Potential conflict of interest: Nothing to report.
Liver fibrosis is usually progressive, but it can occasionally be reversible if the causative agents are adequately removed or if patients are treated effectively. However, molecular mechanisms responsible for this reversibility of liver fibrosis have been poorly understood. To reveal the contribution of bone marrow (BM)-derived cells to the spontaneous regression of liver fibrosis, mice were treated with repeated carbon tetrachloride injections after hematopoietic reconstitution with enhanced green fluorescent protein (EGFP)-expressing BM cells. The distribution and characteristics of EGFP-positive (EGFP+) cells present in fibrotic liver tissue were examined at different time points after cessation of carbon tetrachloride intoxication. A large number of EGFP+ cells were observed in liver tissue at peak fibrosis, which decreased during the recovery from liver fibrosis. Some of them, as well as EGFP-negative (EGFP−) liver resident cells, expressed matrix metalloproteinase (MMP)-13 and MMP-9. Whereas MMP-13 was transiently expressed mainly in the cells clustering in the periportal areas, MMP-9 expression and enzymatic activity were detected over the resolution process in several different kinds of cells located in the portal areas and along the fibrous septa. Therapeutic recruitment of BM cells by granulocyte colony-stimulating factor (G-CSF) treatment significantly enhanced migration of BM-derived cells into fibrotic liver and accelerated the regression of liver fibrosis. Experiments using transgenic mice overexpressing hepatocyte growth factor (HGF) indicated that G-CSF and HGF synergistically increased MMP-9 expression along the fibrous septa. Conclusion: Autologous BM cells contribute to the spontaneous regression of liver fibrosis, and their therapeutic derivation could be a new treatment strategy for intractable liver fibrosis. (HEPATOLOGY; 2007:213–222.)
Liver fibrosis is a pathological condition characterized by an excessive deposition of collagen and other components of extracellular matrix. It was originally considered progressive and irreversible, but several clinical studies have shown that it can be reversed if the causative agents are adequately removed or if patients are treated effectively.1–5 However, molecular mechanisms responsible for this reversibility have been poorly understood, except that apoptosis of activated hepatic stellate cells (HSCs), the major producers of collagen and tissue inhibitors of matrix metalloproteinase (MMP) in fibrotic liver, has been implicated in the regression of experimental liver fibrosis.6
We have been studying the expression of MMPs in the liver—in particular, their role in the regression of liver fibrosis.7–9 Carbon tetrachloride (CCl4)-induced liver fibrosis is a well-known animal model, in which resolution of dense fibrous bands occurs after cessation of CCl4 intoxication.1 Using this model, we have found that MMP-13 gene coding for the major interstitial collagenase in rodents is transiently expressed in the early phase of resolution process.10 However, most of the MMP-13 gene-expressing cells were negative for known markers of liver resident cells, except that only some of them coexpressed α-smooth muscle actin (α-SMA), a marker of activated HSCs.10 We therefore speculated about the possible contribution of bone marrow (BM)-derived cells to MMP-13 expression.11
The present study was conducted to reveal a more direct contribution of BM-derived cells to the regression of liver fibrosis. For this purpose, mouse BM was reconstituted with enhanced green fluorescent protein (EGFP)-expressing cells, and the animals were subsequently treated with repeated CCl4 injections. The results indicated that a large number of EGFP+ cells migrated into fibrotic liver, and some of them—in addition to EGFP− liver resident cells—expressed MMP-13 and MMP-9. Granulocyte colony-stimulating factor (G-CSF) treatment significantly enhanced migration of BM-derived cells and accelerated the regression of liver fibrosis. Interestingly, overexpression of hepatocyte growth factor (HGF), together with G-CSF treatment, synergistically increased MMP-9 expression along the fibrous septa, indicating their potential application for the treatment of liver fibrosis.
C57BL/6 mice were purchased at 6 weeks of age from CLEA Japan Inc. (Tokyo, Japan). Transgenic mice overexpressing human HGF under control of the mouse albumin enhancer and promoter have been reported previously.12 C57BL/6 transgenic mice that ubiquitously expressed EGFP by the CAG expression unit13 were used as BM donors, whereas C57BL/6 wild-type mice and transgenic HGF mice were used as recipients in BM transplantation. All animals received humane care, and the experiments were approved by the Animal Experiment Committee of Tokai University.
A total of 45 wild-type mice and 25 transgenic HGF mice were administered a single 9.5-Gy dose of whole body irradiation. Five million unfractionated BM cells from transgenic EGFP mice were injected intravenously to the irradiated recipients. Recipient mice showing more than 70% of the EGFP+ chimerism in the peripheral blood determined by fluorescence-activated cell-sorter scanner analyses were used for further experiments. Some mice were injected with BM cells without whole body irradiation as a control.
Experimental Liver Fibrosis.
Two months after transplantation, when BM was completely reconstituted with EGFP+ cells, recipient mice started to be injected subcutaneously with 1 mL/kg body weight of CCl4 mixed with olive oil every 3 days for 90 days.14 They received CCl4 by changing the sites of injection to avoid necrosis of the local skin and to obtain invariable results. Control mice were injected with olive oil only. During the last 7 days before the final CCl4 injection, some mice were subcutaneously injected every day with 100 μg/kg body weight of recombinant human G-CSF (Kyowa Hakko Kogyo Co. Ltd., Tokyo, Japan), whereas control animals were injected with 0.1% mouse serum albumin in phosphate-buffered saline. Liver tissues were obtained from three to four mice in each group on the indicated days after the last CCl4 injection, and the sections were subjected to Sirius red-Fast green FCF staining.15 The degree of liver fibrosis was semiquantified by measuring the relative areas of fibrosis with the aid of computer software as described previously.10
Total RNA was isolated from liver tissue, and complimentary DNA was synthesized by reverse transcription. The PCR primers and conditions to detect MMP-13, MMP-9, and GAPDH gene expression are shown in Table 1. The number of PCR cycles for detecting each gene expression was determined to semiquantify the amount of messenger RNA within a linear range.
Table 1. Primers and RT-PCR Conditions Used for Detection of MMP-13, MMP-9, and GAPDH Gene Expression
Enzymatic activity of MMPs in liver tissue was detected by using MMP in situ Zymo-Film (Wako Pure Chemical Industries Ltd., Tokyo, Japan) as described previously.16 A serial section was incubated in the presence of an MMP inhibitor, 1,10-phenanthroline (Wako Pure Chemical Industries Ltd.), to exclude nonspecific proteolytic activities.
EGFP+ cells were observed and analyzed using a confocal laser-scanning microscope, LSM 510 META (Carl Zeiss, Jena, Germany). To detect MMP-13 expression, sections were incubated with anti-mouse MMP-13 antibodies (Lab Vision, Fremont, CA) using the Mouse-on-Mouse Immunodetection Kit (Vector Laboratories, Burlingame, CA). After the secondary antibody reaction with biotinylated anti-mouse IgG, sections were incubated with streptavidin-conjugated Alexa Fluor 660 (Molecular Probes, Eugene, OR). For the detection of MMP-9, sections were incubated with anti-mouse MMP-9 antibodies (R&D Systems, Minneapolis, MN), followed by a reaction with Alexa Fluor 555-conjugated anti-goat IgG (Molecular Probes). Other antibodies recognizing CD34 (BD PharMingen, San Jose, CA), Ov-6 (a generous gift from Dr. Stewart Sell), F4/80 (Serotec, Raleigh, NC), α-SMA (Sigma, St. Louis, MO), α-fetoprotein (Santa Cruz Biotechnology Inc., Santa Cruz, CA), or albumin (Biogenesis Ltd., Poole, UK) were also used with appropriate secondary antibodies to analyze differentiation of BM cells. The emission fingerprinting method17 was used to distinguish specific fluorescent signals from the background autofluorescence.
Data are expressed as the mean ± SEM. The Mann-Whitney U test was used to compare differences between groups. A P value of less than 0.05 was considered statistically significant.
Detection of Specific Fluorescent Signals in Fibrotic Liver Tissue.
Confocal laser-scanning microscopic examination revealed a large amount of autofluorescence in fibrotic liver tissue, especially in the portal areas (Fig. 1). Because the autofluorescent signals were barely distinguishable from the specific EGFP signals, we employed an emission fingerprinting method as described previously.17 Meta-analyses of fluorescence over a wide range of wavelengths successfully identified specific fluorescence with a single peak of a certain wavelength and eliminated the background autofluorescence. The method enabled the analysis of actual EGFP+ cells in the liver with high specificity (Fig. 1).
BM Cells Migrated Into Fibrotic Liver.
We first examined the degree of fibrosis and presence of EGFP+ BM-derived cells in CCl4-treated mouse liver. After a total of 30 repeated CCl4 injections over 90 days, bridging fibrosis connecting the neighboring portal areas and central veins was invariably formed in mice (Fig. 2B), which are generally considered less sensitive to CCl4 than rats. There were no EGFP+ cells detected in control mouse liver without whole body irradiation in either untreated or CCl4-treated condition (data not shown). In contrast, a small number of EGFP+ cells were observed in the nonfibrotic liver tissue of another control mice, who received only olive oil after BM reconstitution (Fig. 2A,D). A considerable number of EGFP+ cells migrated into fibrotic liver 2 days after the last CCl4 injection, the time point of peak fibrosis (Fig. 2E). Then, they decreased within several days (Fig. 2F) in parallel with the resolution of fibrous bands (Fig. 2C). The distribution of EGFP+ cells was essentially overlapped with the portal areas and the fibrous bands, except that there were a relatively small number of cells present within the parenchyma as well. Interestingly, the EGFP+ cells exhibited different morphological features depending on their localization in the liver. Most of the cells observed in the portal areas had small cytoplasm with a round-shaped nucleus (Fig. 3A), whereas cells present within the parenchyma showed a thin mesenchymal cell-like appearance (Fig. 3C). Cells located along the fibrous septa exhibited an intermediate semiround or spindle shape (Fig. 3B).
BM-Derived Cells Differentiated in Fibrotic Liver.
Next, we explored which cell lineages could be differentiated by the BM-derived EGFP+ cells. There were very small (5-10 μm in diameter), round-shaped cells with a high nuclear/cytoplasmic ratio clustering in the periportal areas. Some of them coexpressed EGFP and CD34 (Fig. 4A), indicating that they were stem/progenitor cells derived from BM. In addition, there was a limited number of EGFP+ cells (0.5 ± 0.2 cells per portal tract) showing positive staining for Ov-6, a marker of oval cells (Fig. 4B). However, none of the EGFP+ cells coexpressed α-fetoprotein or albumin (data not shown). On the other hand, approximately half of the cells present within the parenchyma were stained positive for F4/80, a marker of macrophage/Kupffer cells (Fig. 4C), but negative for α-SMA staining (Fig. 4D).
MMP-13 Was Transiently Expressed in Fibrotic Liver.
Next, we analyzed MMP expression in BM-derived cells to investigate their possible contribution to the spontaneous regression of liver fibrosis. Gene expression of MMP-13, the major interstitial collagenase in rodents, was barely detected in control normal liver tissue. It showed the highest levels of expression 2 days after the last CCl4 injection (Fig. 5A). Immunofluorescence study revealed that MMP-13–expressing cells were observed mainly in the portal areas (Fig. 5B). Approximately half of them coexpressed EGFP, indicating their BM origin. MMP-13–expressing cells clustering in the periportal areas (Fig. 5B) exhibited features similar to the CD34-positive stem/progenitor cells shown in Fig. 4A. However, the MMP-13–expressing cells did not coexpress CD34, indicating that they are not at exactly the same differentiation stage. Other MMP-13–positive cells observed in the portal areas had slightly larger cytoplasm with a semiround shape (Fig. 5B). They were negative for any known markers of liver resident cells, including α-SMA (data not shown).
MMP-9 Was Expressed During Recovery From Liver Fibrosis.
In contrast to the transient MMP-13 expression, MMP-9 was expressed in several different kinds of cells throughout the resolution process. Overall, approximately half of the MMP-9–positive cells coexpressed EGFP. MMP-9–expressing cells observed in the portal areas were neutrophils with segmental nuclei (Fig. 6A), whereas those in the parenchyma were F4/80-positive Kupffer cells (Fig. 6B). In addition to those known MMP-9 producers, some EGFP+ cells present in the periportal areas coexpressed CD34 and MMP-9 (Fig. 6C), indicating that they were hematopoietic and/or hepatic stem/progenitor cells. MMP-9–expressing cells with a semiround shape were also observed along the fibrous septa (Fig. 6D), which were negative for any known markers of liver resident cells. Enzymatic activity of MMPs was confirmed via in situ zymography (Fig. 7A), which was completely diminished in the presence of an MMP inhibitor (Fig. 7B).
G-CSF Treatment Enhanced Migration of BM Cells and Regression of Liver Fibrosis.
Administration of G-CSF into CCl4-treated recipient mice significantly increased the number of EGFP+ cells migrating into liver 2 days and 5 days after the last CCl4 injection (Fig. 8A -B). In parallel with this enhanced migration of BM-derived cells, the mean fibrotic area in G-CSF–treated mice was significantly smaller than that in untreated animals 5 days after the last CCl4 injection (Fig. 8C-D). These results indicated the accelerating resolution of fibrous bands by G-CSF treatment.
G-CSF and HGF Synergistically Enhanced MMP-9 Expression.
In the last set of experiments, we evaluated the effect of G-CSF and HGF on MMP expression. G-CSF treatment and/or HGF overexpression did not affect MMP-13 expression (data not shown). Treatment with G-CSF enhanced MMP-9 gene expression in both wild-type and transgenic HGF mice, whereas HGF overexpression itself did not affect the gene expression (Fig. 9A). Then, immunofluorescence study was performed to examine cellular localization of MMP-9. G-CSF treatment significantly increased the number of MMP-9–expressing cells present along the fibrous septa, mostly EGFP+ cells derived from BM (Fig. 9B-C). In contrast, and consistent with the results of RT-PCR, overexpression of HGF did not change the number of MMP-9–expressing cells by itself (data not shown). However, when transgenic HGF mice were treated with G-CSF, they synergistically enhanced MMP-9 expression along the fibrous septa (Fig. 9B-C). Interestingly, the combination of these 2 factors significantly increased MMP-9 expression not only in the EGFP+ BM-derived cells but also in EGFP− liver resident cells (Fig. 9C).
There is a dynamic balance between the production and degradation of collagen, and a disruption of this equilibrium results in fibrosis in various organs, including the liver. Thus, either suppression of accelerated collagen production or relative overexpression of MMP could be useful to reverse liver fibrosis. Indeed, adenovirus-mediated gene transfer of human MMP-118 or MMP-819 suppressed progression of experimental fibrosis in rats. In the present study, we examined whether autologous BM cells served as a source of MMPs and contributed to the spontaneous regression of liver fibrosis. The results indicated that a large number of BM-derived cells migrated into fibrotic liver, some of which expressed MMP-13 and MMP-9 during the regression process. In addition, enzymatic activity of MMPs was confirmed via in situ zymography.
EGFP expression system has a great advantage to detect the fluorescent signal with extremely high specificity, which may be lost when using immunohistochemical staining of EGFP. On the other hand, observation of EGFP and other specific fluorescence is difficult due to the presence of autofluorescence in the case of CCl4-induced fibrotic liver, leading to misinterpretation of the obtained results.20, 21 To overcome this problem and identify only the specific fluorescence, we used an emission fingerprinting method and successfully eliminated the background autofluorescence. The obtained results are therfore convincing in terms of specificity of fluorescent signals (Fig. 1).
In addition to their potential to differentiate into parenchymal hepatocytes or fuse with resident cells, BM-derived cells sometimes differentiate into mesenchymal cell lineages such as endothelial cells,22, 23 Kupffer cells,24 and HSCs,25–27 indicating either their pluripotential or the presence of many kinds of progenitors in BM. In the present study, although some of the BM-derived cells migrating into fibrotic liver were CD34-positive or Ov-6–positive stem/progenitor cells, they seldom differentiated into parenchymal hepatocytes. Instead, a small population of BM-derived cells expressed F4/80, a marker of macrophage/Kupffer cells. In contrast to the results of previous studies, there were few if any BM-derived cells differentiating into either mature endothelial cells or activated HSCs. Those findings might be attributed to the differences in the experimental models and severity of liver injury between our study and others, as well as to the specificity of EGFP signal described above.
In addition to its collagenase activity, transient MMP-13 expression by BM-derived cells may trigger a cascade of signal activation, including the expression of other MMPs. Consistent with this hypothesis, it was previously shown that forced expression of human MMP-8, another interstitial collagenase, increased expression of MMP-2, MMP-3, and MMP-9 as well as HGF, and suppressed experimental liver fibrosis in rats.19 Similarly, our own study revealed that adenovirus-mediated gene transfer of human MMP-13 into rat fibrotic liver increased expression of MMP-9 and HGF, and subsequently accelerated resolution of fibrous bands (Watanabe et al., manuscript submitted for publication). The lack of cells coexpressing MMP-13 and α-SMA does not deny that HSCs have the potential to produce MMP-13, but it may reflect their heterogeneity in the origin and phenotype.
In contrast to the transient MMP-13 expression, MMP-9 was expressed for a longer period over the resolution process. In addition to its ability to degrade collagen and other matrix components, hepatic MMP-9 is also known as a potent factor to recruit BM cells to injured liver.28, 29 It is therefore postulated that MMP-9 expression in BM-derived cells in fibrotic liver further enhances their own migration and accelerates the resolution of fibrous bands. It should be noted that the expression of MMP-13 and MMP-9 was observed not only in the EGFP+ BM-derived cells but also in the EGFP− liver resident cells. Considering that EGFP+ chimerism was higher than 70%, it is unlikely that the finding merely represented the relative ratio of EGFP+ cells in BM. Instead, it is possible to argue that BM-derived cells migrating into fibrotic liver secrete unknown factors that initiate the MMP activation cascade in themselves and the neighboring resident cells.
After confirming the migration of BM-derived cells and their MMP expression, we next sought the factors that accelerate recruitment of BM cells and stimulate MMP expression in fibrotic liver. G-CSF treatment significantly enhanced migration of BM cells into fibrotic liver and increased their MMP-9 expression. This was followed by accelerated resolution of fibrous bands. These findings suggest that G-CSF not only recruits BM cells into the peripheral blood but also affects their migration and phenotypic change in the fibrotic liver. It has been recently reported that G-CSF–primed hematopoietic stem cells or G-CSF itself accelerates recovery and improves survival of mice with CCl4-induced liver injury/fibrosis.30
In addition to MMP-9, HGF is also known as an important factor to recruit BM cells to injured liver.28 The present study showed that HGF overexpression by itself did not enhance MMP-9 expression. However, overexpression of HGF, together with G-CSF treatment, synergistically stimulated MMP-9 expression in fibrotic liver. Interestingly, the combination of those two factors increased the number of both EGFP+ and EGFP− MMP-9–expressing cells. These results may suggest the yet unknown synergistic mechanism between G-CSF and HGF.
It has been reported recently that infusion of BM cells from syngeneic mice into animals with cirrhosis resulted in migration of those cells into fibrotic liver and suppression of the fibrosis by expressing MMP-9.31 These results indicate that BM transplantation could be a treatment option for advanced cirrhosis in humans. However, if the therapeutic derivation of autologous BM cells and their differentiation into MMP-expressing cells can be effectively achieved, it will obviously be a safer and less invasive approach for the treatment of liver fibrosis.
In conclusion, the present study has shown that BM-derived cells, as well as liver resident cells, express MMP-13 and MMP-9 and contribute to the spontaneous regression of liver fibrosis. G-CSF treatment significantly enhanced migration of BM cells into fibrotic liver and, in combination with HGF, synergistically stimulated MMP-9 expression in both BM-derived cells and liver resident cells. Work is now in progress in our laboratory to determine the nature of MMP-13–expressing cells and the initial factors stimulating its expression, as well as the mechanisms of interaction between BM-derived cells and liver resident cells. Those studies not only lead to a better understanding of pathophysiological roles of BM-derived cells in the liver, but also contribute to the establishment of a potential new treatment strategy for intractable liver fibrosis.
We are indebted to Dr. Stewart Sell for providing us with valuable antibodies; Dr. Masaru Okabe for his generous gift of EGFP mice; Dr. Johbu Itoh for technical assistance and suggestions on confocal laser-scanning microscopic examination; and Dr. Scott L. Friedman for his helpful advice and critical suggestions.