Engrafted PD-MSC decreases the hepatic vein diameter and improves the liver regeneration in the CCl4-injured rat model
Engrafted PD-MSCs were detected in CCl4-injured liver tissues of the PD-MSC transplantation (PD-MSC Tx) group by immunofluorescent staining of a human-specific cytoplasmic marker Stem121. The expression of vWF was detected in hepatic vessels of the control group but not in the NTx group. Interestingly, engrafted PD-MSCs were distributed around the hepatic vein, and vWF expression was higher in the PD-MSC Tx group than in the NTx group (Fig. 1A). In addition, the expression level of human Alu mRNA was significantly elevated in the PD-MSC Tx group compared to that in the non-transplantation (NTx) groups at 1 and 2 weeks; however, the expression was considerably reduced by 3 weeks (p < 0.05) (Fig. 1B).
Compared to those in the control group, CCl4-injured livers in the NTx groups showed severe cirrhosis and increased inflammatory cell infiltration and localization around hepatic vessels. Interestingly, compared to that in the NTx group, the level of cirrhosis in the Tx group was visibly reduced (Fig. 1C). In addition, we investigated whether hepatic vein structures were changed by CCl4-induced liver damage. In general, it is well known that portal hypertension is increased in cirrhotic liver via portal vein thrombosis [23, 24]. As shown in Fig. 1C, the hepatic vein diameter was considerably larger in the CCl4-treated group than in the control group. Furthermore, the diameter of the hepatic vein gradually widened in a time-dependent manner (1, 2 and 3 weeks). However, the hepatic vein diameter was significantly smaller in the PD-MSC Tx group than in the NTx groups at 1, 2 and 3 weeks (Fig. 1D, p < 0.05). Furthermore, the structure of hepatic vessels in the PD-MSC Tx groups were similar to that in the control group, whereas the NTx group showed a very irregular structure of hepatic vessels at 1, 2 and 3 weeks (Supplementary Fig. 2).
Furthermore, we investigated the expression levels of proteins related to hepatic markers, such as albumin (ALB) and hypoxia-inducible factor-1alpha (HIF-1α) to confirm the effect of PD-MSCs on hepatic regeneration in the CCl4-injured rat model. The expression level of ALB at 1, 2 and 3 weeks was significantly lower in the NTx groups than in the control group. However, the expression level of ALB at 1 and 2 weeks was significantly increased in the PD-MSC Tx group compared with that in the NTx group, but there was no significant difference in ALB expression at 3 weeks between the groups (p < 0.05) (Fig. 1E). Compared to the levels in the control group, HIF-1α levels remained decreased up to 3 weeks after injection of CCl4. Interestingly, the expression levels of HIF-1α at 1, 2 and 3 weeks were significantly higher in the PD-MSC Tx group than those in the NTx group (p < 0.05) (Fig. 1F). In our previous reports, increased HIF-1α expression by PD-MSC Tx triggered liver regeneration via an autophagic mechanism [20]. Based on these data, we hypothesized that increased HIF-1α expression, mediated by PD-MSC Tx, could increase VEGF expression, which is a downstream factor of HIF-1α. These results suggest that engrafted PD-MSC promotes the recovery of expanded hepatic vein and regeneration in the CCl4- injured rat model.
PD-MSCs induce angiogenesis and tube formation of endothelial cells in both in vivo and ex vivo co-culture systems
To confirm the effect of PD-MSCs on hepatic angiogenesis, we determined the protein expression profiles of angiogenic markers such as vWF, endoglin, VEGF and VE-cadherin. The expression levels of vWF and endoglin were decreased up to 3 weeks after CCl4 injection compared to the expression levels in the control group. However, the levels of vWF and endoglin at 1 week were significantly higher in the PD-MSC Tx group than in the NTx group (p < 0.05) (Fig. 2A and B). Additionally, at 1 and 2 weeks, the levels of VEGF, which is a key factor in generating new blood vessels after injury, were significantly higher in the PD-MSC Tx group than in the NTx group (p < 0.05) (Fig. 2D). In general, VE-cadherin regulates blood vessel formation by modulating VEGF receptor in endothelial cells [25]. Therefore, we confirmed the protein expression of VE-cadherin in the CCl4-injured rat model. Compared to the expression levels in the control group, the expression level of VE-cadherin at 1, 2 and 3 weeks was suppressed after CCl4 injection. However, VE-cadherin expression levels at 1, 2 and 3 weeks were dramatically higher when PD-MSCs were transplanted (PD-MSC Tx group) than when these cells were not transplanted (NTx group) (p < 0.05) (Fig. 2E). These results suggest that PD-MSCs promote restoration of pathophysiological blood vessels in CCl4-injured rats through various angiogenic factors including VE-cadherin. In addition, these data suggest that vessels restored by PD-MSC Tx are involved in hepatic regeneration, with these vessels modulating the microenvironment in injured liver tissues.
Next, using a modified aortic ring assay, we confirmed whether PD-MSCs promote hepatic regeneration by upregulating new vessel formation ex vivo. In the tube formation assay using aortic samples, the length of branch outgrowth from the aorta was significantly decreased by treatment with CCl4 in the absence of PD-MSC co-culturing (Co-free group) compared to control culture conditions (p < 0.05) (Fig. 2C and F). In addition, the number of 5-bromo-2’-deoxyuridine (BrdU)-positive hepatocytes was significantly decreased by CCl4 treatment only but not control treatment conditions (p < 0.05) (Fig. 2C and G). However, tube formation was significantly recovered by co-culture with PD-MSCs compared to CCl4 treatment only (p < 0.05). Additionally, hepatocyte proliferation was significantly higher in the PD-MSC co-culture group than in the CCl4 treatment only group (p < 0.05) (Fig. 2C and G). Furthermore, hepatocyte proliferation with vessel recovery by PD-MSCs was similarly observed in the in vitro co-culture system (Supplementary Fig. 3). These results suggest that engrafting PD-MSCs into a damaged liver promotes hepatocyte proliferation by inducing the formation of new vessels with adjacent hepatic vein in the CCl4-injured rat model.
PD-MSC Tx increases the expression and nuclear translocation of β-catenin in CCl4-injured rats
The Wnt signaling pathway is known to regulate angiogenesis of endothelial cells. Specifically, Wnt and β-catenin regulate the formation of capillary-like networks and endothelial cell proliferation [26]. Therefore, we investigated whether PD-MSCs promote Wnt/β-catenin signaling in CCl4-injured rat liver. Axin2 regulates the transcriptional activity of Wnt/β-catenin signaling [27]. The expression level of Axin2 mRNA was higher in the CCl4-damaged liver group, including the NTx group, than in the control group. However, Axin2 mRNA expression level at 1 week was significantly higher in the PD-MSC Tx group than in the NTx group (p < 0.05) (Fig. 3A). Also, we evaluated the protein expression of p-LRP6, which is a known upstream regulator of β-catenin, in CCl4-injured liver, and we found that p-LRP6 protein levels were reduced after CCl4 injection. However, at 1 and 2 weeks, the levels of p-LRP6 in the PD-MSC Tx group were significantly higher than those in the NTx group (p < 0.05) (Fig. 3B). Similarly, the protein expression of active β-catenin (β-catenin) at 1, 2 and 3 weeks was reduced in the CCl4-treated and NTx groups. However, the expression of β-catenin at 1 week was dramatically higher in the PD-MSC Tx group than in the NTx group, although there were no significance differences in its expression at 2 and 3 weeks between the two groups (p < 0.05) (Fig. 3C).
To confirm whether PD-MSC Tx-induced upregulation of β-catenin is specifically detected in the hepatic endothelium, we performed immunofluorescence in CCl4-injured rat liver using anti-β-catenin antibody. β-Catenin was localized in the nucleus of hepatic endothelial cells (arrow indicates active β-catenin in endothelial cells) or in the hepatocyte transmembrane, but vWF was only localized in hepatic vessels in the control liver (Fig. 3D). In addition, β-catenin was expressed in only the hepatic endothelium, and its translocation from the cytoplasm to the nucleus was reduced in hepatic endothelial cells in the NTx liver. In contrast, there was a striking increase in β-catenin expression in the nucleus of hepatic endothelial cells, and it was co-localized with vWF in the PD-MSC Tx group but not in the NTx group (Fig. 3D). Thus, through upregulation of Axin2 and pLRP6, PD-MSCs have been suggested to promote β-catenin stabilization by upregulating the expression and translocation of β-catenin into the nucleus of hepatic endothelial cells in CCl4-injured rats.
Wnt pathway gene expression was decreased in human cirrhotic liver
To examine the correlation between liver cirrhosis and the Wnt signaling pathway, we investigated the expression of Wnt signaling-related genes and protein localization in cirrhotic human liver. The localization pattern of β-catenin in human cirrhotic liver was similar to that of the animal model. Interestingly, β-catenin expression was gradually decreased in human liver in a stage-dependent manner (Fig. 4B and G). However, β-catenin was translocated from the cytoplasm to the nucleus in endothelial cells from stage I cirrhotic liver compared to that from stage 0 liver (Fig. 4G). LRP6 mRNA level was also significantly lower in stage III cirrhotic liver than in stage 0 (p < 0.05) and stage I livers (p < 0.05) (Fig. 4A). However, expression of GSK3β, which is a negative regulator of β-catenin signaling, gradually increased in stage III cirrhotic liver compared to stage 0 (p < 0.05) and stage I (p < 0.05) (Fig. 4C). Axin2 expression was significantly higher in stage I cirrhotic liver than in stage 0 liver, but there was no significant difference in stage III cirrhotic liver (p < 0.05) (Fig. 4D). Interestingly, expression levels of c-Myc and cyclin D1, which are transcriptional factors mediated by β-catenin in endothelial cells, were significantly lower in stage III cirrhotic liver than in stage 0 and stage I livers (p < 0.05) (Fig. 4E and F). Based on these results, we hypothesized that the Wnt signaling pathway could be involved in liver cirrhosis or liver regeneration.
Effect of co-culturing PD-MSCs with CCl4-treated HUVECs and rat hepatocytes on Wnt signaling in the in vitro co-culture system
To confirm the effect of co-culturing PD-MSCs with HUVECs and WB-F344 cells on the expression of Wnt signaling markers in vitro, we introduced Wnt inhibitors (BIO: GSK3β inhibitor, IWP: LRP6 inhibitor) into the co-culture system. First, we established 10 nM BIO and 54 nM IWP-2 as optimal concentrations that elicited upregulation or downregulation of β-catenin in the treated group compared to untreated control (p < 0.05) (Supplementary Fig. 4). Protein expression levels of β-catenin and p-LRP6 were significantly higher in the BIO-treated group than in the untreated control group (p < 0.05). In contrast, β-catenin and p-LRP6 protein expression levels in the CCl4-treated group were significantly lower than the levels in the control group (p < 0.05). Furthermore, compared with Co-free conditions, PD-MSC co-culturing significantly increased the expression of β-catenin and p-LRP6 regardless of BIO or CCl4 treatment (p < 0.05) (Fig. 5A and B). In contrast, the expression levels of active β-catenin and p-LRP6 were effectively inhibited by IWP-2 treatment in both the control and CCl4-treated groups (p < 0.05) (Supplementary Fig. 4). However, decreased β-catenin and p-LRP6 expression levels in directly co-cultured cells (HUVECs and WB-F344 cells) were significantly increased by PD-MSC co-culturing regardless of BIO or CCl4 treatment (p < 0.05) (Supplementary Fig. 4).
In addition, β-catenin localization was detected in the transmembrane of rat hepatocytes or the nucleus of HUVECs in the control group, but these localizations were not detected following CCl4 treatment. However, β-catenin localization was rescued when both cell types were co-cultured with PD-MSCs regardless of BIO or CCl4 treatment (Fig. 5C). These results showed that PD-MSCs promote hepatic regeneration in cirrhotic liver via hepatic angiogenesis by activating the β-catenin-mediated Wnt pathway.
PD-MSCs effectively decrease HUVEC permeability through the Wnt pathway in the in vitro co-culture system
To investigate whether PD-MSCs promote endothelial cell permeability through the Wnt pathway, we adopted the endothelial permeability assay (Fig. 6A). Endothelial permeability was significantly blocked by BIO treatment compared to control group (p < 0.05). However, although permeability was significantly increased by damaging cells with CCl4 treatment but not by control conditions (p < 0.05), there was no significant difference between BIO and CCl4 treatment. In addition, endothelial permeability was strongly inhibited by PD-MSC co-culturing in the CCl4-treated group and in the combined BIO and CCl4-treated group, even though there was no synergistic effect of BIO treatment and PD-MSC co-culturing on endothelial permeability in the CCl4-treated group (p < 0.05) (Fig. 6B).
In contrast to the effect of BIO treatment, endothelial permeability was significantly enhanced by IWP treatment compared with that in the control group (p < 0.05) (Fig. 6C). Furthermore, although endothelial permeability was strongly enhanced by co-treatment with IWP and CCl4 (p < 0.05), it was effectively inhibited by PD-MSC co-culturing regardless of IWP or CCl4 treatment (p < 0.05). Furthermore, to investigate whether alteration of endothelial permeability was regulated by PD-MSC co-culture, we inhibited the Wnt pathway by pretreatment of PD-MSCs with IWP before co-culturing with endothelial cells. Interestingly, endothelial permeability was significantly enhanced by co-culturing with IWP-pretreated CP-MSCs compared to control culturing conditions (p < 0.05) (Fig. 6C). Taken together, our results suggest that PD-MSC Tx and the in vitro co-culture system promote hepatic regeneration in the CCl4-injured liver by regulating the vascular structure and endothelial permeability through the Wnt signaling pathway.