Potential conflict of interest: Nothing to report.
Partial hepatectomy triggers hepatocyte proliferation, hepatic matrix remodeling, and hepatocyte apoptosis, all of which are important processes in the regenerating liver. Previous studies have shown an increase in the levels of matrix metalloproteinases gelatinase A (MMP-2) and gelatinase B (MMP-9) after partial hepatectomy. The goal of this study was to investigate the role of MMP-9 in liver regeneration after partial hepatectomy. A 70% hepatectomy or sham laparotomy was performed in wild-type or MMP-9–deficient (MMP-9−/−) mice. Hepatic regeneration was determined by liver weight/total body weight ratios and BrdU staining, which was used to a calculate mitotic index at several times postoperatively. Cytokine and growth factor expression was evaluated by Luminex™ bead–based ELISA and Western blots. Finally, the effect of MMP-9 on apoptosis was measured using TUNEL and caspase expression. The MMP-9−/− animals had a delayed hepatic regenerative response when compared with wild-type controls. The MMP-9–deficient animals expressed significantly less VEGF, HGF, and TNF-α between days 2 and 3 post-hepatectomy. Apoptosis, as measured by TUNEL staining and caspase expression, was decreased in the MMP-9−/−. In conclusion, MMP-9 plays an important role in liver regeneration after partial hepatectomy by affecting matrix remodeling, as well as cytokine, growth factor, and caspase expression. (HEPATOLOGY 2006;44:540–549.)
The liver is a highly unique organ in that it retains the ability to regenerate after injury despite the fact that hepatocytes normally do not actively divide.1 On toxic injury or partial hepatectomy, mature hepatocytes will divide until an organ of similar weight is formed. The liver is mainly composed of hepatocytes, with a smaller percentage of nonparenchymal cells such as Kuppfer and stellate cells.1–4 In the early stages of the hyperplastic response after partial hepatectomy, hepatic histology differs significantly from normal. Hepatocytes are grouped into nonvascularized clusters of 12 to 15 cells, and the amount of extracellular matrix reduced as a consequence of hepatocyte proliferation without concomitant extracellular matrix synthesis.2 Later in the regenerative process, hepatocyte proliferation decreases and stellate cells migrate into the clusters, concurrent with formation of new vascular branches, ultimately restoring normal hepatic histology.1–4
Extracellular matrix expression and degradation during liver regeneration is poorly understood, although the onset of extracellular matrix production after partial hepatectomy correlates with a peak in the expression of transforming growth factor-beta (TGF-β) messenger RNA, which stimulates collagen synthesis.5–7 A link clearly exists between extracellular matrix remodeling and hepatic regeneration. Matrix metalloproteinases (MMPs) are a family of zinc-containing neutral proteinases that are involved in matrix remodeling in both normal and pathophysiological processes.8–10 MMPs are synthesized and secreted, in most cases, as pro-enzymes that are then activated by proteinases. In the liver, several MMPs, including MMP-1 (collagenase-1), MMP-2 (gelatinase A), MMP-9 (gelatinase B), MMP-3 (stromelysin-1), and MMP-11 (stromelysins-3), are expressed in stellate cells or hepatocytes.6, 11 The goal of this study was to determine the role of MMP-9 in liver regeneration after partial hepatectomy.
MMP-9−/− mice to start a breeding colony were a generous gift from the laboratory of Robert Senior and J. Michael Shipley (Washington University, St. Louis, MO). The gelatinase B–deficient animals (MMP-9−/−) are bred on a 129SvEv background and are otherwise genetically identical to the wild-type 129SvEv (MMP-9+/+) animals.12 All experiments were performed in compliance with the standards for animal use and care set by the University of Michigan's Committee for the Use and Care of Animals. Six- to eight-week-old mice were used for the study. Anesthesia was induced with subcutaneous ketamine hydrochloride (100 mg/kg) and maintained with isoflurane inhalation. All animals received intraperitoneal lactated Ringer's solution (40 mL/kg) to replace operative fluid and blood losses. Partial (70%) hepatectomy was performed as previously described.13 Previous studies in our laboratory, using this animal model, have demonstrated that the liver will regain its appropriate weight, that is be equal to that of sham operated control animals of an equal age and body weight, approximately 7 to 9 days after 70% hepatectomy. Liver weight to total body weight ratios were performed at 2, 3, 5, 7, and 10 days postoperatively; bromodeoxyuridine (BrdU) and terminal deoxynucleotidyl transferase-mediated nick-end labeling (TUNEL) staining were performed 2, 3, and 5 days postoperatively. For the liver weight/total body weight studies, animals were weighed immediately before killing; after killing, the liver was excised in toto, immediately weighed, and liver weight to total body weight ratios calculated. For the histological studies, the liver was fixed in 10% neutral buffered formalin and embedded using an automated processor. Liver sections were cut between 3 and 5 μm (depending on study) and stained with hematoxylin-eosin. Trichrome staining was also performed to determine whether differences in matrix breakdown were present; no differences were seen between the MMP-9−/− mice and wild-type mice at any time points. Alternatively, sections were prepared for BrdU or TUNEL analysis. For the Western blot analysis of whole cell lysates, animals were killed in a kinetic fashion; liver samples were obtained and snap frozen in liquid nitrogen until ready for analysis.
In Vivo Measurement of Hepatocyte DNA Synthesis by Bromodeoxyuridine (BrdU) Incorporation.
Two hours before killing, animals were injected intraperitoneally with 30 μg BrdU/gram of body weight. Animals were then killed and liver specimens obtained and processed as described. Sectioned liver tissue was stained using the Amersham cell proliferation kit (Amersham Pharmacia Biotech Limited, United Kingdom). Three animals were used per treatment group per time point, and five separate high-power fields (40×) were analyzed. The number of BrdU-positive cells per high-power field were counted and expressed as the mean ± the standard error of the mean for each group.
Microscopic Detection of Apoptosis.
TUNEL assays were performed by the Histology & Immunoperoxidase Core facility at the University of Michigan, using an ApopTag Peroxidase Kit from Intergen (Purchase, NY). TUNEL assays were performed on paraffin embedded liver sections with counts expressed as the number of apoptotic nuclei per high-power (40×) field. A positive cell was defined as a cell that had the associated nuclear changes along with a TUNEL-positive stain. This decreased nonspecific staining because of rapid cell division (Intergren, Purchase, NY) or from necrosis.14 A minimum of five fields for each liver section were counted, with at least 500 cells visualized.
Liver Lysate Protocol.
Snap-frozen liver tissue was homogenized in T-PER (Pierce Biotechnology, Rockford, IL) supplemented with protease (Roche, Indianapolis, IN) and phosphatase inhibitors (Upstate, Charlottesville, VA). Tissue homogenate was centrifuged at 14,000g for 15 minutes at 4°C (Eppendorf, Hamburg, Germany), and the supernatant was used for the remaining studies. Protein concentration was determined using the bicinchoninic acid method (Pierce).
Cytokine Expression After Partial Hepatectomy.
Liver lysate (25 μg/well) was diluted at least twofold in assay buffer; cytokine expression was measured on a Luminex™ 100 using the Upstate mouse cytokine 10-plex (Upstate) kit following manufacturer's instructions. Liver lysate was analyzed using the “high biotin protocol” outlined in the kit with the addition of two extra washes at each wash step. All samples were run in duplicate. On completion of the protocol, data were collected with a minimum of 100 beads per analyte using StarStation 1.0 (Applied Cytometry, Sacramento, CA). The analyzed data were graphed and statistically analyzed using PrismGraph 4.0 (GraphPad Software, Inc., San Diego, CA).
Measurement of Growth Factors Using Luminex™ After Partial Hepatectomy.
Expression of vascular endothelial growth factor (VEGF), platelet-derived growth factor-BB, fibroblastic growth factor-basic, and granulocyte colony-stimulating factor was measured on the mouse growth factor kit (Biosource International, Camarillo, CA). Liver lysate (25 μg/well) was incubated along with the appropriate beads and detected according to manufacturer's instructions. Data were collected with a minimum of 100 beads per analyte, using StarStation 1.0 (Applied Cytometry). The analyzed data were graphed and statistically analyzed using PrismGraph 4.0 (GraphPad Software).
Measurement of Transforming Growth Factor-β1 Expression by ELISA.
Transforming growth factor beta 1 (TGF-β1) was measured using the commercially available ELISA kit (R&D Systems, Minneapolis, MN) as per the manufacture's instructions. Optical density and expression were measured using absorbance (SpectraMax 190 spectrophotometer, Molecular Devices, Sunnyvale, CA). The analyzed data were graphed and statistically analyzed using PrismGraph 4.0 (GraphPad Software).
Western Blot Analysis of Hepatocyte Growth Factor-α and Caspase 3 and 7.
Liver protein lysate was diluted following manufacture's instructions and was analyzed using the NuPage reducing gel system (Invitrogen, Carlsbad, CA). Liver lysate and molecular weight markers [mixture of SeeBlue2, Magic Mark (Invitrogen)] were loaded onto a 4% to 12 % bis/tris Criterion gel (Bio-Rad, Hercules, CA) and run at a constant voltage (200 V) for 35 minutes in NuPage MES SDS buffer (Invitrogen). The gel was blotted onto nitrocellulose (Bio-Rad) using the Trans-Blot SD (Bio-Rad) system with 2× NuPage transfer buffer (Invitrogen) under constant voltage (15 V) for 1 hour. Membranes were washed, blocked for 1 hour in 5 % w/v nonfat dry milk in Tris-buffered saline with 0.1 % Tween-20 (Sigma, St. Louis, MO), and incubated overnight with either cleaved caspase-3 (#9661), caspase-3 (#9662), caspase-7 (#9492), cleaved caspase-7 (#9491; Cell Signaling Technology, Beverly, MA) or hepatocyte growth factor alpha (HGF-α) (Assay Designs Inc. Ann Arbor, MI). In this study, we specifically measured the HGF-α chain, which is the higher molecular weight chain of the heterodimeric form of HGF, containing the four kringles. Membranes were then washed, incubated with goat-anti rabbit IgG horseradish peroxidase conjugated secondary antibody (Zymed, South San Francisco, CA; diluted 1:2500 in block), and washed again. The secondary antibody was detected using ECL+ (Amersham Biosciences Corp, Piscataway, NJ) and exposing to Hyperfilm ECL (Amersham). Blots were then stripped, blocked with non-fat dry milk, and glyceraldehye-3-phosphate dehydrogenase (GAPDH) was detected as above (antibody #ab8245; AbCam, Cambridge, MA). Developed films were scanned using a HP Scanjet 7400c scanner, and densitometry was performed using UnScanIt software (Silk Scientific, Orem, UT). Band density was normalized against GAPDH loading control, analyzed, and graphed using Prism 4.0 (GraphPad).
For the liver weight/total body weight studies and BrdU analysis, groups of data were evaluated by analysis of variance (ANOVA) by the methods of Student-Newman-Keuls to indicate groups with significant differences. Differences were considered significant if P < .05. Data were analyzed using a PowerPC 7100 computer using the Statview II statistical software package (Abacus Concepts, Inc.).
For measurement of cytokine, growth factors, TUNEL, and caspase assays, the data were analyzed using the one-way ANOVA (Graph Pad Prism 4.0, San Diego, CA). Differences among experimental groups were then compared using the Tukey post-test. Results were considered statistically significant when P < .05. Luminex ™ and ELISA techniques used at least three to four animals per group run in duplicate. For the Western blot analysis studies, each experiment was repeated twice on a minimum of three animals, and representative gels from pooled samples are illustrated.
Histological Differences After Partial Hepatectomy in MMP-9−/− Animals.
Liver histology in MMP-9+/+ and MMP-9−/− mice was analyzed using hematoxylin-eosin staining; liver sections from post-hepatectomy days 2, 3, and 5 are illustrated in Fig. 1. A striking difference between MMP-9+/+ and MMP-9−/− mice is seen on post-hepatectomy day 2, with MMP-9−/− mice showing increased hepatocyte swelling and vacuolization, as compared with MMP-9 +/+ animals. This resolved by day 3, and only slight histological differences between wild-type and the MMP-9–deficient animals are seen at this time point. By day 5, the MMP-9−/− animals show small regions of necrosis that are more prevalent than that seen in the MMP-9+/+ animals at this time point.
Liver Weight/Total Body Weight Ratios.
MMP-9+/+ and MMP-9−/− mice underwent partial hepatectomy or sham laparotomy, and liver weight/total body weight ratios were measured at 2, 3, 5, 7, and 10 days postoperatively. No differences were seen between MMP-9+/+ and MMP-9−/− mice undergoing sham laparotomy (data not shown). In contrast, as shown in Fig. 2, liver weight/total body weight ratios were decreased in the MMP-9–deficient mice at all time points as compared with wild-type animals; this reached statistical significance by post-hepatectomy days 5 and 7 (P < .05). By day 10, the MMP-9–deficient mice had a return of liver weights to near those of the wild-type mice.
BrdU Staining After Partial Hepatectomy.
Hepatocyte proliferation, as determined by BrdU staining, 2, 3, and 5 days post-hepatectomy, was generally slower in MMP-9−/− mice as compared with MMP-9+/+ mice. BrdU staining is most notable on day 2; at this time point, MMP-9+/+ animals have significantly more BrdU-positive cells than MMP-9−/− mice (Fig. 3A-C). Hepatocyte proliferation was most rapid in the wild-type mice at day 2 post-hepatectomy; in contrast, MMP-9−/− mice had a slower and somewhat more prolonged period of hepatocyte proliferation.
Cytokine Expression After Partial Hepatectomy.
Most cytokines measured in this study showed no statistical differences between MMP-9+/+ and MMP-9−/− animals (Fig. 4). A notable exception to this is tumor necrosis factor alpha (TNF-α); on day 2, MMP-9−/− animals expressed significantly less TNF-α as compared with MMP-9+/+ mice. Although TNF levels appear to be increased in the MMP-9−/− animals on postoperative day 1, this did not reach statistical significance; the significance of this finding is therefore not clear.
Growth Factor Expression After Partial Hepatectomy.
No differences were seen in hepatic levels of TGF-β1, platelet-derived growth factor-BB, fibroblast growth factor-basic, or granulocyte colony-stimulating factor at any point in this study. However, MMP-9−/− animals showed less VEGF at day 3 when compared with the MMP-9+/+ mice (Fig. 5A, P < .05).
HGF-α expression was detected in all animals throughout the time course (Fig. 6A). Quantification of the individual animals showed a statistically significant decrease in HGF-α at days 2 and 3 in the MMP-9−/− animals, as compared with wild-type mice (Fig. 6B, P < .01).
TUNEL Staining After Partial Hepatectomy.
Quantification and statistical analysis of the TUNEL staining showed no differences between the MMP-9+/+ and MMP-9−/− animals at day 2, but significantly fewer apoptotic cells in MMP-9−/− animals as compared with the MMP-9+/+ animals by days 3 and 5 (Fig. 7G, P < .01).
Caspase Expression After Partial Hepatectomy.
Caspase 3 and 7 expression and activation (cleavage) was analyzed by Western blot. Pro-caspase 3 expression remains relatively stable throughout the 5 days after partial hepatectomy (Fig. 8A,C), whereas pro-caspase 7 expression seems to decrease in the MMP-9−/− mice versus the MMP-9+/+ mice by day 5 (Fig. 8B,D). Caspase 3 cleavage occurred earlier and to a greater extent in wild-type animals at days 3 and 5, as compared with the MMP-9−/− mice (Fig. 8E). Similarly, cleaved caspase 7 expression was also greater in wild-type animals as compared with the MMP-9−/− mice throughout the time course, with the greatest difference occurring at day 3 (Fig. 8F, P < .01).
When compared with wild-type mice, mice deficient in MMP-9 undergoing 70% hepatectomy showed a delayed regenerative response based on liver weight gain and BrdU staining. The MMP-9–deficient animals also had a decreased apoptotic response as measured by TUNEL staining and caspase activation. These effects may be related to altered matrix remodeling in the MMP-9–deficient animals. The lack of MMP-9 may result in significant aberrations in matrix remodeling or affect the levels or activation of a variety of inflammatory mediators and growth factors.
In adult rats, the effects of HGF on DNA synthesis in the liver in vivo are enhanced by pretreatment of the liver with collagenase, suggesting that matrix remodeling in the early stages of hepatic regeneration may play a role in hepatocyte priming.15 Similarly, pro-MMP-2 and pro-MMP-9 are activated after partial hepatectomy and may contribute to priming hepatocytes for proliferation by modulation of the matrix environment in the remnant liver.16 Prior investigations by Kim and colleagues demonstrate that MMP-2 and MMP-9 are expressed after partial hepatectomy6, 7, 17 and suggest that hepatic matrix remodeling is an important contributor to the process of hepatic regeneration. Rudolf and colleagues have also illustrated that differential regulation of extracellular matrix synthesis in the context of partial hepatectomy helps to control hepatocyte transition from the G0 phase into the active replicative phase and then back into the quiescent phase.7 Additionally, extracellular matrix degradation is critical for G1/S transition and Cdk1 induction.18 The current studies further support this hypothesis, illustrating decreases in hepatic regeneration and hepatocyte proliferation in MMP-9–deficient mice as compared with wild-type controls, suggesting an important supporting role for MMP-9–mediated matrix remodeling in the context of hepatic regrowth after injury.
In many pathological conditions, MMPs such as MMP-9 are upregulated and are believed to be important not only in matrix remodeling but also in modulation of cell activation to alter cytokine levels as well as in the modulation of other proteinase activities.9 Several recent studies by Khokha and colleagues have suggested a very important role for matrix remodeling in the context of partial hepatectomy.19–21 In general, MMPs and inhibitors of MMPs are important for extracellular matrix turnover and remodeling, as well as processing and activation of growth factors and cytokines.19 More specifically, these studies have shown that TIMP-1 is an important negative regulator of hepatocyte growth factor (HGF) activity in the context of partial hepatectomy.20 Furthermore, the lack of TIMP-3 in mice in the setting of partial hepatectomy results in TNF dysregulation, with persistently elevated TNF levels; this results in not only failure of liver regeneration, but also actual liver failure, which resembles chronic active hepatitis in humans.21 Other studies also document that MMPs are important in the upregulation or activation of a variety of cytokines and chemokines; this is an important regulatory function that modulates inflammation and the host immune response. MMP-2 and MMP-9 both have been shown to be important for activation of several chemokines, including interleukin-8, monocyte chemotactic protein-3, monokine-induced interferon-gamma, interferon-inducible protein-10, and epithelial neutrophil activating protein; until these molecules are activated, their pro-inflammatory effects do not occur.22, 23
In general, cytokine expression was not affected by the lack of MMP-9. However, a notable exception to this statement is the significant decrease in TNF-α levels that were noted in the MMP-9−/− mice on postoperative day 2 after hepatectomy; this trend continued on postoperative day 3, although these differences did not reach statistical significance. Prior studies have shown that TNF-α is upregulated after partial hepatectomy, and TNF-α inhibition in this setting significantly decreases the rate of hepatic regeneration, possibly by inhibition of Jun kinase.18, 24–26 This may be an important mechanism related to the delay in hepatocyte proliferation that is seen in this study. In contrast, MMP-9−/− mice had increased levels of TNF on postoperative day 1 after hepatectomy; because the difference between MMP-9−/− mice and wild-type mice on postoperative day 1 did not reach statistical significance, the importance of this finding is not clear. Interestingly, no statistical difference was seen in hepatic IL-6 expression, indicating that extracellular matrix remodeling and IL-6 may be independently regulated.27–30 Outside of TNF-α no major differences were seen in cytokine expression, indicating that differences seen between the wild-type and knockout animals may be attributable to growth factor expression.4, 31–33
MMP-9 has been shown to play an important role in angiogenesis and apoptosis in the cartilaginous growth plates of mice.34 In our model, the MMP-9−/− animals had significantly less VEGF than wild-types at day 3, indicating that MMP-9 may affect liver angiogenesis. VEGF mRNA expression peaks at approximately 48 hours after hepatectomy; however, plasma level and protein expression remain relatively stable.35 Liver expression of VEGF protein peaked at day 3 but remained slightly elevated by day 5 in the wild-type animals. This indicates that MMP-9 is involved in either the release or increased expression of VEGF after partial hepatectomy. Liver regeneration is dependent on the formation of proper vasculature via angiogenesis.2, 36, 37 Additionally, MMP-9 has been implicated in angiogenesis and VEGF expression in osteoclast recruitment and carcinogenesis.38, 39 Therefore, one possible mechanism for decreased hepatic regeneration in MMP-9−/− animals is attributable to diminished or altered vasculature formation kinetics.
Another growth factor implicated in liver regeneration and considered a target for MMP-9 is TGF-β1. TGF-β1 is present in the extracellular matrix and rapidly increases in the plasma after partial hepatectomy.4 TGF-β and its family members have been implicated in liver apoptosis and in liver regeneration.5, 40–44 Our data show a slight increase in TGF-β in the MMP-9−/− animals throughout the time studied; however, the lack of statistical differences despite changes in morphology and regeneration rate indicate that MMP-9 does not affect TGF-β1 and may depend on additional factors or cytokines.45
HGF is a critically important mediator during liver regeneration; it is expressed as a single-chain unprocessed protein that can bind the extracellular matrix and is processed by urokinase plasminogen activator or HGF activator into α and β chains.2 One study has shown that HGF levels actually decrease via consumption by the remnant liver; in the next phase, HGF levels increase significantly through hepatic HGF production.46 Because HGF is expressed as a precursor protein, the use of an HGF-α–specific antibody measures both activation and expression. MMP-9 is involved in the upregulation of HGF after partial hepatectomy. HGF-α expression was statistically greater in the day 2 and day 3 wild-type animals, indicating that MMP-9 may affect expression or activation. The increased HGF expression that is seen at days 2 and 3 may cause VEGF induction on day 3, reinforcing the link between these two growth factors.47 Although HGF has been shown to increase the expression of Bcl-xl,48 an anti-apoptotic protein, levels of activated caspace 3 and 7 were decreased in the MMP-9−/− mice, and the overall rates of apoptosis were also decreased in these animals. If HGF levels are decreased, one would anticipate decreased levels of the anti-apoptotic protein, Bcl-xl, and therefore increased levels of apoptosis; this is in contrast to what was actually observed in this study. Likely the overall decreased apoptotic rate in the MMP-9−/− mice after partial hepatectomy correlates with the slower rate of hepatocyte proliferation. This emphasizes the complexity of the control mechanisms involved in this process: despite the fact that there are likely decreased levels of the anti-apoptotic protein, Bcl-xl, overall rates of hepatocyte apoptosis are decreased. This is probably controlled via other mechanisms functioning independently of Bcl-xl. The link between MMP-9 and apoptosis has been shown only in the cardiac cells of diabetic mice and for neuronal growth.49, 50 In these studies, treatment of IL-6−/− animals with MMP-2/9 inhibitors also resulted in a decrease in apoptosis via TUNEL staining.17 More specific studies investigating the mechanisms involved in this model are not yet available.
In summary, the goal of this study was to investigate the role of MMP-9 in liver regeneration after partial hepatectomy. The overall effect of MMP-9 during liver regeneration is summarized in Supplementary Fig. 1 (Supplementary material is available at the HEPATOLOGY website: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). Compared with wild-type mice, MMP-9 knockout mice showed delayed regenerative response based on liver weight gain and BrdU staining. Expression of TNF, HGF, and VEGF were decreased significantly; however, no differences were seen in TGF-β1 in the MMP-9−/− animals. Additionally, the MMP-9−/− animals had a decreased apoptotic response as measured by TUNEL and caspase activation. These data suggest that MMP-9 is important for liver regeneration after partial hepatectomy, and its mechanisms are likely through the upregulation of TNF, HGF, VEGF, and altered regulation of both proliferative and apoptotic pathways.