Although previous evidence indicates close involvement of CD147 in the pathogenesis of liver fibrosis, the underlying molecular mechanisms and its therapeutic value remain largely unknown. In the present study, we investigated the biological roles of CD147 in liver fibrosis and assessed its therapeutic value as a target molecule in the CCl4-induced liver fibrosis mouse model. We found that CD147 was highly expressed in both hepatocytes and SECs (sinusoidal endothelial cells) in fibrotic liver tissues. Additionally, it was significantly associated with the fibrosis stage. TGF-β1 (transforming growth factor β1) was found to be mainly responsible for the up-regulation of CD147. Bioinformatic and experimental data suggest a functional link between CD147 expression and VEGF-A (vascular endothelial growth factor A)/VEGR-2 (VEGF receptor 2) signalling-mediated angiogenesis in fibrotic liver tissues. Furthermore, we observed that the CD147-induced activation of the PI3K (phosphoinositide 3-kinase)/Akt signalling pathway promotes the production of VEGF-A in hepatocytes and expression of VEGFR-2 in SECs, which was found to enhance the angiogenic capability of SECs. Finally, our data indicate that blocking of CD147 using an mAb (monoclonal antibody) attenuated liver fibrosis progression via inhibition of VEGF-A/VEGFR-2 signalling and subsequent amelioration of microvascular abnormality in the CCl4-induced mouse model. Our findings suggest a novel functional mechanism that CD147 may promote liver fibrosis progression via inducing the VEGF-A/VEGFR-2 signalling pathway-mediated cross-talk between hepatocytes and SECs. New strategies based on the intervention of CD147 can be expected for prevention of liver fibrosis.
Previous studies showed that CD147 played multiple functional roles in tumour metastasis, wound healing and tissue remodelling. Recent data has indicated that CD147 is closely involved in multiple fibrotic diseases, including liver fibrosis.
In the present study, we first demonstrated the pro-angiogenesis role of CD147 during liver fibrosis progression. CD147 may promote sinusoidal formation through enhancement of VEGF-A secretion from hepatocytes and VEGFR-2 expression in sinusoidal endothelial cells.
Our study provides the rationale for the therapeutic trials of blocking CD147-induced angiogenesis for the treatment of liver fibrosis.
Liver fibrosis is a common feature of almost all chronic liver diseases arising from infectious, inflammatory or toxic causes . Progressive liver fibrosis eventually leads to cirrhosis, which is characterized pathologically by the formation of regenerative parenchyma nodules. Most previous studies on liver fibrosis have been mainly focused on the fundamental roles of HSCs (hepatic stellate cells), which were recognized as the main matrix-producing cells and as being responsible for excessive deposition of extracellular matrix components during this process . Recently, increasing evidence has shown that enhanced intrahepatic angiogenesis is associated with faster fibrosis progression and thus has been identified as a crucial contributor to thefibrogenesis [3,4]. More significantly, angiogenesis inhibitors have been evaluated in the prevention and treatment of fibrosis, and have been expected as potential potent anti-fibrosis drugs [3,5,6].
Angiogenesis involves a tightly regulated cellular and molecular network. Numerous soluble pro-angiogenic growth factors, such as VEGF (vascular endothelial growth factor) and angiopoietin, are identified to be the most important mediators of this process. It has been reported that the expression of these pro-angiogenic growth factors is significantly increased during liver fibrogenesis, and the suppression of their signalling cascade markedly attenuates liver fibrosis [7,8]. Despite the large amount of studies describing the production of these pro-angiogenic growth factors in liver fibrosis tissues, few previous studies have attempted to elucidate their regulatory mechanisms. A better understanding of these mechanisms could provide more insight into the pathological process of liver fibrosis and facilitate the establishment of new preventive and therapeutic strategies.
CD147, a transmembrane glycoprotein, is widely expressed on the surface of various cells, including cancer cells and activated lymphocytes, as well as epithelial cells. As a metalloprotease inducer, CD147 has been proven to be involved in multiple biological processes, such as immune response, tumour progression and tissue repair. Other groups have demonstrated that CD147 also plays a critical role in the angiogenic function of endothelial cells [9,10]. Moreover, it has been reported that CD147 regulates angiogenesis by increasing the activity of the VEGF/VEGFR (VEGF receptor) axis in malignant and non-malignant diseases [11–13]. Increasing evidence has shown that the biological activity of CD147 is linked to fibrotic diseases, including liver fibrosis [14–18]. However, the mechanisms underlying pro-fibrotic functions and therapeutic implication of CD147 remain to be determined.
The present study aims to validate whether CD147 plays a key role in promoting pro-angiogenic growth factor-mediated liver fibrosis progression and to explore the underlying molecular mechanisms and the therapeutic value of anti-CD147 antibodies in liver fibrosis.
MATERIALS AND METHODS
Cell lines, cell transfection and animal model
Two immortalized human hepatocyte cell lines QSG-7701 and L02 were obtained from the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China), and human SECs (sinusoidal endothelial cells) were obtained from Sciencell Research Laboratories. The overexpression and knockdown of target genes is described in the Supplementary Experimental Procedures. Fifty-nine human liver tissue samples (eight normal and 51 fibrotic tissues) were obtained from the Department of Pathology in the Eastern Hepatobilliary Surgery Hospital, Shanghai, China, between 2011 and 2012, with signed informed consents. One experienced pathologist who was unaware of clinical and genetic data reviewed all biopsies for the fibrosis stage using the METAVIR score. The CCl4-induced liver fibrosis mouse model, which is widely used in periportal and pericentral angiogenesis studies , was established. The detailed approaches are described in the Supplementary Experimental Procedures. All animals were treated humanely according to protocols approved by the Fourth Military Medical University Committee on Animal Care and Use.
H&E (haematoxylin and eosin), Masson's trichrome and Sirius Red staining
Intrahepatic collagen deposition was determined by using H&E, Masson's trichrome and Sirius Red staining as described previously [3,5,6]. Briefly, liver tissues were fixed in 10% (w/v) buffered formalin, embedded in paraffin and sectioned at 5 μm thickness. Sections were stained with H&E, Masson's trichrome or Sirius Red solution. The Sirius Red staining was quantified on the basis of the optical density per standard area in each section. Six randomly chosen views of each section from a single animal were selected, and the optical density for a given area was measured and averaged.
Microarray data series collection
Five public microarray data series (GSE25097, GSE14323, GSE33258, GSE11536 and GSE14520) were obtained from the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). The characteristics of the five microarray series are provided in the Supplementary Experimental Procedures.
qRT-PCR (quantitative real-time reverse transcription–PCR)
RNA extraction, cDNA synthesis and qRT-PCR were performed as described previously . Primer sequences are listed in Supplementary Table S1.
IHC (immunohistochemistry), IF (immunofluorescence) and WB (Western blotting)
Liver tissues and cell lines were processed for IHC, IF and WB as described previously . Antibodies used in the present study are listed in Supplementary Table S2. The IHC scoring method is provided in the Supplementary Experimental Procedures.
Human VEGF Quantikine ELISA kit (R&D Systems) was used following the directions of the manufacturer. Conditioned medium (200 μl) was collected from triplicate samples.
Cell viability analysis
Cell viability was monitored using the MTT (Sigma–Aldrich) method as described previously .
Wound healing assay
Scratch wounds were created in monolayer cells using a sterile pipette tip. After 24 h of incubation, cells were observed using an inverted Televal microscope (Olympus). Wound width was estimated at 0 and 24 h using NIH ImageJ software.
Endothelial cell tube formation assay
An in vitro 2D Matrigel endothelial capillary-like structure formation assay was performed as described previously . Briefly, SECs were seeded on top of fibrin gel in serum-free medium. After incubation for 24 h, the formation of capillary-like structures was photographed, and the measurement of the capillary tube length was carried out using ImageJ software.
SEC motility was measured using a modified Boyden chamber assay. Cells were seeded on the upper chamber of the insert, and 10 ng/ml VEGF165 was placed in the lower chamber wells as a chemoattractant. After incubation for 48 h, cells were fixed, stained with Gentian Violet (Sigma–Aldrich), and counted under a microscope.
The collagen content of the liver tissues was quantified using a hydroxyproline detection kit (Jiancheng Institute of Biotechnology) according to the manufacturer's instructions. Briefly, 30–100 mg of wet liver tissues was lysed at 95°C for 20 min. Following adjusting the pH to within the range 6.0–6.8, the lysates were incubated with reagents which were provided by the kit. Finally, the absorbance of the reaction products was detected at 550 nm light wavelength. Hydroxyproline concentration was calculated using a hydroxyproline standard and normalized by the weight of mouse liver tissue.
Differences between normally distributed repeatedly measured data were analysed with repeated-measurement ANOVA, whereas abnormally distributed repeatedly measured data were analysed with Friedman's test. Student's t test or one-way ANOVA was used to analyse the difference of normally distributed continuous variables, whereas a Mann–Whitney U test and a Kruskal–Wallis H test were applied for abnormally distributed continuous variables. Relationships between measured variables were tested by Pearson correlation analysis. P<0.05 was considered to be statistically significant.
CD147 expression was up-regulated and associated significantly with liver fibrosis
We first assessed the expression level of CD147 in human liver tissues. Our data showed that liver fibrosis tissues had a significantly increased CD147 expression at the mRNA and protein levels when compared with normal liver tissues (Figures 1A and 1B). The increased mRNA expression was confirmed by public GEO microarray data analysis (Supplementary Figure S1). Moreover, immunohistochemical staining revealed a specific distribution pattern of CD147 in fibrotic liver tissues. The intensive staining was clearly observed not only on the hepatocellular membrane, but also in the perisinusoidal space (Figure 1C). In addition, the expression level of CD147 was significantly increased in both the hepatocyte membrane and perisinusoidal space when compared with normal tissues (Figure 1D). Liver perisinusoidal space primarily contains three types of non-parenchymal cells: HSCs, SECs and Kupffer cells. To explore further the cell-specific localization of CD147 in the perisinusoidal space, immunological detection was performed using different cellular markers [α-SMA (α-smooth muscle actin), for HSCs, CD31 for SECs and CD68 for Kupffer cells]. As shown in Figures 1(E) and 1(F), CD147 exhibited a clear co-localization with the SEC marker CD31, but not with α-SMA or CD68.
CD147 expression is up-regulated in fibrotic liver tissues and located on the plasma membrane of hepatocytes and SECs
We estimated CD147 expression levels in human and mouse liver tissues at different fibrotic stages. Our results showed that elevated CD147 expression was closely correlated with the liver fibrosis stage. As shown in Figures 1(G) and 1(H), human liver tissues at an advanced stage of fibrosis exhibited a notably increased CD147 expression compared with those at an early stage of fibrosis. The increased mRNA expression was confirmed by public GEO microarray data analysis (Supplementary Figure S2). Similar results were also obtained in a CCl4-induced mouse liver fibrosis model, indicating that mouse liver fibrosis progressed with CCl4 treatment time, during which CD147 expression was gradually increased (Figures 1I and J). Since CD147 was expressed on both hepatocytes and SECs, we separately analysed the correlation between CD147 expression on these two distinct cell types and liver fibrosis stage. Remarkably, CD147 expressed on each type of cell was significantly associated with the liver fibrosis stage in both human and mouse tissue samples (Supplementary Figure S3). Together, these results strongly suggest that CD147 expression was positively correlated with the liver fibrosis stage.
Up-regulation of CD147 expression was mainly mediated by TGF-β1 (transforming growth factor β1) in both hepatocytes and SECs of liver fibrosis tissues
To explore the underlying mechanism by which CD147 was up-regulated in liver fibrosis tissues, we first evaluated the relationship between the expression of CD147 in hepatocytes and SECs and found that there was a strong positive correlation between them in both human and mouse liver fibrosis tissues (both P<0.001) (Supplementary Figure S4). On the basis of this finding, we propose that there must be certain key factors in the liver fibrotic microenvironment which are responsible for CD147 up-regulation on both cells. A previous study found that TGF-β1 up-regulates the CD147 expression through Slug . We therefore analysed the relationship between mRNA expression of CD147 and TGF-β1 in human liver fibrosis tissues using three public GEO microarray datasets. As expected, there was a strong positive correlation between them (Figure 2A). Moreover, our results showed that exogenous TGF-β1 treatment led to up-regulation of CD147 at the mRNA and protein level both in human hepatocytes and SECs in a dose- and time-dependent manner, demonstrating the modulating effect of TGF-β1 on CD147 expression (Figures 2B–2E). These results suggest that TGF-β1 might be a key factor responsible for increased CD147 expression in these two types of cells.
Up-regulation of CD147 expression is mainly mediated by TGF-β1 in both hepatocytes and SECs
CD147 expression was correlated with VEGF-A/VEGR-2 signalling-mediated angiogenesis during liver fibrosis progression
We explored whether CD147 was involved in intrahepatic angiogenesis during liver fibrogenesis by comparing the expression profile of CD147 with that of 165 ARGs (angiogenesis-related genes) derived from the public microarray dataset GSE14520. As shown in Figure 3(A), a coefficient density plot of Pearson's correlations between CD147 and ARGs indicated a significant shift to the right when compared with that between CD147 and the overall genome, suggesting an excess of positive correlation between ARGs and CD147. Subsequently, the significant positive correlation between mRNA expression of CD147 and angiogenesis marker genes, such as CD31, CDH5, CD105 and vWF (von Willebrand's factor), was also observed in microarray dataset GSE14520 (Supplementary Figure S5). Moreover, immunohistochemical staining analysis also confirmed that CD147 expression was significantly correlated with the CD31-positive staining area (r=0.635, P<0.001) (Figure 3B).
CD147 expression is correlated with intrahepatic angiogenesis
Next, we explored the potential signalling pathways involved in CD147-induced intrahepatic angiogenesis using KEGG pathway analysis. Our results indicated that ARGs, which significantly correlated with CD147 (cut-off P<0.001) in public dataset GSE14520 were mostly enriched in the VEGF signalling pathway (P=3.38×10−60) (Supplementary Table S3). Further analysis indicated that VEGF family member VEGF-A, but not VEGF-B, VEGFR-2 (KDR/Flk-1), but not VEGFR-1 (Flt-1), and VEGFR-3 (Flt-4) exhibited significant expression correlation with CD147 (r=0.335 and 0.227 respectively; both P<0.001) (Figure 3C and Supplementary Figures S6A–S6D). In addition, our data also showed that there was no significant correlation between mRNA expression of CD147 and angiopoietin 1 or 2, two members of another key angiogenesis regulator family (Figures S6E and S6F). Finally, immunochemical staining analysis confirmed our findings, indicating that VEGF-A expression was remarkably increased with liver fibrosis progression (Figures 3D and 3E). As expected, the VEGF-A expression score was significantly correlated with CD31-positive vessel number (Figure 3F). Furthermore, significant correlation was also observed between VEGF-A and CD147 expression in liver fibrosis tissues (Figure 3F). All of these data indicate that the VEGF-A/VEGFR-2 signalling pathway contributes to CD147-induced intrahepatic angiogenesis, thus leading to liver fibrosisprogression.
CD147 promoted VEGF-A expression and secretion via the PI3K (phosphoinositide 3-kinase)/Akt signalling pathway in hepatocytes
Considering that expression of VEGF-A was mainly in hepatocytes and correlated with CD147 expression, we sought to determine whether CD147 was able to modulate the production of VEGF-A in hepatocytes. As shown in Figure 4(A) and Supplementary Figure S7, CD147 expression was significantly changed in the cell lines indicated. Consequently, our results indicate that VEGF-A expression was obviously decreased at both the mRNA and protein level in hepatocytes with CD147 knockdown and vice versa in cells with CD147 overexpression (Figures 4A and 4B). Moreover, ELISA assays showed that CD147 up-regulated the soluble VEGF-A level in the culture supernatant of hepatocytes (Figure 4C).
CD147 promotes VEGF-A production via the PI3K/Akt signalling pathway in hepatocytes
Several previous studies have reported the activation of the PI3K/Akt signalling pathway by CD147 and the importance of this pathway in VEGF-A production [9,24]. We therefore hypothesized that CD147 might induce VEGF-A expression in hepatocytes by the PI3K/Akt signalling pathway. To validate our hypothesis, we analysed the activity of the PI3K/Akt signalling pathway in hepatocytes with CD147 knockdown or overexpression. As shown in Figure 4(D), CD147 expression remarkably promoted Akt phosphorylation and VEGF-A expression. We further applied an Akt-specific inhibitor, LY294002, to determine whether this pathway was responsible for the CD147-mediated VEGF-A production. As expected, phosphorylated Akt in hepatocytes was significantly decreased after exposure to LY294002, and the VEGF-A expression and secretion were inhibited in a dose-dependent manner (Figures 4E and 4F, and Supplementary Figure S8). All of these results suggest that CD147 promoted the expression and secretion of VEGF-A via the PI3K/Akt signalling pathway in hepatocytes.
CD147 enhanced VEGF-A-mediated proliferation and migration of SECs through up-regulating VEGFR-2 expression
We explored the functional role of CD147 in SECs and found that the in vitro proliferation capacity of SECs was considerably increased when CD147 was overexpressed and was decreased when CD147 was knocked down (Figure 5A). In agreement with the data from the proliferation analysis, cell migration and tube formation assays demonstrated a significantly higher pro-angiogenesis activity in CD147-overexpressing SECs (P<0.05) (Figures 5B and 5C). In vitro invasion assays showed that up-regulation of CD147 resulted in an enhanced chemotaxis response of SECs to recombinant hVEGF165 (human VEGF165) and induced morphologically typical changes indicating higher invasiveness (Figure 5D). Furthermore, we also found that the up-regulation of CD147 resulted in a significant increase in VEGFR-2 expression in SECs. However, expression of VEGFR-1 was not affected (Figure 5E). Interestingly, the expression of VEGFR-2 as well as the phosphorylation of Akt was up-regulated in SECs with CD147 overexpression. Moreover, we also found that inhibition of the PI3K/Akt signalling pathway by the PI3K-specific inhibitor LY294002 significantly suppressed the expression level of VEGFR-2 (Figure 5F). In addition, the VEGFR-2-specific inhibitor Ki8751 significantly suppressed cellular proliferation, migration, tube formation and chemotaxis activity of CD147-overexpressed SECs (Figure 5G). These results indicate that CD147 overexpression might augment in vitro proliferation and migration in SECs via up-regulating VEGFR-2 expression.
CD147 promotes VEGF-A-mediated proliferation, migration tube formation and chemotaxis activity of SECs through up-regulating VEGFR-2 expression
Anti-CD147 antibody attenuated liver fibrosis progression via inhibiting VEGF-A/VEGFR-2-mediated angiogenesis
We extended our study to evaluate the anti-fibrotic effect of blocking CD147 with a specific mAb  in the CCl4-induced mouse fibrosis model (Supplementary Figure S9). As shown in Figure 6(A) and Supplementary Figure S10, the blocking of CD147 clearly attenuated collagen formation in the mouse liver fibrosis model. WB and hydroxyproline assay confirmed that the content of collagen was significantly decreased in liver tissues from the mice treated with anti-CD147 mAb when compared with those treated with anti-mouse IgG mAb (Figure 6B and 6C). Furthermore, we detected CD31 staining in liver tissues from different groups and found that there was a significant decrease in CD31-positive staining area in anti-CD147 mAb treatment groups when compared with anti-mouse IgG mAb groups (P<0.05) (Figure 6D). Next, we detected the mRNA expression of VEGF-A and VEGFR-2 in the mouse liver fibrosis models. Our results showed that treatment with anti-CD147 mAb significantly decreased VEGF-A and VEGFR-2 mRNA expression (Figure 6E). All of these data clearly indicate that the CD147/VEGF-A/VEGFR-2 pathway greatly contributes to liver fibrosis progression, and that this pathway is a potential therapeutic target in liver fibrosis treatment (Figure 6F).
Blocking of CD147 attenuated liver fibrosis progression via inhibiting VEGF-A/VEGFR-2 signalling-mediated intrahepatic angiogenesis
In the present study, we confirmed the up-regulation of CD147 expression, which was mainly mediated by TGF-β1, on the plasma membrane of both hepatocytes and SECs in fibrotic liver tissues. We found further that up-regulated CD147 was significantly associated with liver fibrosis stage. We suggest a new functional mechanism, in which CD147 may promote liver fibrosis progression via inducing intrahepatic angiogenesis mediated by VEGF-A/VEGFR-2 signalling. In addition, we demonstrated the anti-fibrotic effect through blocking CD147-mediated angiogenesis in the CCl4-induced mouse fibrosis model.
Hepatocytes have been demonstrated to be major sources of CD147 in fibrotic liver tissues [17,26]. A recent study by Calabro et al.  has provided convincing evidence for cellular localization of CD147. They found that CD147 was abundantly expressed on hepatocytes, but not HSCs, which is consistent with our results. Moreover, in the present study, our data also clearly show that CD147 is highly expressed on hepatic sinusoids. However, the sinusoidal expression of CD147 and its relationship with liver fibrosis has been scarcely investigated. With respect to the cellular location of CD147 in hepatic sinusoids, we found that CD147 appeared to mainly co-localize with the SEC marker CD31, but not with the HSC marker α-SMA or the universal marker of leucocytes (including monocytes/macrophage lineage cells, neutrophil granulocytes and T-cells) CD68 , indicating a SEC-dominant localization. There are three major cell types existing in and around hepatic sinusoids, namely SECs, HSCs, and Kupffer cells. In addition, circulating leucocytes can reside in hepatic sinusoids. It has been reported that quiescent HSCs represent 5–8% of the total sinusoidal cells  and only approximately 30% of them are activated in fibrotic livers . Meanwhile, the proportion of activated CD8+ T-cells in fibrotic livers is no more than 2% . Compared with the low number of HSCs and CD8+ T-cells, SECs are the major cell population in number (approximately 70%) among liver sinusoidal cells  and contribute 26% to the total cell membrane surface in livers . Thus SEC-dominant localization of CD147 is much more rational and consistent with its extensive IHC staining in the perisinusoidal space. Likewise, CD147 has also been reported to be expressed in vascular endothelial cells from many organs, including heart, lung, brain and kidney [33–35]. The synchronistic localization in hepatocytes and SECs suggests key functional roles of CD147 in reciprocal interactions between these two cells to facilitate liver fibrosis.
Hepatic microvascular abnormalities, including capillarization of hepatic sinusoids and pathological angiogenesis, have been considered to be important mediators of liver fibrosis [36,37]. It has been found that hepatic microvascular abnormalities notably correlate with the degree of liver fibrosis. Moreover, hepatic fibrosis progression could be prevented by drugs that specifically inhibit microvascular abnormality . Hepatic microvascular abnormality has been suggested to begin with the activation of SECs within hepatic sinusoids, and is characterized by a significant increased number of CD31-positive blood vessels located within the liver lobules . Data from other groups have demonstrated CD147-induced angiogenic responses by regulating activities of endothelial cells in tumours and non-malignant diseases [9–11]. In the present study, we found that CD147 expression was correlated with an intrahepatic CD31-positive staining area during liver fibrosis progression, which gave rise to a hypothesis that intrahepatic microvascular abnormality in fibrotic liver may be attributable to synchronistic elevated expression of CD147 on hepatocytes and SECs. This issue was clearly confirmed by our bioinformatic and experimental analyses, exhibiting a functional link between CD147 and intrahepatic microvascular abnormality in liver fibrosis, although further details still need to be elucidated.
Hepatocytes, as the dominant parenchymal cells in the perisinusoidal situation, release various growth factors and cytokines to SECs, thus contributing to many biological and pathological processes, such as hepatic injury , inflammation , angiogenesis  and regeneration . Among these factors, two key pro-angiogenic mediators, VEGF-A and its receptor VEGFR-2, are considered to be the main players in the cross-talk between hepatocytes and SECs . Researchers have long acknowledged the prerequisite role of the VEGF-A and VEGFR-2 interaction in liver fibrogenesis, and observed that blocking their interaction using specific neutralizing antibodies significantly attenuates liver fibrosis progression . However, molecular regulatory mechanisms of this interaction have not been demonstrated. It has been widely accepted that CD147-mediated cell–cell interactions exist in various tissue types, such as cancer , nervous  and immune . Our results which indicated the synchronistic expression of CD147 in both hepatocytes and SECs suggested a regulatory role for CD147 in hepatocyte–SEC interactions.
Previous studies have demonstrated that the interaction of CD147 with integrin β1 MIDAS (metal-ion-dependent adhesion site) motif activates the downstream PI3K/Akt signalling pathway in HCC (human hepatocellular carcinoma) cells, and subsequently enhances the invasion and proliferation capacities of HCC cells . Moreover, CD147 stimulates VEGF production via the PI3K/Akt pathway in fibroblast cells . In the present study, we found that the exogenous expression of CD147 in hepatocytes promoted the expression and secretion of VEGF-A via the PI3K/Akt pathway. It has been reported that CD147 could lead to the activation of the PI3K/Akt pathway in activated human umbilical vein endothelial cells, therefore playing an important role in angiogenesis . We then hypothesized whether similar mechanism existed in SECs to mediate CD147-induced up-regulation of VEGFR-2 and its angiogenic capability. As expected, the expression of VEGFR-2 as well as phosphorylation of Akt was up-regulated in SECs overexpressing CD147. Moreover, inhibition of the PI3K/Akt signalling pathway using the PI3K-specific inhibitor LY294002 significantly suppressed the VEGFR-2 expression level. Our data therefore suggest that the PI3K/Akt signalling pathway mediates the effect of CD147 on VEGFR-2 expression in SECs. The present study reveals a novel mechanism in that the CD147-induced activation of PI3K/Akt signalling in both hepatocytes and SECs synergistically promotes microvascular abnormality in the process of liver fibrosis.
It has been revealed that TGF-β1, the most potent profibrogenic factor, is indispensable for angiogenesis . Its genetic deficiency leads to the defective vasculogenesis and thus causes 50% embryonic lethality . Additionally, antagonistic peptides of TGF-β1 have also exhibited anti-angiogenesis activity in vivo . Although the pro-angiogenesis activity of TGF-β1 is clearly demonstrated, the molecular events underlying its pro-angiogenesis effects have not been completely elucidated. Several studies have revealed that TGF-β1 induces the expression of VEGF and/or other angiogenic factors in epithelial or other cell types [51,52]. Furthermore, TGF-β1 has been reported to regulate CD147 expression in many cell types, including hepatocytes . Our data further indicate that TGF-β1 was responsible for the overexpression of CD147 on both hepatocytes and SECs, which clearly contributes to a synergistic effect between these two neighbouring cells.
Anti-angiogenesis has been considered to be a useful strategy for antifibrotic therapy . Since CD147 plays such an important role in hepatic microvascular abnormality and liver fibrosis, we then set out to ascertain whether anti-CD147 therapy would also be beneficial to inhibit the disease progression in an animal liver fibrosis model. Notably, blocking CD147 by an antibody exhibited an anti-fibrotic effect on CCl4-induced liver fibrosis. Meanwhile, we found that mRNA expression of VEGF-A and VEGFR-2 was inhibited and the intrahepatic CD31-positive staining area was reduced, which implied that the microvascular abnormality induced by CD147 was attenuated. According to previous studies, capillarization of SECs was permissive for HSC activation and fibrosis [53,54]. Thus, except for the pro-angiogenesis effect of CD147 during liver fibrosis progression, we cannot rule out the possibility that CD147 induces the capillarization of SECs to attenuate the HSC activation status . In contrast with our results, a recent study demonstrated that anti-CD147 antibody treatment enhanced liver fibrosis using a similar fibrosis model . The discrepancies between their and our results could possibly be explained by differences in either the anti-CD147 antibody and treatment protocol or the liver fibrosis mouse model used.
In summary, our findings provide strong evidence for the pro-angiogenesis role of CD147 during liver fibrosis progression. CD147 may promote the formation of sinusoid through increasing VEGF-A secretion from hepatocytes and VEGFR-2 expression in SECs. Given that CD147 is the upstream regulator of VEGF signalling and is involved in multiple steps of the fibrosis process, it is desirable that targeting CD147 may prove to be effective treatment in suppressing liver fibrosis.
Jinliang Xing and Hongxin Zhang contributed to the concept and design of the study. Kai Qu, Zhaoyong Yan, Jing Zhang, Qichao Huang, Ping Qu, Xinsen Xu, Peng Yuan and Xiaojun Huangcontributed to the generation, collection, assembly, analysis and/or interpretation of data. Kai Qu, Zhaoyong Yan, Jing Zhang, Yongping Shao and Jinliang Xing contributed to the drafting or revision of the paper. Jinliang Xing, Hongxin Zhang and Chang Liu approved the final version of the paper.
This work was supported by Program for New Century Excellent Talents in University, International S&T Cooperation Program of China [grant number S2011GR0239] and National Natural Science Foundation of China [grant numbers 81071896 and 81270505].
human hepatocellular carcinoma
haematoxylin and eosin
hepatic stellate cell
quantitative real-time reverse transcription–PCR
sinusoidal endothelial cell
α-smooth muscle actin
transforming growth factor β1
vascular endothelial growth factor
These authors contributed equally to this work.