19 groups with 6 rats each: 9 PH groups, 9 SO (sham-operated) groups and 1 control group. After ether anaesthesia, rats in PH groups were subjected to two-thirds liver resection according to the method described by Higgins and Anderson (1931). The rats after PH were allowed ad libitum access to food and water for another 2, 6, 12, 24, 30, 36, 72, 120 and 168 h until they received hepatic perfusion. Rats in SO groups were treated as above, but livers were not excised. Six rats in control group, as the 0-h samples for SO and PH groups, received the perfusion immediately after the surgical removal of left and median lobes. The experiments strictly observed the Regulations for the Administration of Affairs Concerning Experimental Animals in China.

2.2. Isolation of KCs

After the treated rats were anaesthetized by ether and sterilized with 75% alcohol, the abdominal cavity was opened to expose the portal vein, and inferior vena cava both below and above the liver were ligated after hepatic portal vein catheterization. Single cell suspensions of liver cells were prepared according to the following protocol. First, the liver was perfusion-fixed using D-Hank's solution at 37°C at 1–2 ml/min. When the liver surface turned white, perfusion continued with 0.05% collagenase solution at 1 ml/min. Then, the perfused livers from six rats were collected together and cut into pieces and digested with 0.05% IV-type collagenase at 37°C for another 15 min. The filtration is done using a 200-mesh net, and the filtered solutions were centrifuged thrice at 80 g for 3 min, and the resulting precipitate was harvested and washed twice with PBS buffer at 4°C. Cell concentration was adjusted to 1×108 cells/ml. Mixed-cell suspension of 6 ml were spread on the surface of 4 ml 60% Percoll to centrifuge at 200 g at 4°C for 5×2 min. The supernatant, dominated by KCs-enriched NPCs (non-parenchymal cells), was collected by centrifugation at 200 g for 2×2 min. NPCs re-suspended in an equal volume of PBS were layered on the Percoll solution (Pharmacia) and again were centrifuged at 300 g for 2 min. The lower layer, i.e. the high-density NPCs fraction, was diluted with PBS buffer and centrifuged at 300 g for 5 min to pellet the cells. The harvested sediment was the pellet enriched with KCs. The KCs-enriched fraction was mixed with 10 µl/ml rat anti-CD68 antibody, and then incubated with 10 µl/ml rat anti-PE magnetic beads for 15 min at 4°C. Cell suspension was loaded on to the separation column and allowed to flow naturally. Separation column was washed twice in PBS buffer at 4°C; the harvested fractions after removal of the magnetic field were the suspension for KCs.

2.3. Immunochemical assay

A few liver cell suspension and purified KCs were taken and individually fixed with 10% formaldehyde for 30 min were smeared on to a glass slide. When the cell suspension on glass slides was dry, microwave antigen retrieval was carried out. The sections were incubated separately with a 1:2000 dilution (v/v) of ED2 and LYZ antibodies overnight at 4°C, then the 1:5000 (v/v) diluted biotin-labelled secondary antibody was added to incubate for 60 min at 37°C. The system was hybridized with SABC (streptavidin–biotin complex) at 37°C for 30 min. Meanwhile, the paraffin sections of liver tissue at the corresponding time points were prepared for immunohistochemical staining. The results were observed under optical microscope (Gehring et al., 2009).

2.4. Microarray analysis

The total RNAs were extracted from KCs pooled from six animals per group according to the manual of TRIzol® reagent (Invitrogen Corporation) (Norton, 1992). RNA quantity and quality were assessed by agarose electrophoresis and spectrophotometric analysis prior to cDNA synthesis (Scott, 1999). The amplification and biotinylation of probes were done according to Affymetrix recommendations for microarray analysis. Samples were hybridized to Rat Genome 230 2.0 microarray. Once microarray was hybridized, washed and stained, and scanned following the manufacturer's protocol (McClintick et al., 2003; Edenberg et al., 2005). The detection and quantification of hybridization were done using GeneChip scanner 3000 (Affymetrix Inc.) (Kube et al., 2007). All arrays were assessed for ‘array performance’ prior to data analysis. Control transcripts were spiked into the hybridization mixture to control for hybridization efficiency and sensitivity. To minimize potential systematic errors, all stages of the experiment were balanced across experimental groups, i.e. the animals in each group were equal in number within the same time point, and equal numbers of RNA preparations from the representative groups were processed through labelling, hybridizing, washing and scanning protocols on a given time.

2.5. Chip data analysis

The data for each microarray were initially normalized by scaling all signals to an intensity of 400. Affymetrix GCOS 1.4 software was used to obtain absent/present calls. Briefly, present calls required P<0.05 and marginal calls required 0.05<P<0.065; probe sets with P>0.065 were marked as absent. Each microarray was subsequently analysed based on the present (gene expression) for each probe set. To ensure reliability of the analyses, each sample was performed with 3 replicates. Replica quality was assessed by pair-wise comparisons of signal intensities. We chose the mean value of normalized signal intensity of three replicates for further analysis. Two main criteria were used for stringently defining the genes with differential expression between datasets of experimental groups (including SO groups and PH groups) at different time points and that of the control group. The t test method, available as a Web-based resource, was used in the normalized signal intensities to detect the significant differences in gene expression, with a false discovery rate cut-off with ≤0.5%, to identify the genes showing differential expression. As a result, the genes whose expression level in experimental samples differed from that of control samples by a factor of 3 or above were identified as up-regulated; by a factor of 0.33 or below, as down-regulated; between 0.33 and 2.99, as insignificantly expressed genes.

2.6. Quantitative real-time PCR

To assess the array results, real-time PCR was performed. RNA samples for RT–PCR (reverse transcription–PCR) were from the isolated KCs at 10 time points after PH. Double-stranded cDNA template preparation and purification were performed with a RT kit (Promega) according to manufacturer's protocol. The primers sequences were obtained for nine rat genes including apoe, jun, myc, pcna, ttr, cd68, b2m, ubc and gapdh, and synthesized following the protocol described by Wang and Xu (2010). And amplification of target cDNA and the standard curves construction were also in accordance with the method of Wang and Xu (2010).

2.7. 2-DE and MALDI–TOF–MS/MS (matrix-assisted laser-desorption ionization–time-of-flight MS) identification

Total protein was extracted from the isolated rat KCs as in the methods mentioned above, and was homogenized with a mortar and pestle in liquid nitrogen. The KC sample powder was re-suspended in lysis buffer (30 mM Tris/HCl, 2 M thiourea, 7 M urea, 4% CHAPS and ∼0.5 ml/200 mg of tissues). The suspension was centrifuged at 25000 g for 1 h at 4°C, and its protein concentration was determined by a modified Bradford assay. The first dimension IEF (isoelectric focusing) was carried out using 24 cm IPG (immobilized pH gradient) dry strips (GE Healthcare) with a linear pH 3–10 gradient. Proteins were separated by the Ettan IPGhor 3 system (GE Healthcare) using a programmed voltage gradient at 20 μC with a current limit of 50 mA per strip in the dark under the following conditions: 30 V 12 h, 250 V 4 h, 1000 V 2 h, 10000 V linear gradient with in 3 h and rapid 10000–60000 V h for a total of 75 kVh. After IEF, the IPG strips were equilibrated in buffer 1 [50 mM Tris/HCl, pH 8.8, 8 M urea, 30% glycerol, 2% SDS and 0.5% DTT (dithiothreitol)] and buffer 2 containing 4.5% iodoacetamide instead of DTT, in each case for 15 min. The strips were transferred on to SDS/12.5% PAGE gels for the two-dimensional separation (GE Healthcare, Ettan DALT 6 system, 1 W/gel for 1 h and 1 W/gel until end at 15°C, ∼4.5 h). Two-dimensional images were scanned by Image scanner III and analysed by ImageMaster 2D Platinum 7.0 software. Protein identification was performed by MALDI–TOF–MS/MS mass spectrometer (Bruker Dalton, Autoflex III).

2.8. Gene functional annotation and statistical analysis

The normalized signal intensity in Chip analysis was log2-transformed for conveniently classifying the gene expression patterns based on temporal changes. Gene Ontology Database (http://www.geneontology.org/) was used as a standard source for gene annotation information. To examine statistical significance for frequencies of a given functional genes group in each patterns, we used a modified Fisher's exact test to measure the gene-set enrichment in the annotation terms.

3. Results

3.1. Changes in morphology and number of KCs during LR

Immunohistochemical assays showed that ED2- and LYZ-positive KCs in the control group are predominantly localized in the lumen of hepatic sinusoids, adhering to the liver SECs. The number of positive KCs in the remnant tissue increased as regeneration proceeded. Since 72 h post-PH, the number of positive cells began to decrease. From 120 h after PH, the number and location of positive cells in PH groups were similar to the control group (Figures 1A and 1B). Immunocytochemical result showed that most KCs had abnormal morphology, but had distinct cell borders, with ED2- or LYZ-positive cytoplasm-staining, small and round nuclei (Figures 1C and 1D). In contrast with the control, the purified KCs at 2 h post-PH were characterized by round or oval somata and the majority of KCs (96% or above) were stained by ED2 and LYZ in darkly stained cytoplasm with high refractivity. With the proceedings of the regeneration, KCs showed a slightly atypical shape accompanied with weakened cytoplasm staining. Until 168 h, the morphology of isolated KCs resembled that of the control.

3.2. The yield, purity and survival rate of KCs

Based on two-step perfusion, collagenase digestion and Percoll density centrifugation, this study utilized the immunomagnetic beads sorting method to isolate KCs of high yield, purity from the regenerating rat liver. The number of mixed liver cells and KCs at 10 time points was on average at least 4.4×108 cells and 4.5×106 cells per animal respectively (Figures 2A and 2B). The survival rate of well-dispersed liver cells and the purified KCs was at least 90.9 and 94.9% respectively (Figure 2C). Immunocytochemical test showed that the proportion of ED2- and LYZ-positive cells among liver cells were sequentially at least 7.9 and 7.9%; among the isolated KCs, correspondingly, at least 96.3 and 95.2% (Figure 2D). In addition, we performed Western blot assay to detect the changes in levels of KC markers ED2 and Lyz across the cell samples from SO and PH groups. The results showed that there were no significant differences in abundance of either ED2 or LYZ between all the cell samples (Figure 3), indicating that the approach developed resulted in high-efficiency purification.

3.3. Identification of rat LR-related genes

In the present study, we measured gene expression profiles of KCs from 2 to 168 h using rat genome 230 2.0 microarray. Each sample corresponding to one time point was hybridized on to one array. In total, 10 time points were measured. After careful quality control analyses of each chip, Affymetrix GCOS 1.4 software was used to analyse the data. The mean of signal intensity of each gene from three replicates at each time point served as effective value for further analysis. Of 25020 transcripts on each microarray, 6343 genes exhibited a 3-fold change in expressions on at least two repeated microarray corresponding to one sample. To retain those promising genes and enable the statistical significance analysis, we chose 6343 genes for further screening analysis according to filter criteria. The filter left 4924 genes showing similar expression trends at the same time point in 3 independent analyses. To eliminate the influence of surgical operation on gene expression, we compared the transcriptional profiles between SO and PH groups applying ANOVA test method, finding that there were 2334 genes showing significant (P≤0.05) or very significant (P≤0.01) difference between SO and PH groups. These 2334 genes were considered as LR-related genes.

3.4. Time-course expression profiles of KCs from the regenerating rat liver

The 2334 LR-related genes comprised 1407 known genes and 927 unknown genes with unannotated functions. The former contained 676 up- and 731 down-regulated genes, and the latter contained 397 up-, 494 down- and 36 complex-regulated genes. Herein, our analyses focused on the similarity and difference of gene expression profiling among different time points rather than the differences at an indicated time point. For convenience sake, we referred to the log2 ratio of gene expression data between each experimental samples and 0-h sample as expression time difference.

Figure 4 displayed the dendrogram for hierarchical clustering of nine time points after PH according to the expression changes of 2334 LR-related genes. It can be seen that the samples fell into 2 major groups: one on the right contains only 120- and 168-h samples, and the group on the left contains other 7 time points including 2, 6, 12, 24, 30, 36 and 72 h, within which 6- and 12-h cell samples cluster together on one branch; 24- and 30-h samples on the other branch and 36- and 72-h samples on third branch. In addition, 2-h sample and other samples severely differed in the gene expression profiling. Regarding the number of differently expressed genes, the genes differently expressed at 12 (1357 genes) and 24 h (1095 genes) after PH highly outnumber those at other time points, suggesting the relatively sharp changes in transcriptional regulation at early phase of LR. Notably, following partial hepatic resection, the number of down-regulated genes at many time points, including 12, 24 and 36 h, was equal to or even more than that of up-regulated genes.

3.5. Quantitative real-time PCR

We selected 9 genes, including apoe, jun, myc, pcna, ttr, cd68, b2m, ubc and gapdh, to verify the changes in expression levels measured by rat genome 230 2.0 microarray using RT–PCR technology. As indicated in Figure 5, mRNA levels for 7 genes (apoe, jun, pcna, ttr, cd68, ubc and gapdh) were not significantly altered by both methods. Myc, whose expression was up-regulated at 2–24 h post-PH, had an elevated increased mRNA level only after 2 h on the microarray. The gene, b2m, exhibiting a slight decrease at 12–36 h after PH by GeneChip, had a reduced expression at other time points except for 12, 24, 30 and 36 h. Although the expression levels of some genes measured by RT–PCR were different from those by microarray, their expression trends by the 2 techniques were overall similar, confirming the competency of chip analysis for detecting gene expression profiling.

3.6. Temporal gene expression patterns of KCs from regenerating rat liver

Using H-clustering analysis, the expression changes of the 1407 known genes were distinctly divided into up- and down-regulated expression (Figure 6A). According to their temporal regulation, K-means clustering identified 14 TCs (temporal clusters; TC1–TC14) including 7 groups for up-regulated transcripts corresponding to 676 genes whose expressions during LR are listed in Supplementary Table S1 (available at http://www.cellbiolint.org/cbi/036/cbi0360721add.htm; Figure 6B) and 7 groups for down-regulated transcripts corresponding to 731 genes whose expressions are presented in Supplementary Table S2 (available at http://www.cellbiolint.org/cbi/036/cbi0360721add.htm; Figure 6C). Their features are as follows: TC1, transient up-regulation at 12 h and a rapid decrease; TC2, transient up-regulation at 24 h and rapid decrease; TC3, gradual increase with a peak at 12 h followed by a gradual return; TC4, similar to TC3, but peaking at 30 h; TC5 and TC6, gradual increase; TC7, rapid increase and persistence; TC8, down-regulation between 0 and 72 h with a subsequent gradual recovery; TC9 and TC10, down-regulation between 0 and 24 h and between 120 and 168 h; TC11 and TC12, general decrease throughout the regeneration process; TC13, slight increase at early phase followed by down-regulation and TC14, gradual decrease and persistence. According to the similarity of temporal expression patterns, 14 temporal clusters were further categorized into 9 distinct groups: 4 up-regulated patterns were G1 (TC1 and TC2), G2 (TC3 and TC4), G3 (TC5 and TC6) and G4 (C7) and 5 down-regulated patterns were G5 (C8), G6 (TC9 and TC10), G7 (TC11 and TC12), G8 (TC13) and G9 (TC14). These nine expression patterns were, in order, composed of 178, 188, 198, 112, 243, 152, 225, 14 and 97 genes, indicating a relatively uniform distribution of these genes in the other eight expression patterns, except G8. Here, we picked up some genes including fasn from G1, krt8 from G2, capg from G3, aldh1a7 from G4, cmbl from G5, aox1 from G6, inmt from G7, sult1e1 from G8, akr1d1 from G9, and used the 2-DE method to examine the protein expression pattern of these genes whose mRNA level had changed after PH. On the whole, the changes in protein abundance during the treatment were in agreement with changes in transcript abundance for these proteins as determined by microarray analysis (Figure 7), indicating the validity of these microarray analyses.

3.7. Gene function category analysis

We categorized the 1407 known genes based on their functions as possible, referring to Gene Ontology Annotation Database, and classified them into following biological categories: cellular metabolism, energy production, immune response, inflammatory response, cell chaemotaxis, cytokine production, signal transduction, cellular organization and biogenesis, cell proliferation, differentiation, apoptosis, etc. Figure 7 lists 24 overrepresented categories and gene numbers in each category. Apparently, there were a majority of genes encoding hypothetical or function-unknown proteins (15.7%). Among the function-assigned genes, a large proportion of genes (148 genes) were associated with cell differentiation and development. The second largest group of 133 genes was involved in intracellular signalling, which is consistent with the concept that at the early phase mRNA expression is associated with signal propagation. The third largest group containing 294 genes was expected to modulate cellular metabolism (including substance metabolism and energy metabolism). Moreover, a relatively large proportion (up to 180 genes) of 1407 genes was related to immunity, inflammation and defence response, presumably reflecting a drastic change in KC immune functions.

3.8. Gene set enrichment analysis in relation to 8 expression patterns

As described above, 676 up-regulated genes were robustly classified into four groups (G1–G4) of expression patterns. Gene Ontology analysis showed that they were involved in almost all the 24 biological activities. The frequencies of these functionally categorized genes in each expression pattern are represented with their statistical significance evaluated by Fisher's exact test (Table 1). Together with the results of Figure 4, high frequencies were observed in the following gene distributions: the genes in category ‘cell proliferation’ in the pattern characterized by ‘transient increase and rapid return’ (G1 = TC1+TC2), ‘cytokines, inflammatory response and cell chaemotaxis’ in the pattern ‘gradual increase and gradual return’ (G2 = TC3+TC4), ‘growth factors and cytokines-mediated signal transduction’ in the pattern ‘gradual increase’ (G3 = TC5+TC6); ‘differentiation and development’ in the pattern ‘rapid increase and persistence’ (G4, namely TC7).


Table 1 Significantly enriched gene function categories in 9 expression patterns

Gene expression patterns Gene function categories Numbers (%) of genes included in each group P value
Up-regulated expression
    G1 Cell cycle and proliferation 25 (14.05) 3.6E-03
    G2 Lipid metabolism 22 (11.70) 1.4E-03
Cytokines 8 (4.25) 1.8E-02
Defence response 17 (9.04) 4.7E-03
Inflammatory response 14 (7.45) 5.1E-03
Cell chemotaxis 6 (3.19) 7.1E-03
    G3 Signal transduction 32 (16.16) 6.1E-04
    G4 Differentiation and development 29 (28.57) 1.0E-04
Down-regulated expression
    G5 Cell cycle and proliferation 27 (11.11) 8.7E-04
Cell adhesion 13 (5.34) 2.5E-02
Signal transduction 30 (12.34) 1.4E-02
    G6 Carbohydrate metabolism 8 (5.26) 8.3E-03
Lipid metabolism 20 (13.15) 1.6E-03
Amino acid and derivative metabolism 6 (3.94) 7.4E-03
Secondary metabolism 23 (15.13) 9.5E-07
    G7 Differentiation and development 23 (10.27) 3.3E-02
Apoptosis 16 (7.14) 8.8E-04
Defence response 32 (14.29) 5.2E-04
Immune response 53 (23.66) 2.7E-07
    G8 Secondary metabolism 5 (35.71) 7.3E-04
Transport 6 (42.86) 2.6E-04
    G9 Secondary metabolism 13 (13.40) 2.1E-03
Transport 12 (12.37) 4.7E-02


P values are based on the Fisher’s exact test.



The 731 down-regulated genes were roughly classified into five groups (G5–G9) based on the similarity of expression patterns. GO analysis showed that they were also categorized into the biological activities listed in Figure 8. We evaluated the frequencies of these functionally classified genes in 5 down-regulation patterns, and found that among the genes grouped into G5 (having properties of down-regulation and gradually recovery), a significant frequency was the members for categories ‘cell adhesion’ and ‘signal transduction’; genes for categories ‘metabolisms of carbohydrates, lipids, amino acids and secondary substances’ were significantly enriched in G6 exhibiting the features of gene repression during 0–24 h and during 120–168 h; genes in ‘cell apoptosis’, ‘defence response’ and ‘immune response’ were enriched in G7 characterized by general decrease during LR; whereas genes enriched in both G8 and G9 were the members of the categories ‘secondary metabolism’ and ‘transport’.

4. Discussion

KCs are one of the most important populations of liver macrophages (Zocco et al., 2006). Besides the well-documented immunomodulatory function, KCs also play a pivotal role in LR (Amemiya et al., 2011). Previous studies on the relationship between KC and LR focused on the direct systemic effect of KC on LR using cell depletion method (Murata et al., 2008). To elucidate the role of KC in LR at molecular level, in this study, we developed an improved isolation method based on our past work: on the basis of two-step perfusion and collagenase digestion, we isolated and purified KCs from regenerating rat liver at different time points post-PH using 60% Percoll density centrifugation and immunomagnetic beads methods. It was proven that the yield of KCs was at least 4.5×107 cells/rat depending on this approach. Trypan Blue staining showed that the survival rate of KCs exceeded 96%; immunochemical staining for KC markers ED2 and LYZ identified a high purity (≥96%) of the isolates. It suggested that high purity, survival rate and yield of LSECs were obtained by above methods.

According to GeneChip data, a substantial number of genes underwent the dynamic expression changes in KCs during rat LR. We identified 2334 LR-related genes fulfilling the criteria previously described. This may be the first study revealing such a number of differently expressed genes in KCs. Among 2334 genes, there were 1407 known genes consisting of 676 up- and 731 down-regulated genes, suggesting the active transcriptional regulation post-PH. According to GO analysis, these genes were primarily related to cellular metabolism, differentiation and development (cell adhesion, cellular organization and biogenesis, cell differentiation and development), immunity, inflammation, cell chemotaxis, defence response, etc.

Using Fisher's exact test, among 4 up-regulation patterns (G1–G4), the genes enriched in G1 were involved in cell cycle, such as G1/S and G2/M transition (rbl1, ccne2, gadd45a and gadd45g), cytokinesis (anln, aurkb and tcf19), chromosome segregation (cdc20, nuf2 and pmf1), spindle organization (espl1, aurka, cdca8, ndc80 and ttk) and DNA replication (mcm3, mcm5 and mcm6). According to chip data, these cell proliferation-involved genes showed an increase in expression at 24–30 h after PH. The findings of others have revealed that the peak mitotic activity of KCs occurred at approximately 48 h post-operatively (Widmann and Fahimi, 1975). In view of the time lag of protein translation compared with transcription, our results are consistent with the findings of these authors. Some studies have suggested that liver metabolic homoeostasis is temporarily imbalanced after rat PH, resulting in a drastic augmentation of lipid metabolism (Michalopoulos and DeFrances, 2005). Guo et al. (2006) also found that fatty acids oxidation-related genes were up-regulated after PH. In the current study, lipid metabolism genes were present frequently in G2 based on the result of gene-set enrichment analysis, which obviously differed from the published work. Moreover, a case in point that needs to be mentioned was that lipid metabolism-involved genes that were highly expressed between 0 and 72 h mainly encode the proteins catalysing fatty acid synthesis and sterol metabolism, suggesting that the main function of KC does not essentially provide energy required for hepatic restoration, and that fatty acid synthesis and sterol metabolism may contribute to cell membrane formation. Typically, the regeneration process in rats ends ∼7 days after surgery at which time metabolic activities returned to normal (Mitchell et al., 2005). In addition, some lipid metabolism-related genes, mainly involved in fatty acid catabolism and sterol metabolism, were observed in this study to be significantly enriched in G6 exhibiting the decreased expression at 120–168 h post-PH, implying that the decrease of metabolism at terminal phase was coupled to the end of regeneration. A body of study report that KCs after PH would quickly release many inflammatory agents, which in turn recruit inflammatory cells to induce inflammation (Ma et al., 2006). Our results also showed that genes in categories ‘cytokines’, ‘inflammation’ and ‘cell chaemotaxis’ were enriched in G2. Hence, KCs were very sensitive to PH-induced liver injury, in accordance with the conclusion of others. It is now widely recognized that the biological events discussed above are governed by signal transduction (Brenner, 1998). In this study, the genes in category ‘signal transduction’ are mainly involved in cytokines and chemokines-mediated signal pathway modulating immunity and inflammation (il1a, hbegf and fgfr2), BMP (bone morphogenetic protein) signal pathway and Wnt signal pathway controlling cell growth and differentiation (fstl1, bmp2 and zeb2), and G-protein receptor and Ras signal pathways regulating various biological processes (gprs, gna14, eltd1, ccrl2, rasgrp3 and rfxank) were preferentially distributed in G3, suggesting that signalling activities last the whole regeneration process.

Using the same method, genes of high-frequency distributed in 5 down-regulation patterns (G5–G8) were related to ‘cell adhesion’, ‘cellular metabolism’, ‘apoptosis’, ‘cellular transport’ and ‘immunity and defence response’. The genes that significantly frequently clustered in G5 were mostly the members of the category ‘cell adhesion’, including cell junction genes snx10, etc., integrin genes itgal, itgam, etc., focal adhesion genes ptk2b, etc. and several regulatory genes pscd3. These genes were down-regulated between 6 and 30 h after PH. Cell adhesion could be prominently important in development, morphogenesis and organ regeneration (Kerrigan et al., 1998). Thus the decreased expression of cell adhesion-related genes at early phase of LR may lead to reduced interaction between cells, thus contributing to the proliferation of KCs. From the data in Table 1, the cellular metabolism-involved genes were significantly enriched in the down-regulation patterns. In detail, genes for carbohydrate, lipid, amino acid and secondary metabolisms were highly frequent in G6; at the same time, a large proportion of secondary metabolism-related genes were also enriched in G8 and G9. Notably, carbohydrate metabolism-related genes in G6 were mainly responsible for glycolysis, reinforcing the above conclusion that KCs are not the main cells providing energy for regeneration. GO analysis indicated that the secondary metabolism genes enriched in G6, G8 and G9 encode the products for metabolisms of drug, alcohol and vitamin, and their reduced expressions might suggest the capacity of KCs to the reduced clearance following PH. KCs are accepted as the immunomodulatory cells residing in the liver, and function by synthesizing Fc receptor, complement elements and various antigens (Decker, 1990). In our study, genes for 2 major KC functions, i.e. immune response and defence response, were mostly assigned to G7, characterized by the general decrease during LR. The above 2 categories are mainly composed of CD antigens cd37, cd3d, cd3e, cd4, etc., killer cell lectin-like receptor family genes klrc2, klrd1, klrk1, etc. and MHC family genes rt1-ba, rt1-ce5, rt1-da, etc. Their mRNA abundance is under control after PH with the slight recovery at terminal phase. Seemingly, the immune and defence responses of KCs constantly dropped after surgery. Clearly our conclusion disagrees with previous reports, which needs to be explained by further investigations. However, the findings come entirely from the chip data. The GeneChip technology can only measure the levels of gene transcription, and cannot quantify the amount of protein translated. In future, we will test above results using gene addition, RNAi and protein–protein interaction.

Author contribution

Cunshuan Xu was the leading investigator in the laboratory, and designed the experimental procedures. Xiaoguang Chen collected and collated the research results. Cuifang Chang was mainly responsible for the processing and analysis of ChIP data. Gaiping Wang was mainly responsible for qRT-PCR performance. Wenbo Wang was mainly responsible for the establishment of PH rat model and the isolation of Kupffer cells. Lianxing Zhang was mainly responsible for the Gene Chip Detection. Qiushi Zhu was mainly responsible for immunochemmical assay. Lei Wang was mainly responsible for 2-DE assay and MALDI-TOF-TOF identification.

Funding

This work was supported by the National Basic Research 973 Pre-research Program of China [grant number 2010CB534905].

REFERENCES

Amemiya, H, Kono, H and Fujii, H (2011) Liver regeneration is impaired in macrophage colony stimulating factor deficient mice after partial hepatectomy: the role of M-CSF-induced macrophages. J Surg Res 165, 59-67
Crossref   Medline   1st Citation  

Böhm, F, Köhler, UA, Speicher, T and Werner, S (2012) Regulation of liver regeneration by growth factors and cytokines. EMBO Mol Med 2, 294-305
1st Citation  

Boulton, R, Woodman, A, Calnan, D, Selden, C, Tam, F and Hodgson, H (1997) Nonparenchymal cells from regenerating rat liver generate interleukin-1alpha and -1beta: a mechanism of negative regulation of hepatocyte proliferation. Hepatology 26, 49-58
Medline   1st Citation  

Brenner, DA (1998) Signal transduction during liver regeneration. J Gastroenterol Hepatol 13, (Suppl), S93-5
Medline   1st Citation  

Decker, K (1990) Biologically active products of stimulated liver macrophages (Kupffer cells). Eur J Biochem 192, 245-61
Crossref   Medline   1st Citation  

Desbois-Mouthon, C, Wendum, D, Cadoret, A, Rey, C, Leneuve, P and Blaise, A (2006) Hepatocyte proliferation during liver regeneration is impaired in mice with liver-specific IGF-1R knockout. FASEB J 20, 773-5
Medline   1st Citation  

Dong, Z, Zhang, J, Sun, R, Wei, H and Tian, Z (2007) Impairment of liver regeneration correlates with activated hepatic NKT cells in HBV transgenic mice. Hepatology 45, 1400-12
Crossref   Medline   1st Citation  

Edenberg, HJ, Strother, WN, McClintick, JN, Tian, H, Stephens, M and Jerome, RE (2005) Gene expression in the hippocampus of inbred alcohol-preferring and -nonpreferring rats. Genes Brain Behav 4, 20-30
Medline   1st Citation  

Gehring, S, Sabo, E, San Martin, ME, Dickson, EM, Cheng, CW and Gregory, SH (2009) Laser capture microdissection and genetic analysis of carbon-labeled Kupffer cells. World J Gastroenterol 15, 1708-18
Crossref   Medline   1st Citation  

Guo, F, Nian, H, Zhang, H, Huang, L, Tang, Y, Xiao, X and He, D (2006) Proteomic analysis of the transition from quiescent to proliferating stages in rat liver hepatectomy model. Proteomics 6, 3075-86
Crossref   Medline   1st Citation  

Higgins, GM and Anderson, RM (1931) Experimental pathology of liver resection. Arch Pathol 12, 186-97
1st Citation  

Kerrigan, JJ, McGill, JT, Davies, JA, Andrews, L and Sandy, JR (1998) The role of cell adhesion molecules in craniofacial development. J R Coll Surg Edinb 43, 223-9
Medline   1st Citation  

Kiiasov, AP and Gumerova, AA (2002) The Ito cells in ontogenesis and liver regeneration. Tsitologiia 44, 342-9
Medline   1st Citation  

Kube, DM, Savci-Heijink, CD, Lamblin, AF, Kosari, F, Vasmatzis, G and Cheville, JC (2007) Optimization of laser capture microdissection and RNA amplification for gene expression profiling of prostate cancer. BMC Mol Biol 8, 25
Crossref   Medline   1st Citation  

Ma, W, Wang, ZR, Shi, L and Yuan, Y (2006) Expression of macrophage inflammatory protein-1alpha in Kupffer cells following liver ischemia or reperfusion injury in rats. World J Gastroenterol 12, 3854-8
Medline   1st Citation  

Malik, R, Selden, C and Hodgson, H (2002) The role of non-parenchymal cells in liver growth. Semin Cell Dev Biol 3, 425-31
1st Citation  

McClintick, JN, Jerome, RE, Nicholson, CR, Crabb, DW and Edenberg, HJ (2003) Reproducibility of oligonucleotide arrays using small samples. BMC Genomics 4, -4
Medline   1st Citation  

Meijer, C, Wiezer, MJ, Diehl, AM, Schouten, HJ, Schouten, HJ and Meijer, S (2000) Kupffer cell depletion by CI2MDP-liposomes alters hepatic cytokine expression and delays liver regeneration after partial hepatectomy. Liver 20, 66-77
Crossref   Medline   1st Citation  

Michalopoulos, GK and DeFrances, M (2005) Liver regeneration. Adv Biochem Eng Biotechnol 93, 101-34
Medline   1st Citation  

Michalopoulos, GK (2010) Liver regeneration after partial hepatectomy: critical analysis of mechanistic dilemmas. Am J Pathol 176, 2-13
Crossref   Medline   1st Citation  

Mitchell, C, Nivison, M, Jackson, LF, Fox, R, Lee, DC, Campbell, JS and Fausto, N (2005) Heparin-binding epidermal growth factor-like growth factor links hepatocyte priming with cell cycle progression during liver regeneration. J Biol Chem 280, 2562-8
Medline   

Murata, S, Matsuo, R, Ikeda, O, Myronovych, A, Watanabe, M and Hisakura, K (2008) Platelets promote liver regeneration under conditions of Kupffer cell depletion after hepatectomy in mice. World J Surg 32, 1088-96
Crossref   Medline   1st Citation  

Norton, JN (1992) Total RNA isolation by a rapid centrifugation method. Am Biotechnol Lab 10, 41
Medline   1st Citation  

Roberts, RA, Ganey, PE, Ju, C, Kamendulis, LM, Rusyn, I and Klaunig, JE (2007) Role of the Kupffer cell in mediating hepatic toxicity and carcinogenesis. Toxicol Sci 96, 2-15
Medline   1st Citation  

Scott, RJ (1999) Isolation of whole cell (total) RNA. Methods Mol Biol 49, 197-202
1st Citation  

Stoick-Cooper, CL, Moon, RT and Weidinger, G (2007) Advances in signaling in vertebrate regeneration as a prelude to regenerative medicine. Genes Dev 21, 1292-315
Crossref   Medline   1st Citation   2nd  

Su, AI, Guidotti, LG, Pezacki, JP, Chisari, FV and Schultz, PG (2002) Gene expression during the priming phase of liver regeneration after partial hepatectomy in mice. Proc Natl Acad Sci USA 99, 11181-6
Crossref   Medline   1st Citation  

Takeishi, T, Hirano, K, Kobayashi, T, Hasegawa, G, Hatakeyama, K and Naito, M (1999) The role of Kupffer cells in liver regeneration. Arch Histol Cytol 62, 413-22
Crossref   Medline   1st Citation   2nd  

Taub, R (2004) Liver regeneration: from myth to mechanism. Nat Rev Mol Cell Biol 5, 836-47
Crossref   Medline   1st Citation  

Wang, GP and Xu, CS (2010) Reference gene selection for real-time RT–PCR in eight kinds of rat regenerating hepatic cells. Mol Biotechnol 46, 49-57
Crossref   Medline   1st Citation  

Wang, WB, Fan, JM, Zhang, XL, Xu, J and Yao, W (2009) Serial expression analysis of liver regeneration-related genes in rat regenerating liver. Mol Biotechnol 43, 221-31
Crossref   Medline   1st Citation  

Widmann, JJ and Fahimi, HD (1975) Proliferation of mononuclear phagocytes (Kupffer cells) and endothelial cells in regenerating rat liver. A light and electron microscopic cytochemical study. Am J Pathol 80, 349-66
Medline   

Zocco, MA, Carloni, E, Pescatori, M, Saulnier, N, Lupascu, A and Nista, EC (2006) Characterization of gene expression profile in rat Kupffer cells stimulated with IFN-alpha or IFN-gamma. Dig Liver Dis 38, 563-77
Crossref   Medline   


Received 14 February 2011/2 March 2012; accepted 27 March 2012

Published as Cell Biology International Immediate Publication 27 March 2012, doi:10.1042/CBI20110104


© The Author(s) Journal compilation © 2012 International Federation for Cell Biology


ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)