Abnormal Distribution and Upregulation of Tenascin-C mRNA Expression in the Liver of 90% Hepatectomy Rats are Related to Post-hepatectomy Liver Failure

Abstract

Background:

Post-hepatectomy liver failure (PHLF) is the most serious complication after extended hepatectomy, with a high perioperative mortality rate. Matricellular protein tenascin-C (TNC) has been related to blood flow change and inflammation. We aimed to explore the hepatic changes of TNC and its relationship with PHLF.

Materials and Methods:

Rats were divided into seven groups: Sham operation, 90% partial hepatectomy (90PH) as a PHLF model, 85% (85PH), and 68% (68PH) hepatectomy without or with lipopolysaccharide (LPS) treatment. The distribution of TNC protein, the mRNA expression of TNC and other endothelial response genes, and the cellular TNC located in the rat livers were evaluated by immunostaining, quantitative real-time polymerase chain reaction, and RNAscope assay.

Results:

TNC mRNA in the liver tissue was significantly upregulated after hepatectomy and reached the highest level in the 90PH rats compared with those in the 85PH and 68PH rats. This upregulation was closely correlated with the changes of severe liver injury, inhibited liver regeneration, and upregulation of KLF2, ET-1, eNOS, and TM-1 in 90PH rat livers. We found a robust disturbance of hepatic TNC protein distribution in 90PH rat livers with an obvious reduction in hepatic sinusoids but an increase around large blood vessels, which were consistent with the severe liver injury of the 90PH rats. It was also found for the first time that treatment with the endotoxin LPS tremendously exacerbated the above changes of TNC.

Conclusion:

The distribution and expression of TNC mRNA changed dramatically and contributed to the occurrence of PHLF.

INTRODUCTION

Post-hepatectomy liver failure (PHLF) is the most serious complication after extended hepatectomy, with an incidence rate of 1%–35% and a perioperative mortality rate as high as 40%–60%.[1,2] In addition, adult living donor liver transplantation is often complicated by small-for-size syndrome (SFSS) due to the small size of the donor liver, which has similar pathogenesis and clinical manifestations to PHLF.[3] There are currently no good treatment measures except preoperative precise surgical planning or premature induction of hyperplasia of reserved liver lobes by surgical means such as portal vein ligation and associated liver partition and portal vein ligation for staged hepatectomy.[4,5]

Extended hepatectomy-induced excessive portal hyperperfusion leads to liver vascular endothelial injury, microcirculation disturbance, and intestinal endotoxin increase, thus causing nonparenchymal cell activation, inflammatory cell infiltration, inflammatory cytokine release, and liver injury.[6,7] These changes result in insufficient liver regeneration and liver failure.[8] Previous study on liver blood flow found that with the increase of liver resection volume, microcirculation perfusion was greatly increased, especially in the 90% hepatectomy liver failure group, compared with 68% and 85% hepatectomy survival groups.[9] And the expression of some matricellular proteins was closely related to PHLF.[10] The predictive role of tenascin-C (TNC) and cellular communication network factor 3 was suggested for PHLF. A large number of studies have shown that TNC expression is regulated by mechanical stress, both during development and in adulthood. Tenascins mediate both inflammatory and fibrotic processes in the pathogenesis of Ehlers-Danlos and heart disease.[11,12] Mechanical compression of human airway epithelial cells induces release of extracellular vesicles containing TNC.[13] TNC is involved in the pathophysiology of aneurysmal and dissecting lesions, in part by protecting the vascular wall from destructive mechanical stress.[14] These studies support the hypothesis that high flow shear stress caused by portal hyperperfusion after extended hepatectomy may be a key factor in the induction of TNC expression.

In liver ischemia-reperfusion injury[15] and steatotic mice,[16] TNC has been proved to play a damaging role by increasing lymphocyte infiltration and cytokine (IL-1β, IL-6, and CXCL2) and adhesion molecules (VCAM-1 and PECAM-1) expression. The residual liver after partial hepatectomy (PH), especially after extended hepatectomy, faces great changes of increased portal blood flow and shear force. This study is designed to explore in detail the changes in TNC distribution and expression following different volumes of hepatectomy, and to clarify its relationship with PHLF. This will provide a better understanding of the complexity and function of matricellular proteins and suggest new ideas for the prevention and treatment of PHLF and SFSS.

MATERIALS AND METHODS

Animals

Adult male Sprague–Dawley rats (250–300 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). They were housed in a standard animal laboratory and kept on a 12-h day/night cycle. All experimental procedures were conducted in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animal of the National Institutes of Health and were approved by the Committee on Ethics of Animal Experiments of the Hospital (ethical number: 2016-X12-16).

Experimental design

The rats were divided into the following four groups: 68%, 85%, and 90% partial hepatectomy (68PH, 85PH, and 90PH, respectively) groups, and a sham operation (Sham) group. In the lipopolysaccharide (LPS) treatment experiment, Sham, 85PH, and 90PH groups that received LPS treatment were included. At 12 and 24 h after hepatectomy, the rats were sacrificed by euthanasia. Liver tissue and blood were collected for further analysis. For each hepatectomy group, ten rats were used to determine the 7-day survival rate, and five to eight rats were used for sample collection at each time point. Five rats were included in the Sham group. LPS (Escherichia coli 0127:B8) was given via inferior vena cava at 0.5 mg/kg body weight after completion of the 85% and 90% hepatectomy (n = 6 per group). At approximately 12 h after hepatectomy and LPS administration, liver tissue and blood were collected for further analysis.

Surgical procedure for rat hepatectomy

Rats were fasted overnight but given free access to water before the procedure. The anesthesia was administered as continuous 1.5% (vol) isoflurane/oxygen inhalation. Partial hepatectomies were performed according to a previously established method.[9] The 68PH procedure was performed by resecting the left lateral lobe (30%) and the median lobes (38%). For the 85PH procedure, the left lateral, median, and inferior right lobes (10%) as well as the caudate lobes (8%) were resected. For the 90PH procedure, all the liver lobes except the caudate lobe (8%) and caudate process (2%–3%) were resected. Briefly, after the laparotomy, the Glisson’s sheath of each target liver lobe was exposed and ligated with 6-0 silk at the liver parenchyma near the lobe pedicle. Then, each liver lobe was ligated with 4-0 silk in two sections by piercing through the liver parenchyma near the inferior vena cava. Finally, the ligated liver lobe was resected and weighed. After peritoneal closure, the rats recovered in warm boxes with free access to food, 20% glucose water, and fresh drinking water. All rats were given a subcutaneous injection of meloxicam 2 mg/kg (0.1 ml/100 g) for 1–3 days based on the sample collection time for pain relief after surgery.

Assessment of liver regeneration

The liver regeneration rate was calculated as the ratio of the remnant liver weight at the time of sacrifice to the whole liver weight before hepatectomy. The whole liver weight was estimated by dividing the weight of the resected left and middle liver lobes by 0.68. The liver regeneration marker Ki-67 was assessed by immunohistochemistry (IHC).

Blood biochemical analysis and enzyme-linked immunosorbent assay (ELISA)

Blood samples were obtained from the inferior vena cava before the liver was removed. Serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin (TBIL) were determined using an automated procedure with a Cobas 8000 serum analyzer (Roche Diagnostics, Mannheim, Germany). The serum TNC level was determined with Tenascin-C Large (FNII-B) Assay Kit (Immuno-Biological Laboratories Co., Ltd., Tokyo, Japan).

Immunostaining

In immunofluorescence staining, the frozen liver tissue sections were fixed with cold methanol and acetone (1:1) for 15 min, then blocked with 0.5% bovine serum albumin for 30 min. The sections were then incubated with primary antibodies against TNC (4F10TT, Immuno-Biological Laboratories Co., Ltd., Tokyo, Japan), α-SMA (1:600, #19245, CST), fibronectin (FN) (1:200, ab268020, Abcam), desmin (1:200, ab32362, Abcam), or CD68 (1:200, ab31630, Abcam), respectively, at 4°C overnight. Then the sections were washed and incubated with rabbit secondary antibodies labeled with Alexa Fluor 488 or 594 (1:400, Jackson Immunoresearch, West Grove, PA). Nuclei were visualized with DAPI. Fluorescence microscopy was used to observe the sections.

In IHC staining, the paraffin-embedded sections were first treated with citrate antigen retrieval solution and then incubated at 4°C overnight with anti-Ki-67 antibody (1:50, #550609, BD Biosciences, USA), or CD68 antibody (1:200, ab31630, Abcam). Next, the sections were incubated with secondary antibody (PV-6000, ZSGB-BIO, China) for 60 min. The sections were stained with diaminobenzidine to detect positive signals, and nuclei were counterstained with hematoxylin. The Ki-67 labeling index was determined as the percentage of Ki-67-positive hepatocytes per total number of hepatocytes in five random visual fields (×400).

Quantitative real-time polymerase chain reaction

Total RNA was extracted from 50–60 mg frozen liver tissues using an RNASimple Total RNA Kit (Tiangen Biotech, Beijing, China) according to the manufacturer’s protocol. Reverse transcription was performed with 4 µg of RNA using a RevertAid First Strand cDNA Synthesis kit (Fermentas International Inc., Burlington, Canada). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed with the SYBR Premix Ex Taq II (TaKaRa Bio Inc., Dalian, China) on a StepOnePlus Real Time PCR system (Applied Biosystems, CA, USA). The relative gene expression levels were normalized to hypoxanthine–guanine phosphoribosyltransferase expression, calculated using the 2^-∆∆Ct method, and expressed as the fold change versus the Sham group. The primer sequences are shown in Table 1.

Table 1.

Primer sequences for quantitative reverse transcription polymerase chain reaction gene amplification

Gene	Forward (5′-3′ orientation)	Reverse (5′-3′ orientation)
TNC	GACTGCTCAGGTGGAGAGTTT	ATGGCCAAGGCTTCTGAGTC
KLF2	ACTTGCAGCTACACCAACTG	CTGTGACCCGTGTGCTTG
ET-1	CATCTGGGTCAACACTCCCG	GGCATCTGTTCCCTTGGTCT
eNOS	TATTTGATGCTCGGGACTGC	AAGATTGCCTCGGTTTGTTG
TM-1	ACCAGTCGCCTCCACTTT	TTCTCGCACGGCTTCTC
IL-1β	CACCTCTCAAGCAGAGCACAG	GGGTTCCATGGTGAAGTCAAC
TNFα	AAATGGGCTCCCTCTCATCAGTTC	TCTGCTTGGTGGTTTGCTACGAC
MIP-1α	GCGCTCTGGAACGAAGTCT	GAATTTGCCGTCCATAGGAG
HPRT	GCTGAAGATTTGGAAAAGGTG	AATCCAGCAGGTCAGCAAAG

TNFα=Tumor necrosis factor-α, IL-1β=Interleukin-1β, MIP-1α=Macrophage inflammatory protein-1α, HPRT=hypoxanthine–guanine phosphoribosyltransferase

RNAscope assay

The RNAscope assay followed the manufacturer’s instructions (Advanced Cell Diagnostics, USA). Paraffin-embedded liver tissue sections were treated with xylene (twice for 5 min) and ethanol (twice for 2 min), then dried by baking at 60°C for 2 min. Then, target retrieval was carried out with sequential steps of 15 min at 100°C, hydrogen peroxide for 10 min at 25°C, and RNAscope Protease IV for 30 min at room temperature. In between each treatment, the samples were washed twice with distilled water. RNAscope probes for TNC (Cat No. 465021, Advanced Cell Diagnostics, USA) and positive controls were applied. The samples were incubated for 2 h at 40°C in a HybEZ oven and then with sequential reagent steps of TNC. They were then counterstained with Gill’s hematoxylin, dried for 15 min at 60°C, treated with Fast Red for 10 min at 25°C in the dark, and mounted with Catamount permanent mounting medium (Vector Labs, USA). RNAscope assay was also combined with immunostaining for desmin or CD68 when colocalization was needed.

Statistical analysis

All results are expressed as the mean ± standard deviation. Student’s t-test was used to compare two groups. Using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA), comparisons of more than two groups were performed by one-way analysis of variance (ANOVA) and Tukey’s post-hoc test of multiple comparisons after the normality and equal variance across groups and time points were tested. If an equal variance test failed, statistical analyses were performed with Games-Howell multiple comparisons tests. P values < 0.05 were considered statistically significant.

RESULTS

90PH was an extended hepatectomy and hyperperfusion-induced fatal liver failure model in rats

All the rats that experienced 68PH and 85PH survived the observation period of 7 days, but only about 50% rats survived in the 90PH group [Figure 1a]. Based on the liver resection volume and survival rate, we considered 68PH as a model of major hepatectomy used for liver regeneration study, and 85PH as an extended hepatectomy model without causing death, while 90PH as a lethal liver failure model induced by extended hepatectomy and hyperperfusion.[9] Compared with 68PH and 85PH rats, 90PH rats showed overall increased serum ALT, AST, and TBIL, and decreased liver regeneration and Ki-67 labeling index [Figure 1b]. Extended hepatectomy-induced hyperperfusion and liver injury are the main causes of fatal liver failure in 90PH rats.

Figure 1.

Establishment of an extended hepatectomy-induced fatal liver failure model in rats. (a) Comparison of the survival rate, serum ALT, AST, and TBIL levels, liver regeneration, and Ki-67 labeling index of rats underwent sham operation or 68%, 85%, and 90% partial hepatectomy (68PH, 85PH, and 90PH). (b) H and E stains show different liver injury and immunostains show different numbers of Ki-67 positive hepatocytes in the liver of Sham, 68PH, 85PH, and 90PH rats at 24 h after hepatectomy. Values of 6–8 animals per group are given in means ± SD. ANOVA and post-hoc comparison: ^&P<0.05 and ^&&P<0.01 versus Sham, ^$P<0.05 and ^$$P<0.01 versus 68PH, ^#P<0.05 and ^##P<0.01 versus 85PH. SD=Standard deviation, ALT=Aminotransferase, AST=Aspartate aminotransferase, TBIL=Total bilirubin

Dramatic changes of tenascin-C expression and distribution in 90PH rats were related to the fatal liver failure

Our previous microarray data suggested that TNC was a significantly differentially expressed gene following different volumes of hepatectomy in rats. In this study, qRT-PCR was first used to assess the liver TNC mRNA expressions in the rats following 68PH, 85PH, and 90PH. In contrast to sham rat livers, in which TNC mRNA expression is minimally detected by qRT-PCR, TNC mRNA of the 90PH group showed great upregulation at 12 and 24 h after hepatectomy compared with the Sham, 68PH, and 85PH groups [Figure 2a]. TNC mRNA expressions of the 68PH and 85PH groups were also significantly higher than those of the Sham group, and the 85PH group is significantly higher than the 68PH group, although they were still much lower than those of the 90PH group.

Figure 2.

Detection of the changes of TNC in rat livers that underwent sham operation or 68%, 85%, and 90% partial hepatectomy (68PH, 85PH, and 90PH). (a) qRT-PCR detection of TNC mRNA in the rat livers. (b) Immunofluorescence staining of TNC in the rat livers. (c) ELISA detection of the serum TNC. n=6–8 for each group. ^&P<0.05 and ^&&P<0.01 versus Sham, ^$P<0.05 versus 68PH, ^#P<0.05 and ^##P<0.01 versus 85PH. qRT-PCR=Quantitative reverse transcription polymerase chain reaction, TNC=Tenascin-C

However, the changes of TNC protein in the liver observed by immunofluorescence staining were not well correlated with its mRNA level. We found that an obvious positive immunofluorescence signal of TNC protein was predominantly visible throughout the hepatic sinusoidal system with fiber-like prolongations and branches in Sham rats, although its mRNA level was low [Figure 2b]. Some show dendritic shapes of varying thickness. There is less positive signal around the central vein and portal vein. Interestingly, we find that the positive signals within the sinusoids are significantly reduced in the 90PH group at 12 and 24 h after hepatectomy compared with the Sham, 68PH, and 85PH groups. TNC in 90PH liver shows an intermittent network of discontinuous branches. However, positive staining of vascular in the portal and central vein is obviously increased in 90PH rats but very weak in the Sham group. We think the reduction of TNC in the hepatic sinusoids may be due to the disruption and release of the TNC network into the blood because of hyperperfusion, while the increase of TNC protein around the large blood vessels was consistent with the increase of TNC mRNA in the liver after hepatectomy. In particular, 90PH resulted in specific changes of TNC mRNA and protein.

Then we used ELISA to detect the serum TNC level, and found that there was a significant increase in serum TNC in all the rats after hepatectomy [Figure 2c]. This increase in serum TNC may be a combination of the release of existing TNC in the hepatic sinusoids and the upregulation of perivascular TNC expression. Anyway, extended hepatectomy resulted in a robust change of hepatic TNC distribution in the rat liver, especially in 90PH rats. The significant reduction of TNC as an extracellular matrix protein (ECM) in the hepatic sinusoids demonstrates a kind of disturbance of hepatic microvascular structure in 90PH rats.

The upregulation of TNC mRNA correlates with endothelial response gene in 90PH rat livers

KLF2, ET-1, eNOS, and TM-1 are well-known transcription factors or genes associated with high perfusion blood flow or high shear stress-induced endothelial injury. We found that the mRNA expressions of KLF2, ET-1, eNOS, and TM-1 were also upregulated and closely correlated with TNC mRNA expression in 90PH rat livers [Figure 3]. These results suggest that the greatly upregulated hepatic TNC mRNA expression in 90PH rats is also related to the high perfusion blood flow or high shear stress in 90PH rat livers.

Figure 3.

The correlation of TNC gene expression and the expressions of endothelial response genes by qRT-PCR in the rat livers at 24 h after 90% partial hepatectomy (n=8). TNC=Tenascin-C, qRT-PCR=Quantitative reverse transcription polymerase chain reaction

Identify the source cells of tenascin-C expression in rat liver after hepatectomy

As TNC is an ECM, we used RNAscope in situ probe hybridization to detect the source cells of TNC expression. As shown in Figure 4, there are minimal positive cells in the Sham group, but an obvious increase in the 68PH, 85PH, and 90PH groups. Importantly, a large number of TNC-positive cells appeared in the 90PH group, and these positive cells were mainly located around the portal vein and hepatic vein area, as well as along the straight path from the portal area to the hepatic vein. This correlates with the intensified immunofluorescence staining of TNC around this area in the 90PH group [Figure 2b]. These results suggest that TNC distribution is possibly related to the change of hepatic blood flow after hepatectomy, and this abnormal distribution might contribute to the liver injury and failure of 90PH rats.

Figure 4.

RNAscope in situ probe hybridization detects the cells expressing TNC mRNA in rat livers that underwent sham operation or 68%, 85%, and 90% partial hepatectomy (68PH, 85PH, and 90PH). HV=Hepatic vein, PV=Portal vein

Both macrophages and hepatic stellate cells (HSC) have been reported to express TNC. We then combined TNC RNAscope in situ probe hybridization with IHC or immunofluorescence for macrophage marker CD68 to identify the source cells of TNC in 90PH rat livers. As shown in Figure 5, double staining of TNC and CD68 [Figure 5a for IHC and Figure 5b for immunofluorescence] demonstrates partial CD68-positive cells showing co-localization with TNC-positive cells. The double-positive cells gathered around the portal veins. In addition, there are still considerable CD68-positive cells not showing co-localization with TNC. These results suggest that TNC is only expressed by a part of macrophages, which are stimulated by hyperperfusion blood flow, because the co-localized cells are around the portal veins.

Figure 5.

Double staining of macrophage marker CD68 and TNC by RNAscope in situ probe hybridization assay combined with IHC (a) and immunofluorescence (b) in 90% partial hepatectomy rat livers. IHC = Immunohistochemistry

Similarly, double immunofluorescence staining of HSC marker desmin and TNC shows that partial desmin-positive cells co-localize with TNC [Figure 6a]. α-SMA is a marker of activated HSC. Immunofluorescence co-staining of TNC and α-SMA shows no co-localization between them in the Sham and 90PH rat livers [Figure 6b]. α-SMA is mainly located in the wall of large blood vessels, and there is no obvious increase in 90PH rats. However, in 90PH rats, the large vessel wall also shows co-staining of TNC and α-SMA due to upregulation of TNC expression. This result suggests that TNC might contribute to the stability of blood flow as an extracellular matrix.

Figure 6.

Double immunofluorescence staining of hepatic stellate cells marker desmin (a) or activated hepatic stellate cells marker α-SMA with TNC in 90% partial hepatectomy rat livers (b). TNC=Tenascin-C

Tenascin-C might contribute to the stability of hepatic blood flow as a matricellular protein

We further detected the relationship between TNC and other ECM by double immunofluorescence staining. FN is the major non-collagenous glycoprotein in the extracellular matrix and basement membrane. We found that FN is well-distributed in the liver in both Sham and 90PH rats [Figure 7]. In the Sham group, TNC shows aggregated and regional distribution, but in 90PH rat livers, TNC is significantly reduced in the hepatic sinusoids. However, all TNCs are co-localized with FN. TNC may act by binding to FN to support or control the structure or blood flow in some areas of hepatic sinuses. Therefore, the loss of TNC from hepatic sinuses in 90PH rats induced by hyperperfusion might contribute to the injury and microcirculation disorder of hepatic sinuses.

Figure 7.

Immunofluorescence co-staining shows the co-localization of TNC with fibronectin in rat livers that underwent Sham operation or 90% partial hepatectomy (90PH). TNC=Tenascin-C

Lipopolysaccharide treatment significantly enhances tenascin-C expression

The bacterial translocation from the gut or the increase of endotoxin level in the portal blood after extended liver resection are important factors contributing to liver injury and PHLF.[17,18] Here, we detected the influence of LPS on TNC expression in the remnant liver after 85PH and 90PH by giving LPS immediately after PH. As shown in Figure 8a, LPS treatment significantly enhanced TNC mRNA expression in the remnant liver and serum TNC level of 85PH and 90PH rats at 12 h after hepatectomy. LPS also increased the hepatic mRNA and serum protein levels of TNC in the Sham group receiving LPS but without hepatectomy. Immunostaining of the liver tissue sections demonstrated that TNC-positive signals within the sinusoids were significantly reduced after LPS treatment, especially in the 90PH group [Figure 8b].

Figure 8.

LPS enhances TNC expression in the remnant livers of rats that underwent Sham operation or 85% and 90% partial hepatectomy (85PH and 90PH) at 12 h after hepatectomy. (a) qRT-PCR detection of TNC mRNA in the rat livers and ELISA detection of serum TNC. (b) Immunofluorescence staining of TNC protein in the rat livers (×200). n=6 for each group. ^$P<0.05 and ^$$P<0.01 LPS versus without LPS treatment; ^&P<0.05 and ^&&P<0.01 versus Sham; ^#P<0.05 versus 85PH. qRT-PCR=Quantitative reverse transcription polymerase chain reaction, TNC=Tenascin-C, LPS=Lipopolysaccharide

LPS treatment obviously increased the serum levels of liver injury makers ALT, AST, TBIL, and induced several-fold upregulation of inflammatory genes such as tumor necrosis factor-α, interleukin-1β, macrophage inflammatory protein-1α and endothelial response genes ET-1, TM-1, eNOS in the remnant livers of 85PH and 90PH rats compared with the Sham group [Figure 9]. These results suggested that obvious inflammation and portal hyperperfusion-induced endothelial dysregulation occurred following both 85PH and 90PH hepatectomy, and LPS administration markedly enhanced the expression of these injury factors. These results further demonstrate that the abnormal distribution and high expression of TNC contribute to extended liver hepatectomy-induced liver hyperperfusion, damage, and failure.

Figure 9.

LPS treatment increases serum levels of liver injury maker ALT, AST, TBIL (a), and induces upregulation of inflammatory genes TNF-α, IL-1β, MIP-1α (b) and endothelial response gene ET-1, TM-1, eNOS (c) in the remnant livers of rats underwent Sham operation or 85% and 90% partial hepatectomy (85PH and 90PH) at 12 h after hepatectomy. n=6 for each group. ^$P<0.05 and ^$$P<0.01 LPS versus without LPS treatment; ^&P<0.05 and ^&&P<0.01 versus Sham; ^#P<0.05 and ^##P<0.01 versus 85PH. TNF-α=Tumor necrosis factor-α, IL-1β=Interleukin-1β, MIP-1α=Macrophage inflammatory protein-1α, ALT=Aminotransferase, AST=Aspartate aminotransferase, TBIL=Total bilirubin

DISCUSSION

We designed a study previously to identify the genes related to PHLF by using rats that underwent different volumes of hepatectomy and omics analysis. The results revealed that the matricellular protein TNC was a candidate molecule. In this study, we demonstrated that TNC mRNA expression in the remnant liver was extensively upregulated in the lethal liver failure model of 90PH rats. Different from a previous study, which suggested that TNC was a mediator of hepatic ischemia and reperfusion injury,[15] we found that TNC may play a function in maintaining the stability of hepatic microvascular structure and blood flow. The hyperperfusion induced by 90PH leads to the loss of a large amount of TNC in the hepatic sinus region, which disrupts the stability of microvascular structure. At the same time, high perfusion in the large blood vessels leads to rapid upregulation of perivascular TNC expression. Therefore, the excessive disturbance of hepatic microcirculation structure results in more severe liver injury and then inhibits liver regeneration in the 90PH liver failure rats.

In this study, we first found that the changes in hepatic TNC expression and distribution correlated with the volume of hepatectomy. 68PH induced a slight upregulation of TNC mRNA expression and a mild reduction of TNC in the hepatic sinus. In 85PH rats, the trend of TNC changes is more pronounced than in 68PH rats. Notably, the TNC mRNA level in 90PH rats was remarkably higher than that in 85PH and 68PH rats. Meanwhile, the TNC protein in the hepatic sinus was significantly reduced. In addition, we found that the TNC mRNA expression in 90PH rat livers was closely correlated with endothelial response genes such as KLF2, ET-1, eNOS, and TM-1. ET-1 is a potent vasoconstrictive peptide, and its upregulation will result in stellate cell activation and liver injury.[19,20] 90PH rats also showed greatly increased serum ALT, AST, and TBIL, and decreased liver regeneration and Ki-67 labeling index compared with 68PH and 85PH rats. These changes directly resulted in liver failure and death of 90PH rats. These indicators demonstrated that 90PH-induced hyperperfusion and liver injury were the main causes of fatal liver failure in 90PH rats. These results suggested that the greatly upregulated hepatic TNC mRNA expression in 90PH rats might be one important factor responsible for the fatal liver failure of 90PH rats.

We also found that the dramatically increased serum TNC correlated with the volume of resected liver after hepatectomy. This is consistent with the previous study, which found that serum TNC levels were markers of postoperative liver failure in patients.[10,21] Plasma TNC-POD14 is very useful for diagnosing postoperative irreversible liver damage early. High serum TNC levels in patients with inflammatory bowel disease have been correlated with disease severity.[22] However, in this study, we cannot differentiate whether the serum TNC is from the secretion of newly synthesized TNC in the liver or the release of previously existing TNC protein from the liver.

TNC is abundantly expressed during embryogenesis and emerges in adult tissues during conditions associated with tissue repair and in pathological conditions such as chronic inflammation, mechanical stress, and tumorigenesis.[23,24] However, for the first time, we found the plentiful existence of TNC in the normal hepatic sinus region. We used several antibodies and TNC knockout mice to prove this finding. The function of TNC in the normal liver has not been reported. The liver is the only organ that can regenerate in mammals. The plentiful and special distribution pattern of TNC in the normal liver may be related to the maintenance of vascular structure in the hepatic sinus and the function of hepatic nonparenchymal cells. A recent study with Ambystoma mexicanum suggests that the maintenance of a TNC-enriched ECM environment promotes the release of muscle satellite cells and supports a high muscle regenerative capacity.[25] Therefore, we assumed that the decrease in TNC distribution in hepatic sinuses after hepatectomy represented a kind of breakdown of the hepatic sinusoidal microenvironment. However, the increase of TNC around hepatic veins after hepatectomy may be related to the migration of HSC or inflammatory immune cells.

TNC is highly expressed in tumors, and during chronic inflammation and bacterial or viral infection, marking it as a potential factor in the control of the immune system. A cross-interaction between TNC and chemokines has been assumed.[26] It was recently shown that tumor-infiltrating leukocytes are trapped in the stroma by binding TNC to CXCL12, leading to tumor surveillance.[27] TNC may also have a chemotactic effect on inflammatory immune cells, such as mononuclear macrophages and neutrophils. This inflammatory response will further aggravate liver damage in 90PH rats.

Although some studies believe that HSC is the main cell source of TNC expression in liver fibrosis, macrophages and fibroblasts can also express TNC.[28] The cell source of TNC expression in extended hepatectomy has not been determined. In this study, the RNAscope assay clearly showed that TNC was expressed by a specific group of cells after hepatectomy, especially in 90PH rats. These cells are not evenly distributed and are mainly distributed around the central vein, portal vein, and the intermediate area. Double staining showed that some TNC-positive cells co-expressed mononuclear macrophage marker CD68, and some TNC co-localized with HSC marker desmin. We did not find that TNC co-localized with α-SMA, and we also found that there were very few activated stellate cells (α-SMA-positive cells) in the remnant liver after hepatectomy. We speculate that a special type of fibroblast is responsible for the TNC expression after hepatectomy. However, because TNC is a secreted matricellular protein and lacks a specific cell surface marker, it is technically difficult to isolate the source cells of TNC.

Interestingly, we found that TNC binds closely to the ECM protein FN in the rat liver. The binding of other ECM proteins, including collagens, elastins, and proteoglycans, to FN fibrils facilitates the maturation and tissue specificity of the ECM. These connections provide key mechanical and chemical signals to induce specific cell behaviors.[29,30] Here, we highlight specific interactions that may play important roles in mechanochemical signaling of FN fibrils. Our immunofluorescence staining of the rat liver shows that FN is a prominent component of hepatic extracellular matrix, and its fluorescence intensity and distribution did not show an obvious change after different volumes of hepatectomy. For the first time, we found that the distribution of TNC and FN is highly colocalized in the rat liver. This highlighted the specific interactions that might play important roles in the mechanochemical signaling of FN and TNC. Since TNC is mainly distributed axially along the hepatic sinusoid or around the central vein, it may be associated with stabilizing the liver sinus structure or as a migration track of some cells. In 90PH rat livers, TNC in hepatic sinuses is significantly reduced, which may disrupt the stability of the hepatic sinuses blood flow structure. The severe loss of TNC in the hepatic sinuses of 90PH rats might contribute to PHLF. The modules for TNC and FN binding need further study in the future.

Increased portal LPS and bacterial translocation because of intestinal barrier repair in post-hepatectomy injuries have been reported, contributing to PHLF.[17,31] Systemic translocation of indigenous bacteria through a dysfunctional intestinal barrier contributes to the early post-liver transplantation infectious complications. In this study, we found that LPS administration immediately following liver resection greatly increased TNC expression in the rat liver. TNC upregulation was correlated with an increased inflammatory response as well as the dysregulation of endothelial function after LPS treatment. In fact, a growing body of evidence suggests that TNC activates TLR4 signaling, the important signal pathway of LPS, resulting in a greater amount of cytokine secretion, more TNC synthesis, and creating a positive feedback loop to augment inflammation. This suggested that TNC also has a strong impact on the inflammatory response, showing a potential role in PHLF. Thus, this work provides the groundwork for the further investigation of TNC as a potential target for therapy against PHLF.

CONCLUSION

In summary, by comparing the differential expression of hepatic mRNA transcripts and protein between rats that underwent different volumes of hepatectomy, we found that the matricellular protein TNC mRNA was tremendously upregulated in the remnant liver of the 90PH lethal liver failure model. Meanwhile, a robust disturbance of hepatic TNC protein distribution in 90PH rat livers occurred with an obvious reduction in hepatic sinusoids but an increase around large blood vessels, which were consistent with the severe liver injury of the 90PH rats. It was also found for the first time that treatment with the endotoxin LPS tremendously exacerbated the above changes of TNC. This provides a potential target for the treatment of liver failure. However, due to the complexity of the TNC structure and function, and more studies aimed at better understanding the biological complexity of TNC, combined with technological advancements, are needed to generate improved therapeutic modalities to inhibit TNC in both experimental and clinical settings.

Author contributions

Xinlan Ge performed the animal surgery and pathological experiments. Ke Pan conducted the gene and protein expression experiments. Qiyue Yang carried out data collection and analysis. Lijie Gao provided technical guidance. Chonghui Li designed the study and wrote the manuscript. Xinlan Ge and Ke Pan contributed equally to this study.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.