Latency-associated Peptide Degradation Fragments Produced in Stellate Cells and Phagocytosed by Macrophages in Bile Duct-ligated Mouse Liver

Abstract

Transforming growth factor-β (TGF-β) activation is involved in various pathogeneses, such as fibrosis and malignancy. We previously showed that TGF-β was activated by serine protease plasma kallikrein-dependent digestion of latency-associated peptides (LAPs) and developed a method to detect LAP degradation products (LAP-DPs) in the liver and blood using specific monoclonal antibodies. Clinical studies have revealed that blood LAP-DPs are formed in the early stages of liver fibrosis. This study aimed to identify the cell source of LAP-DP formation during liver fibrosis. The N-terminals of LAP-DPs ending at residue Arg58 (R58) were stained in liver sections of a bile duct-ligated liver fibrosis model at 3 and 13 days. R58 LAP-DPs were detected in quiescent hepatic stellate cells at day 3 and in macrophages on day 13 after ligation of the bile duct. We then performed a detailed analysis of the axial localization of R58 signals in a single macrophage, visualized the cell membrane with the anti-CLEC4F antibody, and found R58 LAP-DPs surrounded by the membrane in phagocytosed debris that appeared to be dead cells. These findings suggest that in the early stages of liver fibrosis, TGF-β is activated on the membrane of stellate cells, and then the cells are phagocytosed after cell death:

Introduction

Transforming growth factor-β (TGF-β) is secreted into the extracellular matrix (ECM) as a dimer of a latency-associated peptide (LAP) and is deposited with latent TGF-β binding proteins (LTBPs).^1,2 TGF-β is activated by LAP structural changes, which are induced by integrin binding or protease digestion. Moreover, TGF-β induces transcription via Smad activation. Active TGF-β plays an important role in fibrogenesis, carcinogenesis, and epithelial–mesenchymal transition.^3,4 In liver fibrosis, hepatic stellate cells are activated by TGF-β, consequently promoting fibrosis through collagen secretion. ¹

LAP binds to integrin αVβ6 and αVβ8 by its RGD motif, and active TGF-β is then released from LAP. It has been proposed that active TGF-β is released from structurally altered LAP after integrin binding. ⁵ Furthermore, it was reported that when αVβ8 integrin, glycoprotein-A repetitions predominant protein, and LAP form a complex, it binds to the TGF-β receptor and activates Smad signals without releasing active TGF-β. ⁶

TGF-β is also activated by proteases such as plasma kallikrein, plasmin, and matrix metalloproteases. ⁷ We focused on the molecular mechanism and diagnosis of TGF-β activation induced by the digestion of LAP by plasma kallikrein. Plasma kallikrein digests LAP between Arg58 (R58) and Leu59 (L59). The N-terminal fragment of LAP containing R58 (R58 LAP) remains on the ECM or cell surface, and the C-terminal fragment of LAP containing L59 (L59 LAP) is released into the extracellular space. To detect plasma kallikrein-dependent activation, we generated monoclonal antibodies against R58 LAP and L59 LAP, which recognize the cleaved sites at the LAP N-terminals and C-terminals, respectively (Fig. 1). ⁸ The L59 LAP fragment is released into the blood and is mainly detected in the early F1 stage of liver fibrosis.^9,10 Meanwhile, R58 LAP fragments are deposited on the ECM and can be immunostained with an antibody against R58 LAP, as shown in liver sections of CCl₄-treated and bile duct-ligated liver fibrosis mouse models. ¹¹ Furthermore, R58 LAP has been detected in liver sections of patients with non-alcoholic steatohepatitis (NASH) ¹⁰ and in extrahepatic fibrosis, particularly pancreatic fibrosis. ¹² Therefore, the role of plasma kallikrein in TGF-β activation during fibrosis has been established.

Figure 1.

Inactivated TGF-β is deposited on extracellular matrix through latent TGF-β binding proteins (LTBPs). Cleavage between R58 and L59 within latency-associated peptide (LAP) activates TGF-β. Plasma kallikrein (PLK) mediates the cleavage of LAP and forms LAP degradation products (LAP-DP), namely R58 LAP-DP and L59 LAP-DP. The R58 LAP-DP is deposited into extracellular matrix and stained with anti-R58 antibody in the liver section. Meanwhile, L59 LAP-D is released into the blood and detected by ELISA from the plasma. Abbreviation: TGF-β, transforming growth factor-β; LTBP, latent transforming growth factor-β binding protein.

In this study, we aimed to identify the cell source of LAP-DP formation during liver injury and fibrosis. To achieve this, the time course of plasma kallikrein-dependent LAP degradation was monitored in bile duct-ligated liver fibrosis models using multiple immunohistochemistry tests.

Materials and Methods

Materials

Mouse monoclonal anti-R58 LAP-DP antibody was prepared as described previously ¹¹ or was provided by Cosmo Bio (Tokyo, Japan). The following primary antibodies were purchased from each provider as follows: goat polyclonal anti-Iba1 antibody (ab107159; Abcam, Cambridge, UK), rabbit polyclonal anti-α-smooth muscle actin (αSMA) antibody (ab5694; Abcam), rabbit polyclonal antidesmin antibody (RB-9014-P; Thermo Fisher Scientific, Waltham, MA), rabbit polyclonal antilecithin: retinol acyltransferase (LRAT) antibody (38075; IBL, Gunma, Japan), and anti-mouse C-type lectin domain family 4 (CLEC4F)/C-type lectin superfamily member 13 (CLECSF13) antibody (AF2784; R&D systems, Minneapolis, MN). The following secondary antibodies were purchased from Thermo Fisher Scientific: Alexa Fluor 488-antirabbit IgG (A-21206), Alexa Fluor 594-anti-mouse IgG (A-21203), and Alexa Fluor 647-antigoat IgG (A32849).

Animal Experiments

The mice were handled in accordance with the Animal Experiment Committee of the RIKEN Wako Institute and SLC (Hamamatsu, Japan) with approval numbers W2020-2-001 and BT18107, respectively. C57BL/6J male mice purchased from SLC at 8 weeks of age underwent bile duct ligation (BDL) or sham surgery for 3 and 13 days under 2.5% isoflurane (FUJIFILM Wako Pure Chemical; Osaka, Japan) in air induce anesthesia by SLC and RIKEN, respectively, as described in previous studies.^13,14

Immunohistochemistry

Mice were perfused intracardially with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) at 3 and 13 days post-BDL. The liver was removed and post-fixed for 12 hr at 4C, in Celsius units, in PBS containing 4% PFA. After paraffinization, paraffin blocks were sectioned at a thickness of 4 µm and then costained with antibodies against R58 LAP-DP, macrophage markers (Iba1 and CLEC4F), and hepatic stellate cell markers (αSMA, desmin, and LRAT) after antigen retrieval treatment within the Target Retrieval Solution, pH 9 (Dako; Glostrup, Denmark) using microwaves at 98C, in Celsius units, for 10 min. The following primary antibodies were used: anti-R58 LAP-DP antibody (10 µg/ml), anti-Iba1 antibody (1:1000), anti-αSMA antibody (1:100), antidesmin antibody (1:200), anti-LRAT antibody (1:50) and anti-CLEC4F antibody (2 µg/ml). The primary antibodies were detected with the following secondary antibodies: Alexa Fluor: 488-antirabbit IgG, 594-anti-mouse IgG, and 647-antigoat IgG (1:1000). Nuclei were stained with 4′,6-diamidino-2-phenylindole. Images were taken under a confocal microscopy LSM700 (ZEISS; Jena, Germany) and super-resolution laser scanning confocal microscopy (TCS SP8 STED ONE; Leica Microsystems, Inc., Biberach, Germany).

Results

LAP-DPs are mainly detected in the early F1 stage of liver fibrosis; thus, plasma kallikrein-dependent TGF-β activation is expected to occur at the same stage. To reveal the time course and production sites of LAP-DPs, we performed immunostaining of liver sections prepared from bile duct-ligated liver fibrosis mouse models using anti-R58 LAP-DP antibody. At days 3 and 13 post-BDL, the liver specimen was fixed, and then LAP-DPs were stained with anti-R58 LAP-DP antibody (Figs. 2 and 3).

Figure 2.

Immunostaining of mice liver sections 3 days after bile duct ligation (BDL) or sham operation, R58 (red) and Iba1 (green) and hepatic stellate cell marker (light blue) with DAPI (blue) counterstain for nuclei. (A) αSMA is activated hepatic stellate cell marker. R58 LAP-DPs are not colocalized with αSMA (open arrowhead). (B and C) The sections were stained with quiescent hepatic stellate cell markers desmin in (B) and LRAT in (C) (light blue). R58 LAP-DPs are localized in desmin⁺ LRAT⁺ αSMA⁻ quiescent hepatic stellate cells at day 3 post-BDL (white arrowheads). Scale bar indicates 50 µm. Abbreviations: LAP-DPs, LAP degradation products; LAP, latency-associated peptide; αSMA, α-smooth muscle actin; LRAT, lecithin: retinol acyltransferase; DAPI, 4′,6-diamidino-2-phenylindole.

Figure 3.

Immunostaining of mice liver sections 13 days after bile duct ligation (BDL) or sham operation, R58 (red) and Iba1 (green) and hepatic stellate cell marker (light blue) with DAPI (blue) for counterstaining for nuclei. (A) At day 13 post-BDL or post-sham operation, the liver sections are stained with αSMA (light blue), activated hepatic stellate cell marker. (B and C) The liver sections are stained with quiescent hepatic stellate cell markers, desmin in (B) and LRAT in (C) (light blue). Several R58 LAP-DPs are colocalized with Iba1-positive macrophages at day 13 post-BDL (white arrowheads). Scale bar indicates 50 µm. Abbreviations: LAP-DPs, LAP degradation products; LAP, latency-associated peptide; αSMA, α-smooth muscle actin; LRAT, lecithin: retinol acyltransferase; DAPI, 4′,6-diamidino-2-phenylindole.

At day 3 post-BDL, R58-positive cells positively stained with antidesmin and anti-LRAT antibodies, and negatively stained with anti-αSMA and anti-Iba1 antibodies (Fig. 2). Therefore, after 3 days, the R58-positive cells were determined to be quiescent stellate cells. The cells have several protrusions, and these protrusions were stained with a dot-like pattern depending on the angle of sections (Fig. 2).

Meanwhile, at day 13 post-BDL, an R58-positive large mass was mainly observed inside the Iba1⁺ desmin⁻ LRAT⁻ macrophages (Fig. 3). To clarify the subcellular localization of the R58-positive large mass, the liver sections at day 13 were stained with anti-R58 and anti-CLEC4F antibodies, with the latter as a marker for macrophage membrane antibody. It was observed that the R58-positive large mass was surrounded by a macrophage membrane (Fig. 4A and B). Furthermore, the axial localization of R58 signals in a single macrophage, traced using Leica’s super-resolution microscope system, showed that R58 signals observed on the cell surface were engulfed by macrophages (Fig. 4C). Collectively, R58-positive cell debris seemed to be phagocytosed by macrophages.

Figure 4.

Immunostaining of mice liver sections 13 days after bile duct ligation (BDL) or sham operation, R58 (red) and CLEC4F (green) with DAPI (blue) counterstain for nuclei. (A) Several R58 LAP-DPs are localized in macrophages at day 13 post-BDL. Scale bar indicates 50 µm. (B) Magnified view of A. Scale bar indicates 25 µm. (C) Axial localization of R58 signals in the single macrophage using Leica’s super-resolution microscope system. Layer 1 and layer 2 are upper side and lower side of the macrophage, respectively. The signals of R58 LAP-DPs observed on the cell surface were engulfed by macrophage. Nuclei of macrophages are shown by white arrowheads. Scale bars indicate 10 µm. Abbreviations: LAP-DPs, LAP degradation products; LAP, latency-associated peptide; DAPI, 4′,6-diamidino-2-phenylindole.

Discussion

We showed that granulomatous lesions were mainly stained with anti-R58 antibody 13 days after BDL, which suggests that the lesions consisted of microglia and Kupffer cells. ¹¹ In this study, hepatic stellate cells that were αSMA-negative, desmin-positive, and LRAT-positive had R58-positive staining and appeared linear on day 3 post-BDL. On day 13 post-BDL, several R58-positive stains were observed in a dot-like pattern and were costained with Iba1-positive cells. Furthermore, macrophage membranes with CLEC4F-positive staining suggested that the R58-positive debris from dead cells may have been taken up by macrophages. Thus, in the BDL model, plasma kallikrein-dependent activation of TGF-β occurs relatively early, approximately 3 days after surgery, on desmin⁺ LRAT⁺ αSMA⁻ quiescent hepatic stellate cells.

In contrast, another study showed that after administration of CCl₄ for 12 weeks, R58-positive cells were costained with αSMA and were negative for CD31 and F4/80 in the liver fibrosis model, ¹¹ implying that the cells were activated hepatic stellate cells. This supports the results of this study, which suggest that plasma kallikrein-dependent TGF-β activation mainly occurs on hepatic stellate cells at the early stage of fibrosis, even before the activation of stellate cells. In addition, in the later stages, these R58-positive stellate cells undergo cell death and are taken up by macrophages. On day 3 post-BDL, R58-positive LAP-DPs were costained with desmin and LRAT, while on day 13, R58-positive cell debris was not costained with desmin and LRAT. R58-positive LAP-DPs were deposited on the ECM, whereas desmin and LRAT were localized in cytoplasm. Thus, desmin and LRAT can be easily degraded during cell death. Due to the R58-positive staining in the granulomatous lesions, it is postulated that cell death and macrophage infiltration frequently occur in lesional areas. ¹¹ R58 LAP-DP positivity was observed in macrophages with a crown-like structure in the NASH model. ¹⁵ In the acute phase of liver injury, R58 LAP-DPs are observed in F4/80-positive macrophages after 2 hr of anti-Fas antibody (Jo2) injection. ¹⁶ These suggest that the majority of the stellate cells gradually died 13 days post-BDL, and the resulting R58-positive cell debris was phagocytosed by macrophages.

LAP-DP, which is cleaved in plasma kallikrein and released into the blood, is mainly detected in the F1 stage by the anti-L59 antibody, implying that early fibrosis causes plasma kallikrein-dependent activation of TGF-β. ¹⁰ In patients with chronic liver disease, R58 LAP-DP was robustly expressed in and around sinusoidal cells before regional fibrosis. ¹⁰ Using multiple immunohistochemistry with anti-R58 antibody and various hepatic cell markers in a liver fibrosis model with BDL, we showed that LAP-DPs are mainly produced in quiescent hepatic stellate cells. Compared with the current biomarkers observed in activated stellate cells, these fragments have potential as sensitive biomarkers for early liver fibrosis. Generally, proteolytic TGF-β activation occurs in the early (F1) stage in sinusoidal cells and seems to switch to integrin-associated TGF-β activation reported to be abundant in the late (F3 and F4) stages. This occurs with the enhanced expression of integrins, underlining the possible use of proteolytic TGF-β activation as a biomarker and therapeutic target for early liver fibrogenesis, and shows that integrin activation of TGF-β is required for late-stage fibrogenesis and cancer development.