Abstract
Fibrotic tissue in the liver is mainly composed of collagen. Fibronectin, which is also present in fibrotic matrices, is required for collagen matrix assembly in vitro. It also modulates the amount of growth factors and their release from the matrix. We therefore examined the effects of the absence of fibronectin on the development of fibrosis in mice.
Conditional deletion of fibronectin in the liver using the Mx promoter to drive cre expression resulted in increased collagen production and hence a more pronounced fibrosis in response to dimethylnitrosamine in mice. Exclusive deletion of fibronectin in hepatocytes or normalization of circulating fibronectin in Mx-cKO mice did not affect the development of fibrosis suggesting a role for fibronectin production by other liver cell types. The boosted fibrosis in fibronectin-deficient mice was associated with enhanced stellate cell activation and proliferation, elevated concentrations of active TGF-β, and increased TGF-β-mediated signaling.
In vitro experiments revealed that collagen-type-I production by fibronectin-deficient hepatic stellate cells stimulated with TGF-β was more pronounced, and was associated with augmented Smad3-mediated signaling. Interfering with TGF-β signaling using SB431542 normalized collagen-type-I production in fibronectin-deficient hepatic stellate cells. Furthermore, precoating culture plates with fibronectin, but not collagen, or providing fibronectin fibrils unable to interact with RGD binding integrins via the RGD domain significantly diminished the amount of active TGF-β in fibronectin-deficient stellate cells and normalized collagen-type-I production in response to TGF-β stimulation. Thus, excessive stellate cell activation and production of collagen results from increased active TGF-β and TGF-β signaling in the absence of fibronectin.
In conclusion, our data indicate that fibronectin controls the availability of active TGF-β in the injured liver, which impacts the severity of the resulting fibrosis. We therefore propose a novel role for locally produced fibronectin in protecting the liver from an excessive TGF-β-mediated response.
Introduction
The development of liver fibrosis is one of the early steps in the pathogenesis of advanced liver disease and liver failure. Hepatic stellate cells play a key role in this process by producing extracellular matrix molecules that become incorporated in a distorted fibrogenous network hindering normal function in hepatocytes [1]. Transforming-growth-factor-β (TGF-β) has been described as a major stimulator of stellate cell activation, and hence, extracellular matrix production [2].
Fibronectin is one of the molecules produced by hepatic stellate cells [3]. It is also part of the extracellular matrix [4]. Several isoforms have been described and many functions have been attributed to fibronectin. Fibronectin is important for the assembly of a collagen matrix in vitro [5], [6]. Its continuous presence also supports matrix integrity, both in vitro and in vivo [5], [7]. It further regulates cell proliferation and cell cycle progression [8]. We have shown that two isoforms of fibronectin reflect the severity of liver fibrosis in patients with chronic hepatitis C raising the possibility that fibronectin itself may play a role in the pathogenesis of fibrosis [9].
In order to define the role of fibronectin in the liver on the development of liver fibrosis we undertook the following study in which fibronectin was conditionally deleted in various cell types in mice. Liver injury was induced using dimethylnitrosamine (DMN) and experiments on liver tissue from these mice were complemented with in vitro experiments in isolated stellate cells. Our results show that deletion of fibronectin leads to an increase in stellate cell activation, both at baseline and after stimulation with TGF-β. This is due to an increase in TGF-β bioavailability and results in a more pronounced fibrosis.
Results
Deletion of fibronectin in hepatic stellate cells, hepatocytes, Kupffer cells and endothelial cells
Hepatic stellate cells were isolated from control mice (CT) and conditional knockout mice carrying the Mx promoter attached to cre recombinase (Mx-cKO). All mice were homozygote for the floxed fibronectin gene. The Mx promoter was activated in vivo in Mx-cKO mice and through the resulting expression of the attached cre enzyme the homozygote fibronectin floxed genes present in the cells were deleted (Figures 1A and 1B) [10]. 97±2% of the isolated hepatic stellate cells stained for either desmin or glial fibrillary acidic protein (GFAP), both of which are markers used for identifying stellate cells [11] suggesting a high purity. Deletion of fibronectin was confirmed at the DNA (Figures 1A and 1B), mRNA (Figure 1C) and protein level (Figure 1D: Western blot; Figure 1E: ELISA of conditioned media) in cells and conditioned media. Thus, activation of the Mx promoter successfully deleted fibronectin in hepatic stellate cells. In addition to deleting fibronectin in stellate cells, the activated Mx promoter can efficiently delete fibronectin in hepatocytes, which are responsible for the production of circulating fibronectin. Therefore, loss of fibronectin in hepatocytes in mice that express Mx-cre and are homozygote for floxed fibronectin can be easily confirmed by measuring circulating plasma fibronectin, which in this case confirmed successful deletion (Figure 1F). Other cell types that can be affected by the activation of the Mx promoter are Kupffer and endothelial cells. The identity of isolated Kupffer cells was confirmed by staining with F4/80 and absence of staining with desmin or GFAP (data not shown). ELISA of conditioned media confirmed deletion of fibronectin to 14% of control values (Figure 1G). The identity of isolated endothelial cells was confirmed by positive staining with acetylated-LDL and absence of staining with F4/80, desmin and GFAP (Data not shown). ELISA of conditioned media also showed a decrease in fibronectin protein to 19% of control values (Figure 1H).
Induction of liver fibrosis
We next induced liver fibrosis in cKO mice in which the Mx promoter was activated (Mx-cKO) and control (CT) mice using dimethylnitrosamine (DMN). ELISA of circulating fibronectin confirmed successful deletion (Figure 2A) [12], [13]. To exclude the possibility that some fibronectin was generated in fibrotic livers by intronic or alternate promoters we performed quantitative RT-PCR for mRNA encoding the C-terminal part of fibronectin, the EDA, or the EDB domain. There was more than 90% decrease in all three products in fibrotic Mx-cKO mice compared to CT mice (Figure S1). Fibronectin content was examined in the liver by Western blotting (Figure 2B), which showed a difference in fibronectin amounts between fibrotic CT and fibrotic Mx-cKO. This was further confirmed by ELISA of liver lysates (Figure 2C). No significant increase in the amount of fibronectin was seen after induction of liver fibrosis in cKO animals compared to healthy cKO mice however, further supporting the conclusion that fibronectin deletion in the liver was successful in vivo.
Effect on extracellular matrix accumulation
DMN treatment resulted in increased sirius-red staining (i.e., extracellular matrix) in fibrotic-Mx-cKO mice when compared to fibrotic CT mice (Figure 2D). Further, collagen-type-I staining suggested that more collagen accumulated in fibrotic Mx-cKO mice (Figure 2E-upper panel). This was confirmed using a biochemical method measuring the amount of hydroxyproline, which is a component and hence a marker for collagen, in liver lysates (Figure 2E-lower panel). Using this method, healthy Mx-cKO showed an increase in collagen (1.3-fold) compared to healthy controls (p<0.05), while treatment with DMN resulted in a larger increase in collagen-type-I in fibrotic-Mx-cKO (1.6-fold) compared to fibrotic-CT (p<0.05).
Role of apoptosis
The development of liver fibrosis in response to DMN is preceded by increased cell death leading to signals that promote extracellular matrix accumulation [14], [15]. Fibronectin may exert antiapoptotic effects [15], [16]. We therefore sought to determine whether loss of fibronectin in Mx-cKO mice is associated with increased apoptosis and hence increased fibrosis [14], [15]. No difference in the total number of apoptotic cells (stained with TUNEL or with cleaved caspase 3 (ASP175)) between fibrotic Mx-cKO and fibrotic CT livers could be detected (Figure 2F and data not shown). Similarly, we could not detect any differences in the levels of the liver enzymes AST or ALT that reflect hepatocyte necrosis (Figure S2). Thus the increase in matrix accumulation in Mx-cKO mice is not due to increased apoptosis and hence it was not due to loss of an antiapoptotic effect of fibronectin in the liver.
Defining the cell type associated with increased matrix accumulation in Mx-cKO mice
Hepatocytes
The Mx promoter is able to delete fibronectin in stellate cells, hepatocytes, Kupffer cells and endothelial cells as shown above [15] (Figure 1). In order to exclude a role of fibronectin produced by hepatocytes themselves on the development of fibrosis in Mx-cKO mice we deleted fibronectin in hepatocytes exclusively using the albumin promoter (Alb-cKO). Loss of circulating fibronectin due to successful deletion in hepatocytes was confirmed using ELISA (Figure 3A). The degree of liver fibrosis in Alb-cKO and CT was similar, as determined by measuring the amount of collagen (Figure 3B). Therefore, the absence of fibronectin in hepatocytes in Alb-cKO does not affect the development of fibrosis. We also administered circulating plasma fibronectin to both CT and Mx-cKO mice, and were able to normalize the levels of circulating fibronectin in Mx-cKO mice (Figure S3). This did not affect the degree of fibrosis in CT mice or prevent the increased matrix accumulation in Mx-cKO mice in comparison to fibrotic CT mice (Figure 3C). It follows that lack of fibronectin production in hepatocytes of Mx-cKO mice, and loss of circulating fibronectin do not contribute to enhanced fibrogenesis in these mice.
Inflammatory cells
We then sought to determine whether inflammatory cells were involved in mediating increased fibrosis in Mx-cKO mice. An increase in inflammatory cells (CD45+) was noted in fibrotic mice compared to healthy mice but there was no difference between Mx-cKO and CT in either the healthy or the fibrotic pairs (Figure 3D). This indicates that fibronectin does not affect the infiltration of inflammatory cells in response to DMN.
Hepatic stellate cells
In order to examine the role of stellate cells on fibrosis development in Mx-cKO mice staining for α-SMA was performed. α-SMA has been used as a marker for stellate cell activation but it also stains smooth muscle cells [17]. There was a significant increase in the number of α-SMA-positive cells in both fibrotic and healthy Mx-cKO mice compared to fibrotic and healthy controls respectively (Figure 3E). This increase was associated with an increase in proliferation (as evidenced by co-staining for GFAP and Ki-67) in fibrotic Mx-cKO mice compared to fibrotic CT mice (5% vs. 1%, p<0.05) (Figure 3F). In contrast there was no difference in apoptosis (as evidenced by co-staining for GFAP and TUNEL and for GFAP and cleaved caspase 3 (ASP175)) between fibrotic Mx-cKO mice and fibrotic CT mice (Figure S4 and data not shown). We therefore conclude that the increase in matrix accumulation in Mx-cKO is associated with an increase in proliferation and activation of stellate cells in vivo.
Establishing a causal relationship between deletion of fibronectin in stellate cells and increased matrix production
Response of cKO stellate cells to TGF-β
As shown previously, in vivo data had suggested enhanced stellate cell activation and increased collagen production in both healthy and fibrotic Mx-cKO mice compared to their respective controls (Figures 3E and 2E). In order to firmly establish a role for stellate cells lacking fibronectin in mediating increased matrix production in vivo, these cells were isolated and examined in vitro. α-SMA expression as measured by quantitative real time PCR (qPCR) revealed that Mx-cKO cells had a higher baseline expression of α-SMA after 24 hours of culture on plastic and even more so after 48 hours compared to controls (FN fl/fl or Alb-cre FN fl/fl lacking fibronectin in the circulation only) (Figure 4A), suggesting increased activation. Collagen-type-I relative mRNA expression in cultured stellate cells at baseline was also significantly increased in Mx-cKO cells when compared to CT or Alb-cKO, as measured by qPCR (Figure 4B). We then stimulated stellate cells with TGF-β because it is a major fibrogenic mediator in the liver [2], [18]. Stellate cells cultured on plastic for 24 hours after isolation in the presence of fibronectin-depleted serum were treated with different doses of TGF-β for an additional 24 hours. This resulted in significantly higher collagen-type-I mRNA expression in Mx-cKO compared to both types of control cells (CT and Alb-cKO) at all doses tested (Figure 4C). For example, at 10 ng/ml of TGF-β there was 5-fold induction in Mx-cKO cells compared to 2-fold induction in CT cells in relative mRNA expression of collagen-type-I after correcting for baseline differences (p<0.05, Figure 4C). In contrast, stimulation of hepatocytes with TGF-β failed to show an increase in collagen-type-I expression or a difference between Mx-cKO and CT cells at the three doses tested (data not shown). Lastly, stimulation of either Kupffer cells or endothelial cells failed to show a difference in the response to TGF-β between CT and Mx-cKO cells (Figures S5 and S6). This shows that the absence of fibronectin increases the responsiveness of stellate cells to TGF-β stimulation.
Specificity of increased responsiveness of cKO cells
We then sought to determine whether the increase in responsiveness of cKO cells was specific for TGF-β, and therefore performed three sets of experiments. In the first set of experiments, we examined whether the increase in responsiveness to TGF-β is translated into increased intracellular signaling specific for TGF-β. Therefore, translocation of SMAD3 to the nucleus was studied [18], [19]. At baseline the percentage of nuclei in which SMAD3 translocation to the nucleus took place (shown as yellow due to overlay of red SMAD3 and green DAPI staining) was higher in cKO cells (baseline CT: 8% vs. cKO: 16%, p<0.005). Consistently, after stimulation more translocation took place in cKO cells (CT: 32% vs. cKO: 49%, p<0.0001 for CT and cKO compared to the respective baseline and p<0.005 compared to each other) (Figure 4D).
In the second set of experiments we sought to determine whether cytokines other than TGF-β cause increased collagen mRNA expression in cKO cells. Neither treatment with PDGF-BB nor EGF reproduced the effects seen with TGF-β (Figure 4E and data not shown).
In the third set of experiments TGF-β signaling was inhibited in both CT and cKO cells to determine whether a more pronounced inhibition ensued in cKO cells. Adding an inhibitor of TGF-β signaling (SB431542) to cKO cells one hour before treating with TGF-β resulted in a significant decrease in collagen-type-I mRNA expression compared to TGF-β treatment alone in both CT and cKO cells (p<0.05, Figure 4F). Furthermore, the inhibitor resulted in collagen expression in inhibitor-treated cKO cells that was equivalent to inhibitor-treated CT cells (p = NS, Figure 4F). Unstimulated cells showed similar relationships (Figure S7).
We then asked whether this specific response was associated with increased TGF-β-receptor mRNA expression. Indeed, we found that the relative mRNA expression of TGF-β-receptor-II increased in stimulated cKO cells in vitro (Figure 4G) while the relative mRNA expression of TGF-β-receptor-I (Alk-5) was not significantly altered between cKO and CT cells in vitro (Figure 4H).
In summary, the increase in responsiveness in cKO was specific for TGF-β as was shown in three different experiments.
TGF-β amounts in cKO stellate cells
Because of increased responsiveness of cKO cells at baseline we sought to examine whether TGF-β amounts were different between cKO and CT cells. Active TGF-β protein concentrations in conditioned media from hepatic stellate cells cultured for 24 hours on plastic without stimulation were increased in Mx-cKO cells compared to CT as determined by ELISA and confirmed in a bioassay (p<0.05, Figure 5A and data not shown). The amounts of total TGF-β (including matrix TGF-β) were not significantly different between both groups and consequently the percentage of active to total TGF-β was significantly increased in Mx-cKO cells (Figure 5B). This increase could not be directly attributed to an increase in TGF-β mRNA amounts, since TGF-β relative mRNA expression was similar between cKO and CT cells (p = NS, Figure 5C).
Establishing a causal relationship between increased active TGF-β levels and increased responsiveness to TGF-β
The finding of an increase in active TGF-β levels does not explain why stellate cells lacking fibronectin exhibit increased sensitivity compared to control cells when stimulated with equivalent concentrations of TGF-β in vitro (as shown in Figure 4C). To address this we sought to determine whether TGF-β signaling and collagen-type-I mRNA expression and consequently responsiveness of cKO stellate cells could be affected by increased exposure to TGF-β. In fact, a positive feedback loop of TGF-β on the stellate cells has been proposed [20]. We therefore examined in vitro whether repeated exposure of wildtype stellate cells to TGF-β leads to increased TGF-β-receptor-II mRNA expression and SMAD3 translocation to the nucleus, which then in turn will lead to increased collagen-type-I expression. Increased TGF-β-receptor-II mRNA expression and SMAD3 translocation were more pronounced after three exposures to TGF-β (-24, -8, -4 hours) than after a single exposure (-4 hours) (p<0.05, Figures 5D and 5E). This was associated with an increase in collagen-type-I mRNA expression (p<0.05, Figure 5F). These findings thus suggest that an increase in available active TGF-β is associated with increased responsiveness of stellate cells to TGF-β.
Changes in TGF-β protein amounts and signaling in cKO mice
Based on the described in vitro experiments, increased collagen production by the cKO stellate cells was due to an increase in active TGF-β and hence increased responsiveness of these cells. In order to determine whether an increase in available TGF-β was the culprit for boosted fibrosis in Mx-cKO mice, total TGF-β content in fibrotic-Mx-cKO and fibrotic-CT livers was determined by treating mechanically disrupted liver lysates with hydrochloric acid (HCl) to release all TGF-β including matrix-bound TGF-β. However, no difference could be detected (p = NS, Figure 6A). In contrast, active TGF-β in liver lysates was increased in the livers of fibrotic Mx-cKO mice as measured by ELISA (p<0.05, Figure 6B). Because TGF-β that is bound to the latency-associated peptide and hence is inactive, can be detected by antibodies used in ELISA tests, we performed a bioassay that detects only free unbound TGF-β [21]. Similar relationships were obtained (data not shown). The increase in the amount of active TGF-β was associated with enhanced TGF-β-receptor-II expression by Western blotting in livers from fibrotic Mx-cKO mice (Figures 6C and 6D) but no change in TGF-β-receptor-I expression in vivo (Figure S8). TGF-β signaling was also augmented as evidenced by increased SMAD3 phosphorylation in fibrotic Mx-cKO livers in vivo shown by Western blotting (Figures 6E and 6F). Presumably this ultimately led to boosted collagen production shown in Figure 2E. Hence, these in vivo findings replicate the findings in stellate cells in vitro and implicate the increase in active TGF-β as the culprit for enhanced collagen production in fibrotic Mx-cKO mice.
The role of the presence of fibronectin around the cells
The previous data suggest that the absence of fibronectin in the matrix surrounding the stellate cells leads to increased active TGF-β and increased responsiveness of cKO stellate cells. Thus one would expect that the presence of fibronectin in the surrounding of the stellate cells would have two consequences: first, it will result in a decrease in active TGF-β in the media and second, it should protect the cKO cells from excessive stimulation by TGF-β. Wells were therefore pre-coated with fibronectin. The amount of active TGF-β in the conditioned media of stellate cells cultured for 24 hours on fibronectin decreased significantly both in control cells (p<0.05), and in cKO cells (p<0.05, Figure 7A). Thus coating with fibronectin reduced the amount of active TGF-β in the media of cKO cells, but did not completely normalize it. Nevertheless, the decrease was significant compared to controls cultured on plastic (p<0.05). The second consequence, namely that fibronectin coating should lead to normalization of the responsiveness of cKO cells was also confirmed. The TGF-β-induced increase in mRNA expression of TGF-β-receptor-II and collagen type I was blunted in cKO cells cultured on fibronectin (uncoated vs. coated Mx-cKO: p<0.05, coated cKO vs. coated CT: p = NS, coated cKO vs. CT on plastic: p = NS, Figures 7B and 7C). These data thus show a complete restoration of normal responsiveness of cKO cells in the presence of fibronectin in the cell surrounding. This effect was specific for fibronectin, because coating with collagen failed to affect the amount of collagen-type-I mRNA expression in CT or cKO cells after stimulation (Figure 7D).
In order to determine whether the formation of fibrils was enough to normalize the response of stellate cells to TGF-β we used stellate cells from mice that express only one fibronectin gene with a mutated RGD binding domain [22]. These cells were obtained from mice with the following genotype: Mx-cre/- FN-RGE/fl. Activating Mx-cre results in deletion of the floxed fibronectin allele, and the stellate cells can only produce fibronectin that is able to form fibrils but is unable to activate integrins via RGD (RGE/-). Treatment of these cells with TGF-β failed to induce an increased response, as measured by collagen-type-I mRNA expression compared to CT (Figure 7E). This suggests normalization of the response of stellate cells to TGF-β. Indeed, the concentration of active TGF-β in the conditioned media from untreated stellate cells was similar between the cells that produce normal fibronectin (CT) and those that produce the mutated fibronectin (RGE/-) (Figure 7F). This indicates that binding of fibronectin to integrins through the RGD domain is not required for normalization of the stellate cells response to TGF-β, while providing fibronectin fibrils is enough for normalization of the response.
Based on these findings we conclude, that the presence of fibronectin fibrils in the matrix surrounding stellate cells significantly diminishes the amount of active TGF-β released and restores a normal response to TGF-β.
Discussion
The principal findings of this study are: 1) Deletion of fibronectin in the liver results in increased collagen at baseline and in response to induction of experimental fibrosis. 2) Deletion of fibronectin in hepatic stellate cells leads to increased collagen-type-I at baseline and after activation. 3) These effects are associated with increased active TGF-β protein and cellular response to TGF-β, are mediated by the SMAD pathway, are not due to increased TGF-β mRNA expression, but are associated with increased TGF-β-receptor-II mRNA expression. 4) Inhibiting TGF-β signaling or providing fibronectin to stellate cells in vitro normalizes the heightened responsiveness of cKO stellate cells. The increase in fibrosis in mice lacking fibronectin in the liver is thus due to an increase in active TGF-β, most likely resulting from a failure to incorporate TGF-β in the matrix in the absence of fibronectin. These findings show that in vitro fibronectin production by hepatic stellate cells is critical for modulating the amount of available bioactive TGF-β and hence the cell responsiveness to TGF-β. This offers a possible explanation for the in vivo finding that the presence of fibronectin seems to protect against exacerbation of fibrosis.
Fibronectin is required in vitro for the assembly of a collagen matrix [5], [6]. In its absence a reduction in collagen matrix assembly would be expected. This, however, was not the case in our model. Immunohistochemistry of liver sections (data not shown) and ELISA data (Figure 2C) show a residual amount of fibronectin in the matrix, despite the deletion of fibronectin in the circulation and stellate cells by >95%. Presumably, this residual amount originates from the time before induced deletion of fibronectin at 4 weeks postnatally, and might be sufficient to support collagen assembly. It is also possible, that despite the apparent successful deletion of fibronectin the remaining undeleted circulating fibronectin is enough to support normal collagen matrix assembly even though it is much diminished, since circulating fibronectin contributes to tissue fibronectin [7], [23]. A related issue is whether the decrease in the amount of surrounding fibronectin may modulate the synthesis of collagen. Our in vitro data show that this is the case since cKO stellate cells unable to produce fibronectin showed a decrease in collagen expression when cultured on fibronectin-coated plates compared to cells cultured on plastic (Figure 7C). Unfortunately, the degree of deletion of circulating fibronectin using the albumin promoter at 91% was less pronounced than using the Mx promoter at 97% in relation to control mice. Theoretically this residual circulating fibronectin might have been enough and hence could have resulted in equivalent degrees of fibrosis in hepatocyte specific conditional knockouts and their controls, but this is unlikely because we could not detect a dose response effect of circulating fibronectin on fibrosis, and normalizing circulating levels of fibronectin failed to prevent the enhanced fibrosis in Mx-cKO mice compared to fibrotic CT mice. Another cell type that is affected by the deletion of fibronectin using the Mx-promoter is the endothelial cell. Sinusoidal endothelial cells exert a profibrotic effect particularly in the early stages of liver fibrosis [24]. It has been shown that early in fibrosis these cells produce a fibronectin isoform called EDA fibronectin that can be detected in sinusoids in vivo as well as in vitro in endothelial cell cultures. Furthermore, endothelial cells in culture form a fibronectin matrix that accelerates hepatic stellate cell activation [24]. Based on these findings one would have expected a decrease in stellate cell activation in the absence of fibronectin in endothelial cells, and not an increase as presented here. On the other hand, it is conceivable that the absence of fibronectin production by the endothelial cells using the Mx-promoter contributed to the increase in bioavailable TGF-β. At any rate our data clearly implicate the stellate cells both in vivo and in vitro as a culprit for increased collagen production in our model.
The activated stellate cells produce fibronectin isoforms that differ in their characteristics from plasma fibronectin [3]. While plasma fibronectin represents the soluble form of fibronectin, the three isoforms produced by the stellate cells belong to the group of cellular isoforms that are also viewed as being insoluble, since they are found in high concentrations in the tissue. These three isoforms are: EDA, EDB and oncofetal fibronectin (oFN) (defined by an O-glycosylation site in the variable region) [3], [25], [26]. Only little of these isoforms escapes from the matrix and can be detected in the circulation [27], while plasma fibronectin has no demonstrable role in the formation of fibrosis in any reported model. Under certain circumstances plasma fibronectin is involved in extracellular matrix integrity [25], [28], and a number of in vitro studies examining the role of fibronectin in collagen matrix assembly have used plasma fibronectin despite the presence of one report suggesting that the absence of extra-domain-B (EDB) results in fibronectin fibrils that are thinner and shorter [29]. Differences, however, have been reported for the effects on the cells. The EDB domain has been shown to support the proliferation of fibroblasts [29]. The presence of extra-domain-A (EDA) has been implicated in upregulating binding affinity to α5β1-integrin and integrin-mediated signal transduction [8], [30]. It has also been identified as a binding domain for α4β1-integrin [31]. Even though the possibility that an isoform of fibronectin, other than plasma fibronectin may normally inhibit stellate cells cannot be excluded, we are not aware of data to support such a conclusion. If anything there are reports that the presence of fibronectin containing the EDA domain resulted in activation, instead of inhibition of hepatic stellate cells in vitro [24], [32]. Together these data support an indirect effect of fibronectin with regards to TGF-β regulation.
An increase in active TGF-β in the absence of an adequate fibronectin matrix can explain all our findings starting with the increase in baseline production of collagen-tsype-I in healthy as well as in fibrotic Mx-cKO mice, the increase in active stellate cell numbers, and finally the increase in TGF-β signaling [2], [33]. At first glance it may seem surprising that fibronectin is able to blunt the response to TGF-β. It has been unequivocally shown however, that fibronectin plays two roles in TGF-β modification. On the one hand, it is required for the incorporation of TGF-β bound to the latent TGF-β binding protein-1 (LTBP-1) in the matrix [34], and on the other hand, it is required for the αvβ6-integrin mediated activation of TGF-β stored in the matrix [35]. Our in vitro data show an increase at baseline in bioactive TGF-β without an associated increase in TGF-β mRNA in cKO stellate cells. Thus the data suggest that the incorporation of TGF-β in the matrix, and not the production of TGF-β is abnormal resulting in an increase in the amount of available TGF-β. The fact that total TGF-β was similar between CT and cKO fibrotic livers suggests that in the absence of fibronectin in stellate cells total TGF-β is irrelevant to the pathophysiology of fibrosis. It also appears that the total amount of TGF-β is not modulated by the amount of free TGF-β since in the CT livers 7% of the total TGF-β consisted of active TGF-β, while in cKO livers the percentage was 17%.
Recently accepted but not yet published work by Moriya et al. [36] used a similar model and focused on matrix assembly. While we found that there was significant induction of collagen production in Mx-cKO mice at baseline and after DMN and CCl4 administration (Figure 2E and data not shown), Moriya et al. found no difference either at baseline or after CCl4 treatment. Since our finding was reproducible and could be confirmed using two different models of fibrosis induction we currently lack an explanation for this discrepancy. Both works complement each other in that Moriya et al. imply that TGF-β stimulates matrix assembly, while we established both in vivo and in vitro that the amount of active TGF-β originating from the stellate cells is at least partially responsible for increased matrix production in a fibronectin-deficient state.
In summary, our data support the conclusion that in healthy and injured liver, stellate cells require adequate amounts of matrix fibronectin in order to buffer the amount of bioavailable TGF-β and its fibrogenic outcome.
Materials and Methods
Mice
Mice possessing an inducible Mx promoter or an albumin promoter driving cre-recombinase expression were crossed with mice carrying loxP-flanked (floxed) fibronectin (FN-fl/fl) or mice carrying a fibronectin gene with a mutated RGD to RGE sequence in the integrin binding domain and the floxed gene on the second allele (FN-RGE/fl) [15], [22], [37]. Mx was induced at 4 weeks of age using 250 µg polyinosinic-polycytidylic acid (pIpC) injected three times over one week as described [15]. Animals used were of the C57Bl/6 background. Fibronectin in plasma or liver lysates was determined by ELISA [25].
Induction of liver fibrosis
Liver fibrosis was induced by injecting dimethylnitrosamine (DMN) intraperitoneally at 10 µg DMN/g body weight (Sigma-Aldrich) 3x/week for 4 weeks in male mice [13]. Mice in the Mx group started induction of liver fibrosis at the age of 8-9 weeks while mice in the Alb group started induction at the age of 9–11 weeks. Human plasma fibronectin isolated from outdated plasma as described was administered i.p. at a dose of 1 mg/mouse/d [25], [38]. The studies in mice were approved by the animal protection committee of the University of Heidelberg #ma34050, #ma35590, #ma36955, and #A-18/09. Thus, all animal work was conducted according to relevant national and international guidelines.
Determination of the degree of fibrosis
Morphometric analysis of areas stained with sirius-red was performed using the program ImageJ (Wayne Rasband, NIH). Hydroxyproline was measured using a biochemical method as published [39], [40]. To calculate the amount of collagen, we multiplied measured hydroxyproline content by 10 as proposed by Berg in 1982 [39]. Liver enzymes alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured in serum using routine clinical chemistry.
Cell culture
Hepatic stellate cells, Kupffer cells and endothelial cells were isolated from Mx-conditional knockout mice (Mx-cKO) which carry the Mx-promoter attached to cre and are homozygous for the floxed fibronectin gene [15], [37]. For some experiments mice that carry Mx-cre, a mutated fibronectin gene that contains an RGE domain instead of the RGD domain and a floxed fibronectin gene were used (Genotype: Mx-cre/- FN-RGE/fl) [22]. In all Mx mice the Mx promoter was activated in vivo prior to cell isolation as described [10]. Cells from littermates lacking the Mx-cre gene or carrying the Alb-cre gene together with two floxed fibronectin genes were used as controls. Briefly, the liver was perfused and the cells were separated using Nycodenz (Axis-shield) as described [41]. Cells were cultured in DMEM and 10% fibronectin-depleted fetal-calf-serum (FCS) [10]. Depletion of FCS was performed by running FCS on gelatine-sepharose columns 3-5 times. Final concentration of fibronectin was 8-11 ng/ml in the batches used. All experiments were started 24 hours after isolation. Unless noted otherwise, cells were treated with TGF-β1 (1-10 ng/ml), murine platelet derived growth factor-BB (PDGF-BB) (100 ng/ml) (Sigma-Adrich) or murine epidermal growth factor (EGF) (100 ng/ml) (Peprotech). The TGF-β signaling inhibitor (SB431542) (Merck) was used at a final concentration of 10 µM [42]. Treatment with TGF-β1, PDGF-BB or EGF lasted for 24 hours or as described in the figure legends, and treatment with the inhibitor lasted one hour, after which TGF-β was added and left for an additional 2 hours before collecting the samples. Fibronectin used for coating was isolated from outdated human plasma as described [25], and collagen was from human placenta (Sigma-Aldrich). Both were added at 100 µg/ml overnight, because it has been reported that high concentrations of collagen are associated with stimulation of stellate cells [43].
DNA analysis
DNA was isolated from stellate cells using proteinase K and nested PCR was performed using the following primer pairs consecutively: fibronectin forward-5′-TCGCACCCGCTGCGCTGCA-3′, reverse-5′-ATTGTCAAAACAGCCAGCTGCGA-3′, and then forward-5′-AGGGTGTGAGCCGGACAACT-3′, reverse-5′-CAACTGACTTGGTGAGCCTGCA-3′.
RNA analysis
RNA was isolated using RNeasy (Qiagen). Quality was tested using RNA-6000-Pico-Chip (Agilent). Reverse transcription was performed using iScript-Select (BioRad). The following primer pairs were used: fibronectin forward-5̀-GTCAGTGTCTCCAGTGTCTAC-3′, reverse-5′-TGGCTTGCTGGCCAATCAGT-3′, desmin forward-5′-CTGGAGGCTTGGGGTCGC-3′, reverse-5′-CGGATCTCCTCTTCATGCAC-3′, glial fibrillary acidic protein (GFAP) forward-5′-AGCGAGCGTGCAGAGATGAT-3′, reverse-5′- CCGAGGTCCTGTGCAAAGT-3′, GAPDH forward-5′-TGAAGGTCGGTGTGAACGGATTT-3′, reverse-5′-CATGTAGGCCATGAGGTCCAC-3′.
Quantitative real time PCR (qPCR) products were obtained using LightCycler and TaqMan probes, and results were normalized to hypoxanthine-phosphoribosyltransferase (HPRT) after standard curves were established. The choice of HPRT for normalization was based on early experiments showing comparable expression between CT and Mx-cKO liver. Data were evaluated using RelQuant (Roche). The primers used were those suggested by the universal primer library from Roche. The probes were as follows: Collagen-type-I and TGF-β1 probe #15, HPRT probe #95, TGF-β-receptor-I probe#32, TGF-β-receptor-II probe#7, α-SMA probe #21, C-terminal fibronectin probe #66, EDA fibronectin probe #76, and EDB fibronectin probe #31.
TGF-β ELISA and bioassay
Mouse TGF-β1 was determined using an ELISA kit from Promega and confirmed using a bioassay. This was done because the ELISA antibodies may detect TGF-β that is bound to the latency-associated peptide (LAP), and hence inactive. In the bioassay mink lung epithelial cells (MLEC) stably transfected with an expression construct containing a truncated PAI-1 promoter fused to the firefly luciferase reporter gene are responsive only to free active TGF-β, and hence this assay by virtue of its design only detects bioavailable TGF-β.
Conditioned media from stellate cells in culture for 24 hours were used to determine active TGF-β. This was diluted 1∶5 for the commercial assay and 1∶4 for the bioassay. In order to determine total TGF-β in vitro stellate cell plates were heated to 80°C for 10 minutes and then placed on ice. Supernatants including the lysed cells were then assayed (using the kit) or added to the reporter cells in a dilution of 1∶4 (100 µl) for 5 hours and luciferase assayed as described [21]. Bioluminescence was measured in white plates using Wallac-Victor2-1420 (Perkin-Elmer). The amount of active or total TGF-β was then corrected to total protein measured using the BCA kit (Pierce).
In order to determine TGF-β in liver tissue using the commercial assay, liver pieces (50-100 mg in 300 µl DPBS) were mechanically disrupted using a Hybaid-Ribolyser-cell-disrupter, half was then treated with HCl followed by neutralization with NaOH as suggested in the manual to determine total TGF-β. The other half was used without HCl treatment to determine active TGF-β and total protein. Both treated and untreated samples were tested at a final dilution of 1∶10. For the bioassay, liver pieces were disrupted in 500 µl DMEM, half was centrifuged at 13000 RPM for 15 minutes and supernatant used for measuring active TGF-β and total protein, and the second half was heated to 80°C for 10 minutes followed by 10 minutes on ice and 1 minute centrifugation. The supernatant was then used for total TGF-β and total protein determinations. Liver lysates were diluted 1∶400. Bioluminescence was measured as described above and corrected to total protein.
Staining protocols and immunohistochemistry
Extracellular matrix was stained using picro-sirius-red in 5 µm frozen sections of the liver as reported [44]. Cells or liver sections were fixed in 4% paraformaldehyde or ice-cold acetone. The following antibodies were used: rabbit anti-GFAP-IgG (Acris); mouse anti-desmin-IgG (AbD Serotec); rat anti-CD45-IgG and mouse anti-ki-67-IgG conjugated with alexa-555 (BD Pharmingen); rat anti-F4/80 (E-Bioscience); rabbit anti-cleaved caspase 3 (ASP175) conjugated with alexa-488 (Cell Signaling); mouse anti-α-SMA-IgG conjugated with Cy3 (Sigma-Aldrich); rabbit anti-SMAD3-IgG (Cell Signaling); rabbit anti-mouse fibronectin (Millipore); rabbit anti collagen-type-I (Abcam); goat anti-rabbit conjugated with Cy2; goat anti-rabbit conjugated with alexa-555 (Molecular Probes); goat anti-mouse-IgG conjugated with Cy2 (Dianova), and goat anti-rat-IgG conjugated with Cy2 (Dianova). Slides were photographed in the Nikon Imaging Center (University of Heidelberg) and processed using ImageJ (Wayne Rasband, NIH). TUNEL was performed according to the manufacturer's recommendations (Trevigen). Endothelial cells were stained by adding Dil-AcLDL (Invitrogen) at a concentration of 10 µg/ml to living cells for 4 hours at 37°C. Quantification was performed in at least three sections per mouse in at least three mice per group or more as noted in the figure legends. At least 200 positive cells were counted per section and then corrected to the area examined.
Western blotting and immune precipitation
Fibronectin and SMAD3 were detected using the same antibodies as above. In addition, the following antibodies were used: anti-TGF-β-receptor-II (Abcam); anti-TGF-β-receptor-I (R-20) (Santa Cruz); rabbit anti-phospho-SMAD3-IgG (Cell Signaling), and anti-β-Tubulin (Sigma-Aldrich). Western blotting of fibronectin in stellate cells in vitro was preceded by immune precipitation, which was performed using gelatine-sepharose beads added to equivalent protein amounts of cell lysates from cKO and CT cells for 1 hour. After washing, Laemmli buffer was added to the coated beads, which were boiled before being applied to the gel. The beads did not require coating with an antibody because of the avid binding of fibronectin to them. All antibodies were diluted in 5% BSA, except fibronectin which was diluted in 5% milk powder/tris-buffered saline.
Statistical analyses
Analyses were performed using SPSS (V14.0) and first applying ANOVA tests when appropriate. Comparisons were then performed using Student's t-test or Wilcoxon paired test as appropriate. Results are expressed as mean±standard-error-of-the-mean (M±SEM).
Supporting Information
Acknowledgments
We thank Reinhard Fässler for his invaluable input and Stefan Meuer for his continued support. We acknowledge the use of the Nikon Imaging Center at the University of Heidelberg.
Footnotes
Competing Interests: The authors have declared that no competing interests exist.
Funding: This work was funded by the Max-Planck Society and the University of Heidelberg. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References
- 1.Wu J, Zern MA. Hepatic stellate cells: a target for the treatment of liver fibrosis. J Gastroenterol. 2000;35:665–672. doi: 10.1007/s005350070045. [DOI] [PubMed] [Google Scholar]
- 2.Parsons CJ, Takashima M, Rippe RA. Molecular mechanisms of hepatic fibrogenesis. J Gastroenterol Hepatol. 2007;22(Suppl 1):S79–84. doi: 10.1111/j.1440-1746.2006.04659.x. [DOI] [PubMed] [Google Scholar]
- 3.Xu G, Niki T, Virtanen I, Rogiers V, De Bleser P, et al. Gene expression and synthesis of fibronectin isoforms in rat hepatic stellate cells. Comparison with liver parenchymal cells and skin fibroblasts. J Pathol. 1997;183:90–98. doi: 10.1002/(SICI)1096-9896(199709)183:1<90::AID-PATH1105>3.0.CO;2-J. [DOI] [PubMed] [Google Scholar]
- 4.Matsui S, Takahashi T, Oyanagi Y, Takahashi S, Boku S, et al. Expression, localization and alternative splicing pattern of fibronectin messenger RNA in fibrotic human liver and hepatocellular carcinoma. J Hepatol. 1997;27:843–853. doi: 10.1016/s0168-8278(97)80322-4. [DOI] [PubMed] [Google Scholar]
- 5.Sottile J, Hocking DC. Fibronectin polymerization regulates the composition and stability of extracellular matrix fibrils and cell-matrix adhesions. Mol Biol Cell. 2002;13:3546–3559. doi: 10.1091/mbc.E02-01-0048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Velling T, Risteli J, Wennerberg K, Mosher DF, Johansson S. Polymerization of type I and III collagens is dependent on fibronectin and enhanced by integrins alpha 11beta 1 and alpha 2beta 1. J Biol Chem. 2002;277:37377–37381. doi: 10.1074/jbc.M206286200. [DOI] [PubMed] [Google Scholar]
- 7.Bentmann A, Kawelke N, Moss D, Zentgraf H, Bala Y, et al. Circulating fibronectin affects bone matrix, whereas osteoblast fibronectin modulates osteoblast function. J Bone Miner Res. 2010;25:706–715. doi: 10.1359/jbmr.091011. [DOI] [PubMed] [Google Scholar]
- 8.Manabe R, Ohe N, Sekiguchi K. Alternatively spliced EDA segment regulates fibronectin-dependent cell cycle progression and mitogenic signal transduction. J Biol Chem. 1999;274:5919–5924. doi: 10.1074/jbc.274.9.5919. [DOI] [PubMed] [Google Scholar]
- 9.Hackl NJ, Bersch C, Feick P, Antoni C, Franke A, et al. Circulating fibronectin isoforms predict the degree of fibrosis in chronic hepatitis C. Scand J Gastroenterol. 2010;45:349–356. doi: 10.3109/00365520903490606. [DOI] [PubMed] [Google Scholar]
- 10.Geerts A, Niki T, Hellemans K, De Craemer D, Van Den Berg K, et al. Purification of rat hepatic stellate cells by side scatter-activated cell sorting. Hepatology. 1998;27:590–598. doi: 10.1002/hep.510270238. [DOI] [PubMed] [Google Scholar]
- 11.Niki T, De Bleser PJ, Xu G, Van Den Berg K, Wisse E, et al. Comparison of glial fibrillary acidic protein and desmin staining in normal and CCl4-induced fibrotic rat livers. Hepatology. 1996;23:1538–1545. doi: 10.1002/hep.510230634. [DOI] [PubMed] [Google Scholar]
- 12.Constandinou C, Henderson N, Iredale JP. Modeling liver fibrosis in rodents. Methods Mol Med. 2005;117:237–250. doi: 10.1385/1-59259-940-0:237. [DOI] [PubMed] [Google Scholar]
- 13.Yoshida T, Ogata H, Kamio M, Joo A, Shiraishi H, et al. SOCS1 is a suppressor of liver fibrosis and hepatitis-induced carcinogenesis. J Exp Med. 2004;199:1701–1707. doi: 10.1084/jem.20031675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pritchard DJ, Butler WH. Apoptosis--the mechanism of cell death in dimethylnitrosamine-induced hepatotoxicity. J Pathol. 1989;158:253–260. doi: 10.1002/path.1711580314. [DOI] [PubMed] [Google Scholar]
- 15.Sakai T, Johnson KJ, Murozono M, Sakai K, Magnuson MA, et al. Plasma fibronectin supports neuronal survival and reduces brain injury following transient focal cerebral ischemia but is not essential for skin-wound healing and hemostasis. Nat Med. 2001;7:324–330. doi: 10.1038/85471. [DOI] [PubMed] [Google Scholar]
- 16.Zhang Z, Vuori K, Reed JC, Ruoslahti E. The alpha 5 beta 1 integrin supports survival of cells on fibronectin and up-regulates Bcl-2 expression. Proc Natl Acad Sci U S A. 1995;92:6161–6165. doi: 10.1073/pnas.92.13.6161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Yamaoka K, Nouchi T, Marumo F, Sato C. Alpha-smooth-muscle actin expression in normal and fibrotic human livers. Dig Dis Sci. 1993;38:1473–1479. doi: 10.1007/BF01308606. [DOI] [PubMed] [Google Scholar]
- 18.Inagaki Y, Okazaki I. Emerging insights into Transforming growth factor beta Smad signal in hepatic fibrogenesis. Gut. 2007;56:284–292. doi: 10.1136/gut.2005.088690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Massague J, Seoane J, Wotton D. Smad transcription factors. Genes Dev. 2005;19:2783–2810. doi: 10.1101/gad.1350705. [DOI] [PubMed] [Google Scholar]
- 20.Moreira RK. Hepatic stellate cells and liver fibrosis. Arch Pathol Lab Med. 2007;131:1728–1734. doi: 10.5858/2007-131-1728-HSCALF. [DOI] [PubMed] [Google Scholar]
- 21.Abe M, Harpel JG, Metz CN, Nunes I, Loskutoff DJ, et al. An assay for transforming growth factor-beta using cells transfected with a plasminogen activator inhibitor-1 promoter-luciferase construct. Anal Biochem. 1994;216:276–284. doi: 10.1006/abio.1994.1042. [DOI] [PubMed] [Google Scholar]
- 22.Takahashi S, Leiss M, Moser M, Ohashi T, Kitao T, et al. The RGD motif in fibronectin is essential for development but dispensable for fibril assembly. J Cell Biol. 2007;178:167–178. doi: 10.1083/jcb.200703021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Moretti FA, Chauhan AK, Iaconcig A, Porro F, Baralle FE, et al. A major fraction of fibronectin present in the extracellular matrix of tissues is plasma-derived. J Biol Chem. 2007;282:28057–28062. doi: 10.1074/jbc.M611315200. [DOI] [PubMed] [Google Scholar]
- 24.Jarnagin WR, Rockey DC, Koteliansky VE, Wang SS, Bissell DM. Expression of variant fibronectins in wound healing: cellular source and biological activity of the EIIIA segment in rat hepatic fibrogenesis. J Cell Biol. 1994;127:2037–2048. doi: 10.1083/jcb.127.6.2037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kawelke N, Bentmann A, Hackl N, Hager HD, Feick P, et al. Isoform of fibronectin mediates bone loss in patients with primary biliary cirrhosis by suppressing bone formation. J Bone Miner Res. 2008;23:1278–1286. doi: 10.1359/jbmr.080313. [DOI] [PubMed] [Google Scholar]
- 26.Matsuura H, Hakomori S. The oncofetal domain of fibronectin defined by monoclonal antibody FDC- 6: its presence in fibronectins from fetal and tumor tissues and its absence in those from normal adult tissues and plasma. Proc Natl Acad Sci U S A. 1985;82:6517–6521. doi: 10.1073/pnas.82.19.6517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hackl N, Bersch C, Feick P, Antoni C, Franke A, et al. Circulating fibronectin isoforms predict the degree of fibrosis in chronic hepatitis. Submitted. 2009. [DOI] [PubMed]
- 28.Ni H, Yuen PS, Papalia JM, Trevithick JE, Sakai T, et al. Plasma fibronectin promotes thrombus growth and stability in injured arterioles. Proc Natl Acad Sci U S A. 2003;100:2415–2419. doi: 10.1073/pnas.2628067100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Fukuda T, Yoshida N, Kataoka Y, Manabe R, Mizuno-Horikawa Y, et al. Mice lacking the EDB segment of fibronectin develop normally but exhibit reduced cell growth and fibronectin matrix assembly in vitro. Cancer Res. 2002;62:5603–5610. [PubMed] [Google Scholar]
- 30.Manabe R, Ohe N, Maeda T, Fukuda T, Sekiguchi K. Modulation of cell-adhesive activity of fibronectin by the alternatively spliced EDA segment. J Cell Biol. 1997;139:295–307. doi: 10.1083/jcb.139.1.295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Liao YF, Gotwals P, Koteliansky VE, Sheppard D, Van De Water L. The EIIIA Segment of Fibronectin Is a Ligand for Integrins alpha9beta1 and alpha4beta1 Providing a Novel Mechanism for Regulating Cell Adhesion by Alternative Splicing. J Biol Chem. 2002;277:14467–14474. doi: 10.1074/jbc.M201100200. [DOI] [PubMed] [Google Scholar]
- 32.Serini G, Bochaton-Piallat ML, Ropraz P, Geinoz A, Borsi L, et al. The fibronectin domain ED-A is crucial for myofibroblastic phenotype induction by transforming growth factor-beta1. J Cell Biol. 1998;142:873–881. doi: 10.1083/jcb.142.3.873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.ten Dijke P, Hill CS. New insights into TGF-beta-Smad signalling. Trends Biochem Sci. 2004;29:265–273. doi: 10.1016/j.tibs.2004.03.008. [DOI] [PubMed] [Google Scholar]
- 34.Dallas SL, Sivakumar P, Jones CJ, Chen Q, Peters DM, et al. Fibronectin regulates latent transforming growth factor-beta (TGF beta) by controlling matrix assembly of latent TGF beta-binding protein-1. J Biol Chem. 2005;280:18871–18880. doi: 10.1074/jbc.M410762200. [DOI] [PubMed] [Google Scholar]
- 35.Fontana L, Chen Y, Prijatelj P, Sakai T, Fassler R, et al. Fibronectin is required for integrin alphavbeta6-mediated activation of latent TGF-beta complexes containing LTBP-1. FASEB J. 2005;19:1798–1808. doi: 10.1096/fj.05-4134com. [DOI] [PubMed] [Google Scholar]
- 36.Moriya K, Bae E, Honda K, Sakai K, Sakaguchi T, et al. A Fibronectin-Independent Mechanism of Collagen Fibrillogenesis in Adult Liver Remodeling. Gastroenterology. 2011. [DOI] [PMC free article] [PubMed]
- 37.Kuhn R, Schwenk F, Aguet M, Rajewsky K. Inducible gene targeting in mice. Science. 1995;269:1427–1429. doi: 10.1126/science.7660125. [DOI] [PubMed] [Google Scholar]
- 38.Akiyama SK. Purification of fibronectin. Curr Protoc Cell Biol Chapter. 2001;10:Unit 10 15. doi: 10.1002/0471143030.cb1005s02. [DOI] [PubMed] [Google Scholar]
- 39.Berg RA. Determination of 3- and 4-hydroxyproline. Methods Enzymol. 1982;82(Pt A):372–398. doi: 10.1016/0076-6879(82)82074-0. [DOI] [PubMed] [Google Scholar]
- 40.Stegemann H. [Microdetermination of hydroxyproline with chloramine-T and p-dimethylaminobenzaldehyde.]. Hoppe Seylers Z Physiol Chem. 1958;311:41–45. [PubMed] [Google Scholar]
- 41.Liu W, Hou Y, Chen H, Wei H, Lin W, et al. Sample preparation method for isolation of single-cell types from mouse liver for proteomic studies. Proteomics. 2011;11:3556–3564. doi: 10.1002/pmic.201100157. [DOI] [PubMed] [Google Scholar]
- 42.Inman GJ, Nicolas FJ, Callahan JF, Harling JD, Gaster LM, et al. SB-431542 is a potent and specific inhibitor of transforming growth factor-beta superfamily type I activin receptor-like kinase (ALK) receptors ALK4, ALK5, and ALK7. Mol Pharmacol. 2002;62:65–74. doi: 10.1124/mol.62.1.65. [DOI] [PubMed] [Google Scholar]
- 43.Theret N, Lehti K, Musso O, Clement B. MMP2 activation by collagen I and concanavalin A in cultured human hepatic stellate cells. Hepatology. 1999;30:462–468. doi: 10.1002/hep.510300236. [DOI] [PubMed] [Google Scholar]
- 44.Puchtler H, Waldrop FS, Valentine LS. Polarization microscopic studies of connective tissue stained with picro-sirius red FBA. Beitr Pathol. 1973;150:174–187. doi: 10.1016/s0005-8165(73)80016-2. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.