Recent reports have shown the capacity of the bone marrow cell (BMC) to differentiate into a variety of non-hematopoietic cell lineages.1–5 These results indicate that the BMC is an attractive cell source for regenerative medicine compared with tissue-specific stem cells.6 The capacity of the BMC to differentiate into hepatocytes and intestinal cells has been shown by Y-chromosome detection in autopsy analysis of human female recipients of BMCs from male donors.7, 8 Although Lagasse et al. reported that purified hematopoietic stem cells could differentiate into hepatocytes using a fumarylacetate hydrase-deficient model,5 Wagers et al. showed little evidence of plasticity in adult hematopoietic stem cells.9 Thus, although there is still controversy about which part of BMCs can differentiate into hepatocytes, the BMC seems to have the plasticity to differentiate into such cells. From the point of view of therapy, one of the targets of liver disease for BMC transplantation is liver cirrhosis with chronic liver failure. This is an unphysiological condition with excessive deposition of extracellular matrix and a relative lack of parenchymal cells (hepatocytes). Even if BMC transplantation is successful in supplying parenchymal cells, the fate of the extracellular matrix under these conditions is unknown. The present study clearly shows that transplanted BMCs reduce (degrade) carbon tetrachloride (CCl4)-induced liver fibrosis with a significantly improved survival rate.
We investigated the effect of bone marrow cell (BMC) transplantation on established liver fibrosis. BMCs of green fluorescent protein (GFP) mice were transplanted into 4-week carbon tetrachloride (CCl4)–treated C57BL6 mice through the tail vein, and the mice were treated for 4 more weeks with CCl4 (total, 8 weeks). Sirius red and GFP staining clearly indicated migrated BMCs existing along with fibers, with strong expression of matrix metalloproteinase (MMP)-9 shown by anti–MMP-9 antibodies and in situ hybridization. Double fluorescent immunohistochemistry showed the expression of MMP-9 on the GFP-positive cell surface. Film in situ zymographic analysis revealed strong gelatinolytic activity in the periportal area coinciding with the location of MMP-9–positive BMCs. Four weeks after BMC transplantation, mice had significantly reduced liver fibrosis, as assessed by hydroxyproline content of the livers, compared to that of mice treated with CCl4 alone. Subpopulation of Liv8-negative BMCs was responsible for this fibrolytic effect. In conclusion, mice with BMC transplants with continuous CCl4 injection had reduced liver fibrosis and a significantly improved survival rate after BMC transplantation compared with mice treated with CCl4 alone. This finding introduces a new concept for the therapy of liver fibrosis. Supplementary material for this article can be found on the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2004;40:1304–1311.)
Materials and Methods
GFP-transgenic mice (TgN(β-act-EGFP)Osb) were kindly provided by Masaru Okabe (Genome Research Center, Osaka University, Osaka, Japan).10 C57BL6 female mice were purchased from Japan SLC (Shizuoka, Japan). Mice were properly anesthetized during experiments.
Six-week-old female C57BL6 mice were treated with 1 mL/kg CCl4 dissolved in olive oil (1:1) twice a week for 4 weeks. One day (24 hours) after the eighth injection of CCl4, 1 × 105 green fluorescent protein (GFP)-positive BMCs or sorted Liv8-positive or Liv8-negative BMCs (1 × 105 cells) or same volume of saline as a control (described also as mice treated with CCl4 alone) were injected into the tail vein as described previously.11, 12 Mice continued to be treated with CCl4. After 1, 2, 3, or 4 weeks, mice were then sacrificed to assess the extent of liver fibrosis. For examination of the survival rate, mice were treated with CCl4 for 4 weeks and divided into 2 groups (15 mice each) with bone marrow transplantation or the same volume of saline injection. All mice were then treated with CCl4 for a further 25 weeks.
For BMC isolation, GFP-transgenic mice (TgN(β-act-EGFP)Osb) (6 weeks old) were killed by cervical dislocation and the limbs removed. GFP-positive BMCs were flushed with Dulbecco's Modified Eagle medium (DMEM) culture medium with 10% fetal bovine serum (FBS) from the medullary cavities of tibias and femurs using a 25-G needle.
Production of Rat Monoclonal Antibody, Liv8.
Eight-week old WKY/NCrj female rats were immunized in the hind footpads with 100 μg of E11.5 murine fetal liver lysate in complete Freund's adjuvant (0.2 mL). Anti-Liv8 antibodies were raised according to a previously described protocol.13
Fluorescence-Activated Cell Sorter Analysis of Fetal Liver Cells and BMCs Using Anti-Liv8 Antibody.
Prepared mouse fetal liver cells (E11.5) and adult BMCs were reacted with biotin-conjugated anti-Liv8 antibody,12 phycoerythrin-conjugated rat anti-CD45 (Becton Dickinson Bioscience, San Jose, CA), fluorescein isothiocyanate-conjugated anti–c-kit (Becton Dickinson Bioscience), phycoerythrin-conjugated anti–Thy 1 (Becton Dickinson Bioscience), and fluorescein isothiocyanate-conjugated anti-B220 antibodies (Becton Dickinson Bioscience) at the rate of 1 μg per 106 total cells, mixed well, and incubated in the tube for 30 to 40 minutes at 4°C. Following the incubation with the first antibody, the cells were washed twice by 0.02 mol/L phosphate-buffered saline (PBS) and centrifuged at 500g for 5 minutes. Labeled cells were then reacted to allophycocyanin-conjugated streptavidin (Becton Dickinson Bioscience) at the rate of 1 μg per 106 total cells, mixed well, and incubated in the tube for 30 to 40 minutes at 4°C. After that, these were washed out once with 0.02 mol/L PBS and centrifuged at 500g for 5 minutes. The labeled cells were analyzed using FACS Calibur (Becton Dickinson Bioscience).
Preparation of Liv8-positive and Liv8-negative BMCs
Liv8-positive and Liv8-negative BMCs were prepared as described previously.12 Briefly, prepared BMCs were reacted to rat anti-Liv8 immunoglobulin G (IgG) antibody at the rate of 1 μg per 106 total cells, mixed well, and incubated in the tube for 30 minutes at 4°C. Cells were then washed twice by 0.02 mol/L PBS and centrifuged at 500g for 5 minutes. Cells were labeled with rat anti-Liv8 IgG antibody by reacting with goat anti-rat IgG MicroBeads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) at the rate of 20 μL per 107 total cells, mixed well, and incubated for 20 minutes at 4°C. Labeled cells were washed once by 0.02 mol/L PBS and centrifuged at 500g for 5 minutes. These cells were separated into Liv8-positive cells or Liv8-negative cells by the autoMACS magnetic cell sorting system (Miltenyi Biotec GmbH) for 10 minutes per tube.
Tissue Preparation and Immunohistochemistry.
The liver was perfused via the heart with 4% paraformaldehyde to flush out blood cells and incubated with 4% paraformaldehyde overnight. Tissues were then soaked in 30% sucrose for 3 days. Tissues were frozen with liquid nitrogen to prepare for sectioning with a cryostat for immunohistochemistry.
Cells expressing GFP and matrix metalloproteinase (MMP)-9 (or α-smooth muscle actin) were analyzed by both fluorescent microscopy and conventional immunohistochemistry using anti-GFP, anti–MMP-9 (Santa Cruz Biotechnology, Santa Cruz, CA), and anti–α-smooth muscle actin antibodies (Sigma-Aldrich, St. Louis, MO). Tissues were soaked in 0.3% Triton X-100 with 0.05% normal goat serum (NGS) (Chemicon, Temecula, CA) or normal rabbit serum (NRS) (Chemicon) in PBS overnight. The next day, the tissues were put in 500 mL of 10% NGS or NRS in 0.3% Triton X-100 of PBS for 2 hours, then washed with 0.3% Triton X-100 with 0.05% NGS or NRS in PBS for 10 minutes. We soaked the tissues in 1.5% H2O2 in 50% methanol with distilled water for 2 hours. The tissues were then washed in 0.3% Triton X-100 with 0.05% NGS or NRS in PBS. Sections were incubated with anti-GFP and anti–MMP-9 (ICN Pharmaceuticals Inc., Kanagawa, Japan) antibodies. Anti–biotin-conjugated anti-goat IgG, anti-rabbit IgG, biotin-conjugated rabbit anti-goat IgG, and biotin-conjugated rabbit anti-mouse IgG were purchased from Dako Japan (Kyoto, Japan) and used as the secondary antibodies. PAP-goat (B0157), PAP-mouse (B0650), and PAP-rabbit (Z0113) polyclonal antibodies (Dako Japan) were used as third antibodies.
For fluorescent immunohistochemistry, we used Alexa Fluor R 488 and 568 donkey anti–goat- or anti–rabbit- or anti–mouse-IgG (H + L)-conjugated antibodies (Molecular Probe Inc., Eugene, OR) as second antibodies.
For the evaluation of fibrosis, picro-sirius red staining was performed using 0.1% picro-sirius red solution as previously described.14
Quantitative Analysis of Liver Fibrosis.
We quantified the liver fibrosis area with picro-sirius red staining using an Olympus Provis microscope equipped with a CCD camera (Tokyo, Japan), as described previously.15 Briefly, the red area, considered the fibrotic area, was assessed by computer-assisted image analysis with MetaMorph software (Universal Imaging Corporation, Downingtown, PA) at a magnification of ×40. The mean value of 6 randomly selected areas per sample was used as the expressed percent area of fibrosis.
Microarray analysis was performed as described previously.16
Briefly, total RNA of liver was isolated using the Atlas Pure Total RNA labeling system (Clontech Laboratories, Inc.) from mice 1 week after BMC transplantation (n = 3) or from mice treated with CCl4 for 5 weeks (n = 3) according to the manufacturer's recommendations. Differential hybridization analysis was done using an Atlas Mouse complementary DNA expression array (BD Bioscience Clontech, Tokyo, Japan). Complementary DNA probe preparation and hybridization were done according to the manufacturer's recommendations. The array results were scanned with a Strom 840 PhosphoImager (Molecular Dynamics, Sunnyvale, CA) and analyzed with Atlas Image software (BD Bioscience Clontech). The results show the mean values of 3 mice in each group.
Hydroxyproline content was determined by a modification of Kivirikko's method, as previously reported.17 Briefly, liver specimens were weighed, and 20 mg of the freeze-dried sample was hydrolyzed in 6 mol/L HCl at 110°C in an autoclave at a pressure of 1.2 kg force/cm2 for 24 hours. After centrifugation at 2,000 rpm at a temperature of 4°C for 5 minutes, 2 mL of supernatant was mixed with 50 mL of 1% phenolphthalein and 8 N KOH to obtain a total volume of 5 mL of liquid at pH of 7 to 8. Absorbance was measured at 560 nm. The hydroxyproline content of the liver was expressed as micrograms per gram of wet weight.
In Situ Hybridization.
In situ hybridization was performed essentially as described previously.18 Briefly, digoxigenin (DIG)-11-UTP–labeled single-stranded RNA probes were prepared with DIG RNA labeling mix and the corresponding T3 or T7 RNA polymerase (Boehringer Mannheim Japan, Tokyo, Japan) according to the manufacturer's instructions. The mouse MMP-9 probe was a 150–base pair fragment from the 3′ untranslated region cloned in the pBluescript (Stratagene, Tokyo, Japan) vector. In situ hybridization was performed on tissue sections placed on Superfrost Plus slides postfixed in 4% paraformaldehyde in PBS, rinsed in PBS containing 0.1% active diethyl pyrocarbonate, and prehybridized for 2 hours at 58°C in 50% formamide, 5 × SSC (standard saline citrate), and 40 μg of salmon-sperm DNA per milliliter. Hybridization was carried out at 58°C for 16 hours in a humid chamber with 400 ng of DIG-labeled probe per milliliter diluted in the same solution used for prehybridization. After hybridization, the sections were successively washed in 2 × SSC at room temperature for 30 minutes, 2 × SSC for 1 hour at 65°C, and 0.1 × SSC at 65°C for 1 hour. For the reaction of anti-DIG antibodies, slides were preincubated in buffer A (100 mmol/L Tris, 150 mmol/L NaCl [pH 7.5]), and then with an alkaline phosphatase-coupled anti-DIG antibody (Boehringer Mannheim Japan) diluted 1:5,000 in buffer A containing 0.5% Boehringer blocking reagent for 2 hours at room temperature. The slides were washed in buffer A and then preincubated in buffer C (100 mmol/L Tris, 50 mmol/L MgCl2 [pH 9.5]). Alkaline phosphatase was then revealed as described for 16 to 24 hours at room temperature. The enzymatic reaction was stopped with Tris-ethylenediaminetetraacetic acid (EDTA) for 15 minutes. The slides were rinsed in water for several hours and then dried, cleared in xylene, and mounted directly.
In Situ Zymography.
In situ zymography was performed as described.19
The fresh specimens of CCl4 treated with BMC-transplanted liver tissues (1 week after BMC transplantation) were embedded without fixation in Tissue-Tek optimal cutting temperature compound (Miles, Elkhart, IN). Serial frozen sections were made using a cryostat (MicroM, Walldorf, Germany) and mounted on gelatin films that were coated with 7% gelatin solution (Fuji Photo Film, Tokyo, Japan). The films with sections were incubated for 24 hours at 37°C in a moisture chamber and stained with 1.0% amido black 10B. The gelatin in contact with the proteolytic areas of the sections was digested, and thus zones of enzymic activity were indicated by negative staining. The digested areas in the sections were compared with serial sections stained with hematoxylin-eosin. As a control, liver tissues treated with CCl4 alone (5 weeks) were used, and the frozen sections were treated in a manner similar to that already described.
Results are presented as the mean ± SD. Differences between groups were analyzed by 1-way ANOVA.
The survival rate was examined using the Breslow-Gehan-Wilcoxon test.
This experiment was reviewed by the Committee of Animal Experiment Ethics at the Yamaguchi University School of Medicine and was carried out under the guidelines for animal experiments at Yamaguchi University School of Medicine (no. 105).
Five weeks after CCl4 injection, liver fibrosis was already seen (Supplementary Fig. 1A). One week after BMC transplantation (5 weeks after CCl4 injection), BMCs were seen along with the fibers recognized by light red staining (black arrows), different from hepatocytes (Fig. 1A) with sirius red staining. More BMCs were seen after 2 weeks (Figs. 1B and 2A,C) and 3 weeks (Figs. 1C and 2B,D), and large spheroid-shaped cells (blue arrows) and small cells (green arrows) (Fig. 2B) were found in the area presumably occupied by fibers (Fig. 2D), shown by sirius red and GFP staining.
Surprisingly, 4 weeks later, the BMC-transplanted liver clearly showed reduction of liver fibrosis (Fig. 1D) compared with the liver treated with CCl4 alone at 8 weeks (Supplementary Fig. 1D), although CCl4 was injected throughout the experimental period. Quantitative image analysis of liver fibrosis indicated that the percent area of liver fibrosis at 1 week after BMC transplantation was 5.36% ± 0.90%, but at 4 weeks after transplantation it was significantly decreased to 4.16% ± 0.53% (P < .001, n = 8 each; Fig. 3).
Treatment with CCl4 alone for 8 weeks showed an increased hydroxyproline content of 630 ± 93 μg/wet g liver (Table 1). BMC transplantation significantly reduced this to 392 ± 59 μg/wet g liver 4 weeks later (P < .01, n = 8 each). This hydroxyproline content was significantly reduced even compared with that of 1 week after BMC transplantation (494 ± 74 μg/wet g liver, P < .05).
|Treatment(No. of Mice)||Hydroxyproline (μg/g Liver)|
|CCl4, 5 wk (8)||464 ± 93|
|CCl4, 8 wk (8)||630 ± 93|
|CCl4/BMT, 5 wk (8)||494 ± 74|
|CCl4/BMT, 8 wk (8)||392 ± 59*†|
The mouse fetal liver at E11.5 functions as a definitive hematopoietic organ, and Liv8-positive cells of the fetal liver at E11.5 include c-kit–positive immature hematopoietic cells and CD45-positive lymphoid cells. These results indicate that almost all Liv8-positive cells are of hematopoietic origin (Supplementary Fig. 2).
In addition, all c-kit–positive mouse adult BMCs belong to the Liv8-positive fraction, and Liv8-positive BMCs include almost all of the CD45- and Thy-1–positive BMCs, in addition to B220, a marker of B lymphocytes (Fig. 4).
These results strongly suggest that Liv8-positive cells include both immature and mature hematopoietic cells.
Liv8-negative BMCs significantly reduced liver fibrosis compared with that of the liver treated with CCl4 alone for 8 weeks, although Liv8-positive BMCs had no effect on liver fibrosis (Table 2).
|Treatment (No. of mice)||Hydroxyproline (μg/g Liver)|
|CCl4 (8)||687 ± 102|
|CCl4/Liv8-positive (8)||638 ± 94|
|CCl4/Liv8-negative (8)||415 ± 77*|
Microarray analysis of the liver 1 week after BMC transplantation indicated increased expression of MMP-2, MMP-9 and MMP-14 with decreased expression of tissue inhibitor of metalloproteinase-3 (TIMP-3) compared with that of the liver treated with CCl4 alone for 5 weeks (Table 3). Because the expression of MMP-9 was marked, we investigated it in this model.
|CCl4 + BMC vs. CCl4 Alone|
Immunohistochemistry of MMP-9 showed localization of these cells similar to that of transplanted BMCs (Fig. 5A). However, the liver treated with CCl4 alone showed only a few MMP-9–positive nonparenchymal cells (black arrows, Fig. 5B).
The expression of MMP-9 with in situ hybridization coincided with the immunohistochemical staining of MMP-9 (Fig. 5C).
Although double fluorescent-positive cells were not seen in the liver treated with CCl4 alone (Fig. 6A), double-positive yellow-colored cells (black arrows) were seen in the BMC-transplanted liver (Fig. 6B). With high magnification, double fluorescent immunohistochemistry showed the expression of MMP-9 (red) on the GFP-positive (green) cell surface (Fig. 6C).
A double fluorescent (anti-GFP with green color and anti–α-smooth muscle actin with red color) study indicated that a fine network pattern of stellate cells (red) existed in the liver treated with CCl4 alone for 5 weeks (Fig. 7A). Conversely, GFP-positive green-colored cells (green arrow) were seen with a reduced fine network pattern (red arrow) in the liver 1 week after BMC transplantation (Fig. 7B). Double fluorescent-positive yellow-colored cells (yellow arrow), presumably stellate cells, were then seen without the fine network pattern (red arrow) in the liver 2 weeks after BMC transplantation (Fig. 7C). The shape of these cells was different from other GFP-positive cells, and the number of the double-positive cells was very small.
Next, we examined the direct activity of MMP-9 using in situ zymography. Film in situ zymographic analysis revealed strong gelatinolytic activity in the periportal area coinciding with the location of MMP-9–positive BMCs compared with the liver treated with CCl4 alone (Fig. 8).
This gelatinolytic activity was completely blocked by the addition of 1,10-phenanthroline, an MMP inhibitor (data not shown).
Finally, the mice that underwent BMC transplantation with continuous CCl4 injection showed a gradually increased serum albumin level (Supplementary Fig. 3) resulting in a significantly improved survival rate after BMC transplantation compared with mice treated with CCl4 alone (Supplementary Fig. 4).
In this report, transplanted BMCs can degrade collagen fibers and clearly reduce liver fibrosis with strong expression of MMPs, especially MMP-9, as indicated by both in situ zymography and the double staining of GFP and MMP-9 using fluorescent microscopy. The reason for the strong expression of MMP-9 is still unknown. However, Heissing et al.20, 21 recently reported that MMP-9 induced in BMCs released soluble Kit-ligand, which might be related to the transfer of stem cells in BMCs to the proliferative niche. Therefore, MMP-9 in our model could play an important role in the degradation of extracellular matrix and also by releasing some factors, e.g., soluble Kit-ligand, related to the differentiation and proliferation of transplanted BMCs in liver inflammation induced by continuous injection of CCl4. It has also been shown that MMP-9 plays an important role in the migration of mast progenitor cells to inflammatory tissue.22, 23 Therefore, the increased expression of MMP-9 in this study was somehow related to the migration of BMCs to the inflammatory liver.
Film in situ zymography clearly showed that these MMP-9–positive cells possessed high gelatinolytic activity compared with the liver treated with CCl4 alone. Thus, the BMCs that migrated acted in the degradation of liver fibrosis (fibrolysis).
According to our present data, increased expression of MMP-14 (MT1-MMP [membrane-type 1 matrix metalloproteinase]) will contribute to degrading interstitial collagens24 to gelatin that MMP-9 can degrade, resulting in the regression of fibrosis (fibrolysis).
Recently, Kollet et al.25 reported that the expression of MMP-9 was increased with the migration of human CD34+ progenitor cells in CCl4-treated NOD/SCID mice and that an inhibitor of MMP-9 reduced this migration. Thus, proteolytic activity seems to be necessary for the cell migration in addition to matrix degradation activity.
It seems to be very important how many cells can migrate into the damaged liver to degrade fibers, but a recent paper26 reported little evidence of bone marrow-derived hepatocytes in the CCl4-treated liver. However, the dose of CCl4 was only 4% (0.08 mL/kg) of our dose (0.5 mL/kg), and the number of mice used was too small (1 or 2). The reason they did not see the BMCs that migrated is most likely due to the cessation of CCl4 injection after BMC transplantation. Even in our experimental model,11 the cessation of CCl4 after BMC transplantation dramatically reduced the number of BMCs migrating into the damaged liver (I.S., unpublished data, 2003). Thus, the extent of continuing liver damage may limit BMC migration to the liver with matrix degradation activity.
Transplanted BMCs differentiated into albumin-producing hepatocytes with an increased serum albumin level, and the degradation of the extracellular matrix may presumably lead to improved liver function resulting in better survival of mice with BMCtransplantation compared to that of treated with CCl4 alone, although only 1 transplantation of BMCs was performed.
As shown by double fluorescence, our data may also indicate that transplanted BMCs seem to become stellate cells, in agreement with a recent report,27 although the number was very small in our experimental model. This result seems to be contradictory to our result of resolution of liver fibrosis by BMC transplantation because transdifferentiated stellate cells may produce collagens. Our preliminary results indicated reduced messenger RNA expression of type I procollagen, transforming growth factor-β1 (TGF-β1) and no change of hepatocyte growth factor messenger RNA expression in the liver 1 week after BMC transplantation compared with the liver treated with CCl4 alone (I. Sakaida, unpublished data, 2003). As shown in Fig. 7, migrated BMCs seemed to reduce the fine network pattern of activated stellate cells. Thus, transplanted BMCs may affect activated stellate cells by reducing their number—e.g., by leading them to apoptosis. However, further examinations are necessary to determine the exact relationship between BMCs and resident stellate cells.
Our recent data12 indicated that the subpopulation of BMCs, nonhematopoietic cells in bone marrow, separated using an anti-Liv8 antibody, will transdifferentiate into hepatocytes in the liver damaged by CCl4 induction. The present study clearly indicates that this subpopulation of BMCs is also responsible for the resolution of liver fibrosis (fibrolysis) induced by CCl4 treatment.
In conclusion, the present study introduces a new concept for the treatment of liver fibrosis.