SEARCH

SEARCH BY CITATION

Keywords:

  • liver repair;
  • macrophage accumulation;
  • monocyte chemoattractant protein-1;
  • neutrophil recruitment;
  • plasminogen

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

Summary. Background: The involvement of plasminogen in liver repair has been reported, but its exact role in promoting this process is unknown. Objective: To elucidate the underlying mechanism, we examined the dynamics of liver repair by using a reproducible liver injury model in plasminogen gene-deficient mice and their wild-type littermates. Methods: Liver injury was induced by photochemical reaction and the subsequent responses were histologically analyzed. Results: In wild-type animals, the area of the damage successively decreased, and the repair process was associated with macrophage accumulation at its border. Neutrophils were also attracted to the damaged region on day 1 and were evident only at its border by day 4, which spatially and temporally coincided with the expression of macrophage chemoattractant protein-1 (MCP-1). Neutrophil depletion suppressed recruitment of macrophages at the border between the damaged and the normal tissues. These changes were followed by activated hepatic stellate cell accumulation, collagen fiber deposition and angiogenesis at the boundaries of the injured zone. In contrast, in plasminogen gene-deficient mice, the decrease in the area of damage, macrophage accumulation, late-phase neutrophil recruitment, hepatic stellate cell accumulation, collagen fiber deposition and angiogenesis were all impaired. Conclusion: Our data suggest that accumulated neutrophils at the border of the damaged area may contribute to macrophage accumulation at granulation tissue via the production of MCP-1 after liver injury. The plasminogen system is critical for liver repair by facilitating macrophage accumulation and triggering a cascade of subsequent repair events.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

The wound healing response is triggered in various organs, including the liver, after an injury [1]. The healing process can generally be divided into three phases: acute inflammation, granulation tissue formation, and tissue remodeling [2]. In acute inflammation, neutrophils are rapidly recruited to the damaged area in response to numerous mediators released from the damaged tissue, whose levels decrease gradually, coinciding with the increase in macrophage infiltration [2]. Granulation tissue consists of various types of cells such as macrophages and the fiber extracellular matrix (ECM)-producing cells, including fibroblasts and myofibroblasts, supported by newly generated blood vessels [2]. Macrophages are attracted by a variety of chemokines, including monocyte chemoattractant protein-1 (MCP-1) [3]. Activated macrophages release growth factors, which further stimulate the production of ECM and angiogenesis in granulation tissue [2]. The deposited ECM proteins act as a scaffold for angiogenesis [2]. In injured liver, resident hepatic macrophages, known as Kupffer cells, induce the transdifferentiation of quiescent hepatic stellate cells (HSCs) into myofibroblasts, expressing α-smooth muscle actin (α-SMA) [1] and stimulating subsequent ECM production [1]. Disruption of this chain of events leading to liver repair causes chronic liver diseases, including liver fibrosis and cirrhosis [1].

Plasminogen is an inactive proenzyme that is converted to active serine protease plasmin by plasminogen activators, including tissue-type plasminogen activator as well as urokinase [4]. The plasminogen activator-plasmin system has various physiological or pathophysiological roles in thrombolysis, cerebral ischemia, fibrosis, inflammation and wound healing [4–6]. Plasminogen displays high affinity for lysine residues through its kringle domains, which are physiologically relevant for binding to cells and subsequent pericellular proteolysis [4]. Previous studies have shown that plasminogen also plays an important role in liver repair through the proteolysis of ECM and clearance of cellular debris [7–9]. Further, plasminogen suppresses hepatocyte apoptosis [10] and induces hepatocyte proliferation [11]. Moreover, its mode of action in other tissues suggests that plasminogen may affect recovery by acting beyond its proteolytic effect on the ECM.

Various animal models have been established for studying the subsequent responses after liver injury [7–9,12,13]. However, these models allow only semi-quantitative estimation of the response, mostly due to the difficulties in inducing reproducible damage. Therefore, in this study, we established a liver injury model by using a photochemical method, which has been widely used to cause damage in various organs [14–16]. In this method, a combination of intravenously infused Rose Bengal, a photosensitive dye, and local photoillumination produces reactive oxygen at irradiated site. Consequently, the reactive oxygen causes a limited ischemic injury. By using this model, we evaluated quantitatively the wound healing reactions in the liver, including the size of the damaged area, macrophage accumulation, neutrophil recruitment, HSC accumulation accompanied by collagen fiber deposition, and angiogenesis in mice deficient for the plasminogen gene (Plg−/−) compared with their wild-type counterparts (Plg+/+). In addition, fibrin(ogen) as well as MCP-1 were also assessed at the damaged area.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

Animals

Plg−/− mice and their Plg+/+ littermates (controls), each weighing 18−25 g, with mixed backgrounds of C57BL/6J (75%) and 129/SvJ (25%) were used [17]. All experiments were performed in compliance with the guidelines of the International Society for Thrombosis and Haemostasis, and the current institutional rules for the use and care of laboratory animals.

Animal model

Liver injury was induced photochemically by using the method reported by Umemura et al. [16] with some modifications. In brief, the left lateral liver lobe was exposed after laparotomy under isoflurane anaesthesia, and an optic fiber (3-mm in diameter) was allowed to gently touch the center of the lobe. The right jugular vein was cannulated with a polyethylene tube containing 0.9% saline and connected to a microinjection pump. Rose Bengal (10 mg kg−1) (Wako Pure Chem, Osaka, Japan) dissolved in 0.9% saline was infused at a rate of 50 μL per min and was immediately followed by photoillumination with green light (wavelength, 540 nm; illumination intensity, 2.36 × 10−2 W cm−2) for 10 min by using a heat-absorbing light source (model L5178, Hamamatsu Photonics, Hamamatsu, Japan). The incised muscle and skin were then sterilely sutured, and the anesthesia was discontinued. The body temperature was maintained at 37 °C during surgery by using a heating pad.

Analysis of photochemically induced liver injury

The mice were anesthetized with pentobarbital (50 mg kg−1, intraperitoneally) on days 1, 4, 7 and 14 after the surgery. Then, the injured lobe was exposed, its surface was photographed, and the area of damage was quantified by planimetry. After collecting blood samples, the liver was removed and embedded in paraffin following fixation with 4% paraformaldehyde. Then, 4-μm sections were cut vertically starting from the center of the damaged area. After staining with hematoxylin and eosin (H&E), the sections were photographed and the area of damage was again quantified by planimetry. To determine the degree of hepatocyte injury, the plasma level of alanine aminotransferase (ALT) was measured with a colorimetric assay kit (Wako Pure Chem., Osaka, Japan).

Histological analysis

Immunostaining for F4/80 (a macrophage marker), fibrinogen, neutrophil-specific marker, MCP-1, α-SMA (an activated HSC marker), 5-bromo-2-deoxyuridine (BrdU, a thymidine analog) and CD31 (an endothelial cell marker) was performed as described elsewhere [18]. The 4-μm sections were incubated with rat monoclonal anti-F4/80 antibody (AbD Serotec, Raleigh, NC, USA), rabbit polyclonal anti-α-SMA antibody (Novus Biologicals, Inc., Littleton, CO, USA), goat polyclonal anti-mouse fibrinogen antibody (Nordic Immunological Laboratories, Tilburg, the Netherlands), rabbit polyclonal anti-MCP-1 antibody (Abcam, Cambridge, UK), rat monoclonal anti-neutrophil antibody (NIMP-R14; Abcam), or mouse monoclonal anti-BrdU (Chemicon International, Temecula, CA, USA) and rabbit polyclonal anti-CD31 (Abcam). The sections were then incubated with the appropriate secondary antibodies conjugated with horse-radish peroxidase. Positive signals were visualized using a tyramide signal amplification system (PerkinElmer, Waltham, MS, USA). The sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and photographed under a fluorescence microscope (E800; Canon, Tokyo, Japan) with the CCD camera. The fibrinogen immunoreactivity was visualized by diaminobenzidine coloration and the sections were counterstained with hematoxylin. Those double-stained for MCP-1 with either antigens for neutrophil-specific marker or F4/80 were examined under laser scanning confocal microscopy (LSM 510 META; Carl Zeiss, Oberkochen, Germany).

The broadness of the condensed F4/80 immunoreactivity at the border of the damaged area was measured by an image process program (Mac SCOPE; Mitani Co., Fukui, Japan) in a blinded evaluation.

The borders of the damaged area were photographed, and the number of neutrophils was measured as the mean of the number of immunoreactive cells for neutrophil-specific marker in five individual microscopic fields at the borders of the damaged area by a blinded assessor.

To evaluate the accumulation of ECM proteins, the sections were stained by using the Masson trichrome method.

Other experimental methods

Administration or inhibition of plasminogen, fibrinogen depletion or neutrophil depletion and other procedures were performed as described in the Supporting Information.

Statistical analysis

All data are expressed as the mean ± SEM. Statistical significance was evaluated by Student’s t-test and was set at the < 0.05 level.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

Size of liver injury in the photochemical model

After photoillumination, the damaged area was observed as a circular pale zone on the liver surface (Fig. 1A) and reached half of the thickness of the liver lobe on day 1 (Fig. 1C). The area of damage was quantified and is visualized in Fig. 1(B,D). As shown by the small error bars in each graph, the liver injury in the model was highly reproducible. In the mice without Rose Bengal infusion, little area of damage was observed near the surface in contact with the optic fiber (Supplementary Fig. S1). The plasma ALT levels, however, were not increased significantly on day 1 in these mice (without Rose Bengal, 17 ± 0.95 IU L−1; intact, 22 ± 2.7 IU L−1). Therefore, this model was considered suitable for quantitative evaluation of the subsequent responses in the damaged area after liver injury.

image

Figure 1.  Impairment of the recovery of the damaged area after photochemically induced liver injury in Plg−/− mice. (A) Photographs of the liver surface at the damaged area after photoillumination in Plg+/+ mice (upper) and Plg−/− mice (lower). (B) Quantification of the damaged area as shown in (A). The pale surface area was measured by NIH Image. (C) H&E-stained vertical sections collected from the center of the damaged area in Plg+/+ mice (upper) and Plg−/− mice (lower). Scale bars indicate 1 mm. (D) Quantification of the damaged area in the H&E-stained sections. (E) The plasma ALT levels in the Plg+/+ and Plg−/− mice after photoillumination. The data represent the mean ± SEM of three to four mice at each time point. *< 0.05 and **< 0.01 vs. Plg+/+ mice.

Download figure to PowerPoint

Impairment of the recovery of the damaged area after photochemically induced liver injury in Plg−/− mice

The area of damage on the liver surface gradually decreased until 56 days in Plg+/+ mice. On the other hand, this area did not change until 14 days and decreased slightly on days 28 and 56 in Plg−/− mice (Fig. 1A). Thus, the area of damage in Plg−/− mice was significantly larger than that in Plg+/+ mice on days 4–56 after photoillumination (Fig. 1A,B). The same tendency was observed when the damage was quantitatively estimated on the vertical section dissected through the center of the affected region until 14 days in both groups of mice (Fig. 1C,D). On days 28 and 56, Plg−/− mice also showed a decrease in the area of damage, which was more remarkable on the surface (Fig. 1B) than vertically (Fig. 1D). Therefore, the difference in the decrease on the surface was not attributable to a decrease in the area of damage but to the covering of the surface by living cells. No significant difference was observed between the mice with regard to the transient increase in the plasma ALT on day 1 (Fig. 1E), suggesting that the damage was induced temporally until 1 day after photoillumination. These results indicate that plasminogen played a crucial role in the recovery of the damaged tissue after photochemically induced liver injury.

Involvement of plasminogen in the accumulation of F4/80-immunoreactive cells at the border of the damaged area

The H&E-stained sections indicated that granular cells were abundant in the region between the normal and the damaged area in Plg+/+ mice on day 4 after photoillumination (Fig. 2A, left). Therefore, this layer was defined as the granular cell-rich layer. F4/80-immunoreactive cells accumulated in this layer (Fig. 2B, left). In contrast, no such layer was observed in Plg−/− mice on day 4 after photoillumination (Fig. 2A, right), which was consistent with the lower accumulation of F4/80-immunoreactive cells at the border of the damaged area (Fig. 2B, right). As the F4/80-immunoreactive cells were observed inside the damaged tissue in Plg−/− mice, the borderline between the damaged area and the F4/80-immunoreactive cell zone was not clearly visible (Fig. 2B, right). In Plg+/+ mice, the area of F4/80-immunoreactive cells at the border of the damaged tissue decreased gradually from 4 to 14 days after photoillumination and was accompanied by decreases in the area of damage (Fig. 2C) and the granular cell-rich layer (data not shown). In contrast, in Plg−/− mice, the accumulation of F4/80-immunoreactive cells was dramatically attenuated (Fig. 2D) and the extent of tissue damage did not change over time; even on day 56, macrophage accumulation was not observed in these mice (Figs 2C and S2).

image

Figure 2.  Involvement of plasminogen in the accumulation of F4/80-immunoreactive cells at the border of the damaged area. (A) Microphotographs with higher magnification of the H&E-stained vertical sections from the center of the damaged area in Plg+/+ mice (left) and Plg−/− mice (right) on day 4 after photoillumination. Scale bars indicate 100 μm. (B) Microphotographs of F4/80-immunoreactive cells in Plg+/+ mice (left) and Plg−/− mice (right) on day 4 in the adjacent sections of those shown in (A), respectively. The area between the solid lines indicates the granular cell-rich layer in the Plg+/+ mice (B, left). The dotted line indicates the border between the normal and the damaged areas in Plg−/− mice (B, right). Scale bars indicate 100 μm. (C) Microphotographs of F4/80-immunoreactive cells in the vertical section of the damaged area in Plg+/+ or Plg−/− mice. Scale bars indicate 1 mm. (D-F) Quantification of the area of condensed F4/80 immunoreactivity at the border of the damaged area in Plg+/+ and Plg−/− mice (D), in the Plg−/− mice treated with the vehicle or Glu-plamsinogen (Glu-Plg) (E), and in the Plg+/+ mice treated with the vehicle or tranexamic acid (F) on day 4 after photoillumination. The data represent the mean ± SEM of four (D), seven (E) or four (F) mice in each group.

Download figure to PowerPoint

Daily intravenous administration of 10 mg kg−1 Glu-plasminogen significantly increased the area of F4/80-immunoreactive cells in Plg−/− mice on day 4 after photoillumination (Fig. 2E). Therefore, the reduction of macrophage accumulation in these mice was not associated with developmental abnormalities but rather with the acute need for plasminogen to promote this effect. Furthermore, in Plg+/+ mice treated with tranexamic acid, the area of F4/80-immunoreactive cells significantly decreased on day 4 after photoillumination (Fig. 2F), indicating the importance of the kringle domains of plasminogen in the accumulation of F4/80-immunoreactive cells. Altogether, these results demonstrate the critical role of plasminogen in the accumulation of F4/80-immunoreactive cells after liver injury.

Effect of fibrinogen depletion on the accumulation of F4/80-immunoreactive cells

In order to investigate the role of fibrinogen and/or fibrin [fibrin(ogen)] on the accumulation of F4/80-immunoreactive cells at the border of the damaged area, the immunostaining with anti-mouse fibrinogen antibody was performed. As shown in Supplementary Fig. S3, fibrin(ogen) was detected at the damaged area, especially at the border in both genotypes on day 1. Although the deposited fibrin(ogen) decreased together with the decrease in the damaged area until day 14 in Plg+/+ mice, it was observed in Plg−/− mice even at day 14. However, the accumulation of F4/80-immunoreactive cells at the border of the damaged area was comparable between the fibrinogen-depleted Plg+/+ mice and the control group on day 4. That is, the immunoreactive area quantified by Mac SCOPE was 0.86 ± 0.13 mm2 in the fibrinogen-depleted Plg+/+ mice and 0.60 ± 0.07 mm2 in the control group. In the fibrinogen-depleted Plg−/− mice, the granular cell-rich layer was not observed at the border of the damaged area (data not shown). In addition, the accumulation of F4/80-immunoreactive cells was comparable between the fibrinogen-depleted Plg−/− mice and the control group on day 4; their areas were 0.30 ± 0.04 mm2 and 0.32 ± 0.02 mm2, respectively. These results suggest that the accumulation of F4/80-immunoreactive cells might be regulated by the fibrin(ogen)-independent mechanisms.

Impairment of MCP-1-expressing neutrophil accumulation at the border of the damaged area in Plg−/− mice

In the damaged area, numerous neutrophils were observed on day 1 (Fig. 3A,B), but most of them disappeared on day 4 in both genotypes (Fig. 3C,D). Instead, neutrophils were observed mostly in the granular cell-rich layer in Plg+/+ mice (Fig. 3C) and were scarce at the border of the damaged area in Plg−/− mice (Fig. 3D). Although the number of recruited neutrophils was comparable between the genotypes on day 1 (Fig. 3E), it was significantly low in Plg−/− mice on day 4 as compared with Plg+/+ mice (Fig. 3F). The number of recruited neutrophils on day 4 was significantly increased in Plg−/− mice with exogenous Glu-plasminogen treatment (Fig. 3G) and decreased in Plg+/+ mice with tranexamic acid treatment (Fig. 3H). These results indicate that plasminogen participates in neutrophil accumulation in the damaged area on day 4, but not on day 1, after photoillumination.

image

Figure 3.  Impairment of the neutrophil accumulation on day 4 in Plg−/− mice. (A–D) Microphotographs of immunoreactive neutrophils (red) at the border of the damaged area in Plg+/+ mice (A, C) and Plg−/− mice (B, D) on day 1 (A, B) and day 4 (C, D) after photoillumination. The sections were counterstained with DAPI (blue). The dotted lines indicate the borders between the normal and the damaged areas (A, B, D). The area between the solid lines indicates the granular cell-rich layer in Plg+/+ mice (C). Scale bars indicate 50 μm. Similar results were obtained from four mice in each group. (E–H) Quantification of the number of neutrophils at the border of the damaged area in Plg+/+ and Plg−/− mice on day 1 (E) and day 4 (F), in Plg−/− mice treated with the vehicle or Glu-plamsinogen (Glu-Plg) on day 4 (G), and in Plg+/+ mice treated with the vehicle or tranexamic acid on day 4 (H). The data represent the mean ± SEM of four (E, F), seven (G) or four (H) mice in each group.

Download figure to PowerPoint

We next investigated the tissue expression of MCP-1 in our model. MCP-1 immunoreactivity was co-localized with neutrophils in the granular cell-rich layer in Plg+/+ mice on day 4, but not Plg−/− mice, where neither MCP-1 nor neutrophils were present (Fig. 4A,B). Interestingly, the MCP-1-positive neutrophils were abundant near the edge of the damaged area within the granular cell-rich layer (Fig. 4A). In contrast, MCP-1 immunoreactivity did not co-localize with F4/80 immunoreactivity (Fig. 4C,D). MCP-1 immunoreactivity was also detected in normal hepatocytes surrounding the damage in both genotypes (Fig. 4A–D). We also found that the expression level of MCP-1 at the damaged area was significantly lower in Plg−/− mice than in Plg+/+ mice on day 4 after photoillumination (Fig. 4E), which was consistent with the lower number of MCP-1-positive neutrophils in Plg−/− mice. To evaluate the role of MCP-1 in the accumulation of F4/80-immunoreactive cells in the granular cell-rich layer, MCP-1-neutralizing antibody was administered to Plg+/+ mice. This administration significantly decreased in the area of F4/80-immunoreactive cells in the granular cell-rich layer on day 4 after photoillumination (Fig. 4F). These results suggest that neutrophil-derived MCP-1, at least in part, contributed to the accumulation of F4/80-immunoreactive cells in the granular cell-rich layer.

image

Figure 4.  Expression of MCP-1 in the recruited neutrophils at the granular cell-rich layer. (A–D) Double staining for MCP-1 (red) and neutrophil antigen (green; A, B) or F4/80 (green; C, D) at the border of the damaged area in Plg+/+ mice (A, C) and Plg−/− mice (B, D) on day 4 after photoillumination. The arrowheads indicate the co-localization of neutrophils and MCP-1 (A). The inset shows the enlarged image of the small square in (A). The area between the solid lines indicates the granular cell-rich layer in Plg+/+ mice (A, C). The dotted line indicates the border between the normal and the damaged areas in Plg−/− mice (B, D). Scale bars indicate 50 μm. (E) The amount of MCP-1 at the damaged area was determined by ELISA before, and on days 1 and 4 after photoillumination. The data represent the mean ± SEM of five mice in each group. (F) Quantification of the area of condensed F4/80 immunoreactivity at the border of the damaged area in Plg+/+ mice on day 4 after photoillumination. Control IgG (Cont) or MCP-1-neutralizing antibody was daily injected for 4 days immediately after photoillumination. The data represent the mean ± SEM of five mice in each group.

Download figure to PowerPoint

Effect of neutrophil depletion on the macrophage recruitment to the damaged area and the decrease in the size of the damaged area in Plg+/+ mice

We examined macrophage accumulation and the decrease in the size of the damaged area after neutrophil depletion by the administration of neutralizing antibody in Plg+/+ mice as described in the Supporting Information. On day 7, granular cells were observed extensively not only at the border of the damaged area but also in the normal area in the neutrophil-depleted Plg+/+ mice (Fig. 5A,B). The distribution of the granular cells was extensively in line with that of the F4/80-immunoreactive cells on day 7 in these mice (Fig. 5C,D). The damaged area in the neutrophil-depleted Plg+/+ mice was significantly larger than that in the control group on day 7 (Fig. 5E), although it was comparable between the neutrophil-depleted Plg+/+ mice (4.0 ± 0.29 mm2) and the control group (4.1 ± 0.09 mm2) on day 1. These results suggest that the accumulated neutrophils in the granular cell-rich layer on day 4 contributed to recruitment of macrophages to the edge of the damaged area accompanied by a decrease in the size of the damaged area.

image

Figure 5.  Effect of neutrophil depletion on the recruitment of macrophages to the border of the damaged area and the decrease in the damaged area in Plg+/+ mice. (A, B) Microphotographs of H&E-stained vertical sections from the center of the damaged area on day 7 in the Plg+/+ mice treated with control IgG (A) or anti-mouse Ly-6G/Ly-6C (B). (C, D) Microphotographs of F4/80-immunoreactive cells on day 7 in the Plg+/+ mice treated with control IgG (C) or anti-mouse Ly-6G/Ly-6C (D) in the adjacent sections of those in (A) or (B), respectively. Scale bars indicate 200 μm. (E) Quantification of the area of damage in the Plg+/+ mice treated with control IgG or anti-mouse Ly-6G/Ly-6C antibody on day 7. The data represent the mean ± SEM of six mice in each group.

Download figure to PowerPoint

Impairment of activated HSC accumulation and collagen fiber deposition at the border of the damaged area in Plg−/− mice

On day 4, immunoreactive cells against α-SMA were observed in the damaged area and granular cell-rich layer in Plg+/+ mice (Fig. 6A), but in the damaged area and surrounding normal area in Plg−/− mice (Fig. 6B). ECM proteins, stained blue by Masson trichrome, were deposited in the granular cell-rich layer in Plg+/+ mice from 4 days to 14 days after photoillumination (Fig. 6C, upper). These proteins decreased from 7 days to 14 days after photoillumination (Fig. 6C, upper and D), which was associated with a decrease in the area of F4/80-immunoreactive cells in the granular cell-rich layer in Plg+/+ mice (Fig. 2C, upper). In contrast, ECM proteins were not deposited at the border of the damaged area in Plg−/− mice even on day 14 (Fig. 6C, lower and 6E). To examine collagen deposition in the damaged area, we quantified the hydroxyproline contents before and on day 7 after photoillumination in Plg+/+ and Plg−/− mice. The hydroxyproline content was significantly increased on day 7 compared with the one before experiment in Plg+/+ mice, although it was comparable in Plg−/− mice (Fig. 6F). These results suggest that activated HSCs were accumulated in the granular cell-rich layer after liver injury, accompanied by collagen fiber deposition. Furthermore, plasminogen contributed to activated HSC accumulation.

image

Figure 6.  Impairment of activated HSC accumulation and collagen fiber deposition at the border of the damaged area in Plg−/− mice. (A, B) Microphotographs of α-SMA immunoreactivity (red) at the border of the damaged area in Plg+/+ mice (A) and Plg−/− mice (B) on day 4 after photoillumination. The sections were counterstained with DAPI (blue). The area between the solid lines indicates the granular cell-rich layer in Plg+/+ mice (A). The dotted line indicates the border between the normal and the damaged areas in Plg−/− mice (B). Scale bars indicate 50 μm. Similar results were obtained from four mice in each group. (C–E) ECM proteins stained blue by Masson trichrome in Plg+/+ and Plg−/− mice. (D, E) Enlarged photographs of the squares in (C) for Plg+/+ and Plg−/− mice, respectively. Scale bars indicate 100 μm (C) and 25 μm (D, E). Similar results were obtained from four mice in each group. (F) Hydroxyproline contents in Plg+/+ and Plg−/− mice before and on day 7 after photoillumination. The data represent the mean ± SEM of eight mice in each group.

Download figure to PowerPoint

Angiogenesis at the border of the damaged area

BrdU-positive blood vessels were abundant in the granular cell-rich layer in Plg+/+ mice on day 4 after photoillumination (Fig. 7A), although they were less in number at the border of the damaged area in Plg−/− mice (Fig. 7B). In Plg+/+ mice, the total vessel density in the granular cell-rich layer increased time-dependently after photoillumination (Fig. 7C, open circle). The density of newly generated vessels, as indicated by BrdU immunoreactivity, was similar on days 4, 7 and 14 in Plg+/+ mice (Fig. 7C, solid column). The average luminal area of the vessels increased time-dependently in the granular cell-rich layer in Plg+/+ mice (Fig. 7D). These results show that blood vessels were generated constantly in the granular cell-rich layer after liver injury in Plg+/+ mice and the size of the generated blood vessels increased progressively. Furthermore, plasminogen also played a critical role in angiogenesis in the granular cell-rich layer during liver repair.

image

Figure 7.  Angiogenesis at the border of the damaged area. (A, B) Double staining for CD31 (red) and BrdU (green) at the border of the damaged area in Plg+/+ mice (A) and Plg−/− mice (B) on day 4 after photoillumination. The mice received the injection of BrdU twice a day over 4 days before they were sacrificed. The arrows indicate BrdU-positive blood vessels. The inset shows the enlarged image of the small square in (A). The area between the solid lines indicates the granular cell-rich layer in Plg+/+ mice (A). The dotted line indicates the border between the normal and the damaged areas in Plg−/− mice (B). Scale bars indicate 50 μm. Similar results were obtained from four mice in each group. (C, D) Quantification of the blood vessels in the granular cell-rich layer in Plg+/+ mice. The density (C) and average luminal area (D) of the blood vessels in the granular cell-rich layer are shown. The data represent the mean ± SEM of four mice in each group.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

In this study, we established a highly reproducible liver injury model by using a photochemical method to evaluate quantitatively the sequential responses during the process of liver repair, including the decrease in the area of the damaged tissue, macrophage accumulation, neutrophil recruitment, activated HSC accumulation accompanied by collagen fiber deposition, and blood vessel generation. By using this model, we have shown that accumulated neutrophils contribute to macrophage recruitment at the border of the damaged area via the production of MCP-1 during liver repair. Furthermore, we have demonstrated that neither a decrease in the area of the damaged tissue nor macrophage accumulation was observed in Plg−/− mice. In addition, neutrophil accumulation on day 4 after photoillumination and the subsequent responses, including collagen fiber deposition and angiogenesis, were impaired in Plg−/− mice. These findings indicate the essential role of plasminogen for the recovery processes after liver injury.

Though several liver injury models have been used [7–9,12,13], it was difficult to quantify the pathological or pathophysiological processes after the damage. In the present model, however, the area of injury was highly reproducible, as shown by the small standard deviation on day 1, indicating that the recovery processes could be evaluated quantitatively. In addition, the typical recovery processes of tissue repair [1,2], including neutrophil recruitment, formation of granulation tissue consisting of numerous macrophages, activated HSCs, collagen fibers, and blood vessels, and decrease in the area of damage, were observed comparably both in this model and after liver injury by ischemia-reperfusion or traumatic damage (data not shown). Altogether, our data indicate that this model is suitable for quantitative analysis of the subsequent responses in liver repair processes.

In the photochemical method, reactive oxygen generated by the combination of intravenous Rose Bengal injection and photoillumination injures the endothelial membrane at the irradiated site, which subsequently causes ischemic injury through thrombus formation [14–16]. In addition, because Rose Bengal is excreted in bile via hepatobiliary transport [19], it is likely that reactive oxygen generated in irradiated hepatocytes also causes injury.

In Plg−/− mice, the area of damage remained constant on day 56, although it almost disappeared on day 14 in Plg+/+ mice, as we found previously [8]. In general, damaged tissue is cleared by phagocytosis of neutrophils and macrophages [2]. Taken together with previous in vitro evidence that the phagocytic function of macrophages is comparable between Plg+/+ and Plg−/− mice [20], no decrease in the area of damage may result from impaired accumulation of macrophages and neutrophils in Plg−/− mice.

The impaired macrophage accumulation in Plg−/− mice was restored by exogenous plasminogen administration. In addition, tranexamic acid, which binds the kringle domain in plasminogen molecule, suppressed the macrophage accumulation in Plg+/+ mice. These data indicate that plasminogen plays a pivotal role in macrophage accumulation, which is likely to be induced through the binding of its kringle domains to the surface of macrophages. Our findings also suggest that the macrophage accumulation is partially regulated by a MCP-1-dependent mechanism.

In addition, a possibility that fibrin(ogen), a substrate of plasmin, is involved in the macrophage accumulation remained. To clarify this possibility, we have studied the effect of fibrinogen depletion by treatment of batroxobin, a snake venom-derived defibrinogenating drug. However, sufficient depletion of fibrinogen over the experiments was not obtained (see Supporting Information). Therefore, further studies are needed to explore the role of fibrin(ogen) in the macrophage accumulation.

Plasmin is known to contribute to macrophage-related events, including proteolytic activation of MCP-1 [21], migration through activation of matrix metalloproteinase-9 [22], and binding to histone H2B, which is expressed on monocytes [23], and chemoattractant activity against leukocytes [24]. Therefore, the complete lack of these effects is most likely responsible for the impairment of macrophage accumulation in Plg−/− mice.

We observed that the neutrophil recruitment in the damaged area in Plg−/− mice was comparable with that in Plg+/+ mice on day 1. However, neutrophil recruitment to the border of the damaged area, as observed in Plg+/+ mice, was attenuated in Plg−/− mice on day 4 and was restored by exogenous Glu-plasminogen administration. Furthermore, neutrophil recruitment on day 4 was significantly suppressed by tranexamic acid in Plg+/+ mice. These observations are in line with those of a previous study showing that in the late-phase, but not the acute-phase, neutrophil recruitment to biomaterials implanted into the abdominal cavity is impaired in Plg−/− mice [25]. These findings confirm the important role of plasminogen in late-phase neutrophil recruitment during the physiological recovery process after liver injury. Considering the differences in neutrophil recruitment between day 1 and day 4 after photoillumination, the impaired neutrophil recruitment on day 4 in Plg−/− mice might be attributable to reduced levels of several chemoattractants for neutrophils from cells around the damaged area rather than the impairment of migration itself.

Recruited neutrophils are known to produce MCP-1 in other pathological models, including bleomycin-induced lung fibrosis [26] and collagen-induced arthritis [27], suggesting that such neutrophils generally trigger macrophage accumulation in damaged tissue. We found that the recruited neutrophils in the granular cell-rich layer expressed MCP-1 and that MCP-1 neutralization suppressed macrophage accumulation (Fig. 4), which is in line with the above observations [26,27]. In the neutrophil-depleted Plg+/+ mice, macrophages were observed extensively at not only the border of the damaged area but also the normal area (Fig. 5D), accompanied by an impairment of the decrease in the size of the damaged area (Fig. 5E). Considering the data obtained on the neutrophil and/or macrophage recruitment (Figs 4 and 5), complementary action between neutrophils and macrophages coordinated with MCP-1 intervened in the formation of the border layer of the damaged area in normal mice. On the other hand, the diminished macrophage accumulation in Plg−/− mice was thought to have resulted in the defect in activation of MCP-1 by plasmin [21] and macrophage migration [22].

Addendum

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

N. Kawao and N. Nagai designed the research, analyzed the data, and wrote the manuscript; C. Ishida, K. Okada, K. Okumoto and S. Ueshima conducted the research; Y. Suzuki and K. Umemura contributed to establishment of the photochemically induced liver injury model; and O. Matsuo has full responsibility for the present research.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

The authors thank S. Kurashimo (Center for Instrumental Analyses, Central Research Facilities, Kinki University School of Medicine) for supporting the flow cytometric analysis and K. Takenishi for supporting the histological analysis. This study was partly supported by a Grant-in-Aid for Young Scientists (B: 20790182) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, a grant from the High-Tech Research Center at Kinki University, Graduate School of Medicine from the Japan Society for the Promotion of Science (JSPS), and a grant from Medical and Engineering Link at Kinki University.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  11. Supporting Information

Fig. S1. Photographs of the liver surface and H&E-stained vertical sections on day 1 in the absence or presence of Rose Bengal infusion in Plg+/+ mice.

Fig. S2. Microphotographs with higher magnification of the H&E-stained and F4/80-immunostained vertical sections on day 56 after photoillumination.

Fig. S3. Microphotographs of fibrinogen-immunoreactivity at the damaged area after photoillumination.

Data S1. Supplementary Materials and Methods.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
JTH_3870_sm_R1-SupportingInformationfinal.pdf4534KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.