Transcription Profiles of Marker Genes Predict The Transdifferentiation Relationship between Eight Types of Liver Cell during Rat Liver Regeneration.

OBJECTIVE: To investigate the transdifferentiation relationship between eight types of liver cell during rat liver regeneration (LR).
MATERIALS AND METHODS: 114 healthy Sprague-Dawley (SD) rats were used in this experimental study. Eight types of liver cell were isolated and purified with percoll density gradient centrifugation and immunomagentic bead methods. Marker genes for eight types of cell were obtained by retrieving the relevant references and databases. Expression changes of markers for each cell of the eight cell types were measured using microarray. The relationships between the expression profiles of marker genes and transdifferentiation among liver cells were analyzed using bioinformatics. Liver cell transdifferentiation was predicted by comparing expression profiles of marker genes in different liver cells.
RESULTS: During LR hepatocytes (HCs) not only express hepatic oval cells (HOC) markers (including PROM1, KRT14 and LY6E), but also express biliary epithelial cell (BEC) markers (including KRT7 and KRT19); BECs express both HOC markers (including GABRP, PCNA and THY1) and HC markers such as CPS1, TAT, KRT8 and KRT18; both HC markers (KRT18, KRT8 and WT1) and BEC markers (KRT7 and KRT19) were detected in HOCs. Additionally, some HC markers were also significantly upregulated in hepatic stellate cells ( HSCs), sinusoidal endothelial cells (SECs) , Kupffer cells (KCs) and dendritic cells (DCs), mainly at 6-72 hours post partial hepatectomy (PH).
CONCLUSION: Our findings indicate that there is a mutual transdifferentiation relationship between HC, BEC and HOC during LR, and a tendency for HSCs, SECs, KCs and DCs to transdifferentiate into HCs.

Introduction

Mammalian liver is almost unique amongst body tissues in its regenerative capacity (1). The capacity of the liver to regenerate after resection has been known since the late 1800’s (2). From that time onward, numerous scientists have devoted themselves to the study of liver regeneration (LR). However, many questions about LR have not been clearly answered yet, especially with regard to the transdifferentiation activities of different types of liver cell, including hepatocytes (HCs), biliary epithelial cells (BECs), hepatic oval cells (HOCs), hepatic stellate cells (HSCs), sinusoidal endothelial cells (SECs), Kupffer cells (KCs), pit cells (PCs) and dendritic cells (DCs) (3, 4). Of these eight types of liver cell, the latter five cell types are also collectively known as liver non-parenchymal cells. Transdifferentiation means the conversion of one differentiated cell type to another (5). Currently, transdifferentiated cells can be examined using cell function, epigenome, transcriptome, or proteome profiles, or by tracing the expression of markers of the target cell type. Among these methods, measurement of cell specific markers, considered as the potential indicator for identification or tracing the differentiation of specific cell types, is the most utilized approach (6). As for transdifferentiation relationships among different liver cells, at present, many researchers are primarily focused on studying transdifferentation between HOC, HC and BEC and have made significant progress. For instance, many studies have come to the conclusion that HOCs can differentiate into HCs and BECs through the observation that HOC can express markers of both HC and BECs. Briefly, during the course of differentiation of HOC toward HC and BEC, the expression level of HOC markers tends to decrease, while the expression of levels of HC markers (such as ALB, AFP, G6P, HNF4a, KRT18) and BEC markers (such as GGT, KRT7, KRT19) gradually increase (7-10). An in vitro HC culture experiment carried out by Nishikawa et al. (11) showed that, in the course of HC culture, expressions of mature HC markers (such as ALB, HNF1, HNF4α and KRT8) were gradually lost. In turn, some bile duct-specific proteins (such as KRT7 and KRT19) began to be expressed, indicating that HCs have the potential to trans-differentiate into BEC. Additionally, transdifferentiation of mature HCs into biliary cells has been shown to occur in rat liver (12). It has also been reported that rat BECs are capable of undergoing hepatic differentiation upon sequential exposure to liver-specific factors. For example, experimental observation of in vitro rat BEC culturing by Snykers et al. (13) showed that when rat epithelial cells were exposed to a hepatic-stimulating microenvironment, biliary KRT19 and connexin CX43 both gradually declined in expression, and KRT19 expression even disappeared completely. In contrast, expression of HC marker KRT18 persisted throughout the culture process. Furthermore, hepatic HNFβ, AFP, TTR, HNF4, ALB, HNFα, MRP2 and CX32 were also strongly expressed, showing the differentiation capacity of BEC into HC. As mentioned above, this research was mainly carried out on the transdifferentiation relationships among HOC, BEC and HC. However, little is known about whether other transdifferentiation activities exist among the eight types of liver cell. For this reason, in this study we separately isolated the eight types of liver cell at 0, 2, 6, 12, 24, 30, 36, 72, 120 and 168 hours after partial hepatectomy (PH) and examined their transcriptional profiles with Rat Genome 230 2.0 Array. We also emphatically analyzed expression changes in the marker genes of the above liver cell types during the regeneration process, and the potential transdifferentiation relationships among these cell types.

Materials and Methods

Preparation of rats - the 2/3 hepatectomy model

Animals used in this experimental study are Sprague-Dawley (SD) rats that are obtained from the Animal Center of Henan Normal University. A total of 114 cleaning-grade adult rats, aged 10-12 weeks and weighing 190 ± 20 g were randomly divided into nine PH groups, nine sham-operation (SO) groups and one control group with 6 rats per group. Rats in the PH groups underwent an operation for 2/3 PH according to the guideline described by Higgins and Anderson (14). Briefly, the left and median lateral liver lobes were surgically removed, then the hepatectomized rats were allowed free access to food and water for 2, 6, 12, 24, 30, 36, 72, 120 and 168 hours, respectively, and sacrificed by cervical dislocation. Rats in the SO groups were treated as mentioned above, but no liver lobes were removed. The animals in the control group, as in the case of the 0-hour samples for both the SO and PH groups, were perfused immediately after the surgical removal of left and median lobes. At the same time, the rat body weight (g) and regenerating liver weight (g) were noted and the liver coefficient (Lc) was calculated using the following formula: Lc=regenerating liver weight (g)/ body weight (g)×100% (15). All procedures involving rats in this study were performed in accordance with the standard protocols approved by the Ethical Committee of Henan Normal University.

Isolation of different liver cell types

Rats were subjected to abdominal skin disinfection with alcohol after being anaesthetized by inhaling diethyl ether. The abdominal cavity was opened to expose the liver and the superior and inferior vena cava was ligated followed by portal vein cannulation. The dispersion of liver cells and isolation of different liver cell types were performed according to the method described previously (16). The liver was perfused with calcicum-free perfusate preheated at 37˚C until it turned grey, then with a 15 mL 0.05% collagenase IV solution (Invitrogen, USA) instead of perfusate at a flow rate of 1 mL/minutes. After the liver capsule was removed, the perfused liver was cut into small pieces and digested with 0.05% collagenase IV for 15 minutes at 37˚C. After this it was filtered through 200-well nylon netting (Corning, USA) and the liquid was centrifuged (3S-R low speed refrigerated centrifuge, Leica, Germany) at 500 g for 3 minutes. The pellet at the bottom was collected and washed three times in a 4˚C phosphate buffer saline (PBS) buffer to adjust the cell concentration to 1×108 cells/mL. Six mL of the mixed cell suspension was spread onto the surface of 4 mL 60% percoll (Pharmacia, Biotech AB, Uppsala, Sweden) in a 10 mL tube for a single centrifugation at 200 g for 5 minutes at 4˚C. The centrifuged pellets and supernatant were the purified HCs and nonparenchymal cells-enriched supernatant fractions, respectively. The supernatant was mixed with an equal volume of PBS, centrifuged at 400 g for 2×2 minutes at 4˚C. The mixed nonparenchymal cell-rich pellet was collected and adjusted to a concentration of 1×108 cells/mL, then mixed with 10 μL/mL of rat anti-THY1, -GFAP, -CD31, -CD68, CD161a, and -CD11c PE-antibodies (BD Biosciences, USA), respectively. HOCs, HSCs, sinusoidal endothelial cells, KCs, PCs and DCs were identified using the immunomagnetic bead method (17). White intrahepatic bile duct fractions left on the nylon netting were added into the digestive solution containing 0.25% trypsin (Sichuan Deebio Pharmaceutical Co., Ltd, China) and 0.05% collagenase IV, incubated at 37˚C for 50 minutes, and filtered through 200-well nylon netting. The filtered solution was centrifuged at 300 g for 5 minutes. The resulting sediment was the pellet enriched with BECs. BECs were isolated with rat anti-KRT19 PE-antibody as described above.

Immunohistochemical identification of eight types of liver cells

A few fractions of each type of liver cell was taken and fixed on glass slides with 10% formaldehyde (Nanjing Senbei Jia Biotechnology Co., Ltd., China) for 30 minutes, then smeared onto glass slides. Microwave antigen retrieval was undertaken once the cell samples on the glass slide had dried. In relation to HCs, for instance, the slides were incubated separately with anti-ALB and G6P antibody overnight at 4˚C, then with biotin-labeled secondary antibody at 37˚C for 60 minutes. The reacted sections were mixed with streptavidin-biotin complex (SABC, Wuhan Boster Biological Technology., Ltd., China) and incubated at 37˚C for 30 minutes. Finally, 3,3΄-diaminobenzidine (DAB, Wuhan Boster Biological Technology., Ltd., China) was added for staining and the results observed under an optical microscope (Shang Hai Tuo Feng Instrument Co., Ltd., China) (18). Similarly, BECs, HOCs, HSCs, SECs, KCs, PCs, and DCs were respectively identified with anti-KRT18 and GGT1, OC2 and OV6, CD14 and ET-1, LYZ and ED2, DES and VIM, CD8 and CD56, CD86 and CD103 antibodies following the above protocol.

Rat Genome 230 2.0 microarray detection and data analysis

Total RNAs from eight types of liver cell at each recovery time point were extracted one by one, according to the Trizol reagent manual (Invitrogen Corporation, Carlsbad, California, USA) and then purified following the RNeasy mini protocol (Qiagen, Inc, Valencia, CA, USA). The quality of the total RNA samples was assessed by agarose electrophoresis (180 V, 0.5 hours) with a 2:1 ratio of 28S rRNA to 18S rRNA intensities and optical density measurement at 260/280 nm prior to cDNA synthesis (19). RNAs pooled from 6 rats in each group were used as a probe. The probes were amplified and biotinylated [Gene Tech (Shanghai) Co., Ltd., China] according to the Affymetrix recommendations for microarray analysis. Probes were hybridized to the Rat Genome 230 2.0 microarray. Arrays were washed to remove the superfluous hybridization buffer, stained in a GeneChip fluidics station 450 (Affymetrix Inc., Santa Clara, CA, USA) and scanned with a GeneChip scanner 3000 (Affymetrix Inc., Santa Clara, CA, USA). The images obtained were converted into normalized signal values, and signal detections [present call (P), absent call (A) and marginal call (M)] values using Affymetrix GCOS 2.0 software (20). To minimize technical error during the array analysis, the Rat Genome 230 2.0 Array procedure was repeated using three liver cell samples at each time point.

The data for each microarray were normalized by scaling all signals to a target intensity of 200 using GCOS 2.0 software (Affymetrix, USA). Each probe set used in the Affymetrix GeneChip produces a detection call, with present expression (P, requiring P value<0.05) indicating good quality, marginal expression (M, requiring 0.05