Essential involvement of the CX3CL1-CX3CR1 axis in bleomycin-induced pulmonary fibrosis via regulation of fibrocyte and M2 macrophage migration

The potential role of macrophages in pulmonary fibrosis (PF) prompted us to evaluate the roles of CX3CR1, a chemokine receptor abundantly expressed in macrophages during bleomycin (BLM)-induced PF. Intratracheal BLM injection induced infiltration of leukocytes such as macrophages into the lungs, which eventually resulted in fibrosis. CX3CR1 expression was mainly detected in the majority of macrophages and in a small portion of α-smooth muscle actin-positive cells in the lungs, while CX3CL1 was expressed in macrophages. BLM-induced fibrotic changes in the lungs were reduced without any changes in the number of leukocytes in Cx3cr1−/− mice, as compared with those in the wild-type (WT) mice. However, intrapulmonary CX3CR1+ macrophages displayed pro-fibrotic M2 phenotypes; lack of CX3CR1 skewed their phenotypes toward M1 in BLM-challenged lungs. Moreover, fibrocytes expressed CX3CR1, and were increased in BLM-challenged WT lungs. The number of intrapulmonary fibrocytes was decreased in Cx3cr1−/− mice. Thus, locally-produced CX3CL1 can promote PF development primarily by attracting CX3CR1-expressing M2 macrophages and fibrocytes into the lungs.

disease, and amyotrophic lateral sclerosis 12 . Furthermore, the CX3CL1-CX3CR1 axis could maintain β cell function to preserve sufficient insulin secretion 13 . We have previously demonstrated that CX3CL1 protects against polymicrobial sepsis and toxin A-induced enteritis by binding to CX3CR1-expressing macrophages 14,15 .
PF is characterized pathologically by diffuse interstitial inflammation and fibrosis 16 . Moreover, PF can develop from various conditions, such as irradiation injury, oxygen toxicity-related pneumonitis, and scleroderma; however, approximately half of the PF cases are idiopathic. Environmental exposures to various substances 17 , as well as genetic factors 18 , have been proposed to be implicated in the development of idiopathic PF (IPF). However, their contribution to disease pathogenesis remains to be investigated. Aberrant tyrosine kinase activation in IPF has incited the clinical application of nintedanib, a multi-tyrosine kinase inhibitor. Moreover, pirfenidone, an agent with anti-inflammatory and anti-oxidant activities, was also examined for its effectiveness against IPF 19 . While both agents can delay the decline in respiratory function in IPF, they are unable to cure the disease. As a consequence, it is necessary to establish a novel treatment strategy based on the molecular and cellular mechanisms of IPF pathogenesis.
Bleomycin (BLM) can induce lung injuries that mimic the pathological features observed in human IPF. Analysis of BLM-induced lung injuries revealed the contribution of pulmonary macrophages to disease onset [20][21][22][23][24] . The expression of CX3CR1 on monocytes/macrophages prompted us to investigate the roles of the CX3CL1-CX3CR1 axis in BLM-induced lung injury in CX3CR1-deficient (Cx3cr1 −/− ) mice. We revealed that genetic ablation of Cx3xr1 significantly attenuated BLM-induced lung fibrosis.

Results
Intrapulmonary expression of CX3CL1 and CX3CR1 following BLM treatment. We first evaluated gene expression of Cx3cl1 and Cx3cr1 in lungs of WT mice following intratracheal BLM administration. Both Cx3cl1 and Cx3cr1 mRNAs were detected in lungs of untreated WT mice. Expression of Cx3cl1 was significantly upregulated at 1 day and 3 days after the BLM challenge, followed by augmentation in Cx3cr1 expression ( Fig. 1a and b). Double-color immunofluorescence analysis demonstrated that F4/80 + cells, but not α-SMA + fibroblasts and epithelial cells, were major cellular sources of CX3CL1 ( Fig. 1c and Supp. Figure 1). Moreover, CX3CR1 proteins were detected in F4/80 + macrophages (Fig. 1d) and in a small proportion of α-SMA + fibroblasts (Fig. 1e). Thus, intratracheal BLM injection induced F4/80 + cells mainly in the lungs to produce CX3CL1. Upregulation of CX3CL1 can attract F4/80 + macrophages and α-SMA + fibroblasts through interactions with CX3CR1 expressed on these cells.
Pathogenic roles of the CX3CL1-CX3CR1 axis in BLM-induced PF. In order to determine the pathophysiological roles of the CX3CL1-CX3CR1 axis in BLM-induced PF, we challenged WT and Cx3cr1 −/− mice with intratracheal BLM. No significant histological differences were observed between the lungs of untreated WT and Cx3cr1 −/− mice (Fig. 2a). At 21 days after BLM treatment, WT mice exhibited severe destruction of pulmonary alveolar structures and massive collagen deposition in the lungs, as evidenced by positive Masson staining (Fig. 2a). On the contrary, these fibrotic changes were markedly attenuated in Cx3cr1 −/− mice (Fig. 2a). Moreover, Col1a1 mRNA expression was significantly elevated in the lungs of WT mice at 7 and 14 days after BLM treatment; this enhancement was reduced in Cx3cr1 −/− mice (Fig. 2b). The intrapulmonary content of hydroxyproline (Hyp), a major component of collagen, was also consistently increased in WT, but not Cx3cr1 −/− mice, at 21 days after BLM treatment (Fig. 2c). We further challenged bone marrow (BM) chimeric mice generated from WT and Cx3cr1 −/− mice with BLM, as CX3CR1 is expressed by BM-and non-BM-derived cells 25 . BLM treatment increased intrapulmonary Hyp content to a similar level as that in WT and Cx3cr1 −/− mice with WT-derived BM (Fig. 2d). On the contrary, the increased Hyp content was reduced to a similar level in WT and Cx3cr1 −/− mice with Cx3cr1 −/--derived BM. Therefore, BM-derived CX3CR1 + cells may play a major role in the pathogenesis of BLM-induced PF.
Involvement of the CX3CL1-CX3CR1 axis in intrapulmonary M2-macrophage infiltration. As leukocytes are hypothesized to play essential roles in the development of PF 26 , we wanted to evaluate leukocyte infiltration following BLM challenge. We did not observe any differences in the number of granulocytes, macrophages, and T cells between untreated WT and Cx3cr1 −/− mice (Fig. 3). BLM challenge induced infiltration of granulocytes, macrophages, and T cells to similar extents in WT and Cx3cr1 −/− mice (Fig. 3), which was confirmed at 5 days after BLM injection by flow cytometric analyses of bronchoalveolar lavage fluids (BALFs) (Supp. Figure 2). Therefore, absence of the CX3CL1-CX3CR1 axis may trigger compensatory enhancements in the expression of other chemokines, which lead to subsequent maintenance of inflammatory cell infiltration. However, this assumption was negated by the observation that absence of CX3CR1 had few impacts on the intrapulmonary expression of several chemokine and chemokine receptor genes, which were proposed to be involved in the pathogenesis of BLM-induced PF (Supp. Figure 3) 26 . Several lines of evidence demonstrated that functionally polarized macrophage subtypes, M1 and M2, are crucial for inflammation and tissue repair, respectively 27 . M2-macrophages are known to be important for tissue fibrosis 28 . As CX3CR1 was expressed by macrophages recruited into the lungs (Fig. 1d), we evaluated the phenotypes of BALF macrophages in WT and Cx3cr1 −/− mice following BLM challenge. In the BALF of BLM-treated WT mice, the number of CD206 + macrophages (M2-type) dominated over CD86 + macrophages (M1-type). However, in BLM-treated Cx3cr1 −/− mice, the opposite trend was observed (Fig. 4a). Immunofluorescence analysis demonstrated that CD206 and CD68 were simultaneously expressed by CX3CR1-expressing cells in the lung parenchyma after BLM challenge ( Fig. 4b  and c). Thus, the CX3CL1-CX3CR1 axis can play a crucial role in the intrapulmonary recruitment of pro-fibrotic M2-macrophages that eventually contribute to BLM-induced PF.
SCIenTIFIC REpoRtS | 7: 16833 | DOI:10.1038/s41598-017-17007-8 Reduced intrapulmonary fibrocyte accumulation in Cx3cr1 −/− mice. Given the crucial involvement of BM-derived CX3CR1 + cells in lung fibrosis, we focused on an additional cell type present in the BM, the fibrocytes. These cells were identified as CD45 + Col-I + BM cells and hypothesized to be essential for tissue repair and abnormal fibrosis [29][30][31] . We generated BM chimeric mice by transplanting lethally irradiated WT mice with GFP-transgenic (Tg) mouse-derived BM cells. When WT mice bearing GFP + BM cells were intratracheally injected with BLM, GFP + cells was detected in the lungs, indicating that BM cells were recruited into the lungs after BLM treatment. In addition, Col-I was also detected in a substantial proportion of GFP + cells ( Fig. 5a and  b). Moreover, immunofluorescence analyses revealed that both Col-I and CD45 were present in a large proportion of GFP + cells recruited from the BM following BLM treatment (Fig. 5c). These observations suggested that BM-derived CD45 + Col-I + fibrocytes migrate to lungs after BLM treatment. Flow cytometry analysis revealed that CX3CR1 was expressed in approximately 60% of CD45 + Col-I + fibrocytes in the BM of WT mice after BLM treatment (Fig. 5d). We next determined the number of fibrocytes in the BM, peripheral blood, and lung tissues; there were no significant differences in fibrocyte numbers between untreated WT and Cx3cr1 −/− mice ( Fig. 6 and Supp. Figure 4). BLM challenge increased the number of fibrocytes to similar extents in the BM and peripheral blood of WT and Cx3cr1 −/− mice (Supp. Figure 4). Number of fibrocytes in the lungs were markedly increased in WT mice at 12 days after BLM treatment. However, this enhancement in the number of fibrocytes was suppressed in Cx3cr1 −/− mice (Fig. 6). Collectively, these results suggested that locally-produced CX3CL1 can contribute to BLM-induced PF by recruiting CX3CR1-expressing fibrocytes that produce Col-I from the BM into the lungs.
Reduced TGF-β1 expression in mice lacking CX3CR1. TGF-β1 is a potent fibrogenic growth factor in various organs. We examined the cell types expressing TGF-β1 in the lungs after BLM treatment. Quantitative RT-PCR analyses of Cx3cl1 and Cx3cr1 gene expression was carried out. Values represent mean ± SEM (n = 6). *P < 0.05; **P < 0.01, vs. unchallenged lungs (time = 0). (c) to (e) Cell types expressing CX3CL1 and CX3CR1 in lungs of WT mice at 3 days after BLM exposure. Double-color immunofluorescence analyses were performed; representative images from six individual animals are shown here. Signals were merged digitally.
Immunofluorescence analyses showed expression of TGF-β1 in CX3CR1 + cells (Fig. 7a), which suggested that CX3CR1-expressing cells, fibrocytes, and M2-macrophages could be a major source of TGF-β1 in the lungs. Moreover, BLM treatment progressively enhanced TGF-β1 expression at both the gene and protein levels in the lungs of WT mice; this enhancement was markedly suppressed in the lungs of Cx3cr1 −/− mice ( Fig. 7b and c). As a result, the CX3CL1-CX3CR1 axis can have an indispensable role in BLM-induced lung fibrosis by inducing the migration of TGF-β1-producing fibrocytes and M2-macrophages into the lungs.

Discussion
It is highly evident from previous studies that macrophages are indispensable in tissue repair and the subsequent fibrotic changes 32,33 . Robust CX3CR1 expression in macrophages 2,3 prompted several scientific groups to investigate the roles of the CX3CL1-CX3CR1 axis in fibrosis in various organs. Studies have shown that hypertensionand unilateral ureteral obstruction-induced renal fibrosis are attenuated by Cx3cr1 gene deficiency [34][35][36] . On the contrary, the interplay between CX3CL1 and CX3CR1 can prevent carbon tetrachloride-induced liver injury 37,38 . These discrepancies may be explained by organ-specific differences in macrophage populations; CX3CR1-expressing macrophages may exert anti-inflammatory activities, as observed in corneal tissues 39 . Hence, we examined the effect of CX3CR1 deficiency on BLM-induced PF. Our results revealed that CX3CR1 deficiency inhibits BLM-induced lung pathologies, but does not exert any effect on infiltration of inflammatory cells, including macrophages.
Macrophages in the lungs belong to two subpopulations, interstitial macrophages (IMs) and alveolar macrophages (AMs), residing in the interstitial and alveolar spaces, respectively 40 . Ly6 high CCR2 + inflammatory monocytes in the circulation are recruited into lung tissues, most likely in a CCL2-dependent manner, similar to lung metastasis in breast cancer 41 . On the contrary, AMs originate from fetal liver monocytes, and self-maintain throughout their life cycle under steady state 42 . Allergen challenges reduced the number of embryo-derived AMs 43 and enhanced CCL2 expression-induced migration of CCR2-expressing monocyte-derived AMs to replenish the AM loss 44 . Therefore, total number of intrapulmonary macrophages may not differ significantly between BLM-treated WT and Cx3cr1 −/− mice, since BLM may be able to enhance the intrapulmonary expression of the chemoattractant CCL2 for both IMs and AMs to similar extents.
Alternatively, there remains a possibility that CX3CR1 deficiency may have profound impact on macrophage functions. Macrophages can be functionally polarized into two phenotypes, M1 and M2. Skewing of macrophage polarization can contribute to pathogenesis of inflammatory responses, allergic diseases, tissue repair, and tumor progression [45][46][47] . Several lines of evidence indicated that macrophages with the M2 phenotype could promote tissue fibrosis 48,49 . Indeed, CX3CR1-expressing CD68 + cells in the lungs were also positive for CD206 50,51 , a M2-specific marker. This indicated that CX3CR1 + macrophages are M2-macrophages. Moreover, M2-macrophages were reduced in BLM-challenged lungs of Cx3cr1 −/− mice. This was concomitant with a reciprocal increase in CD68 + M1-macrophages. Therefore, absence of CX3CR1 can have profound effects on macrophage polarization rather than total macrophage numbers in the lungs after BLM challenge.
Myofibroblasts display characteristics of both fibroblasts and smooth muscle cells, as represented by their ECM-synthesizing capacity and contractile cytoskeleton, respectively 52 . As a result, myofibroblasts have indispensable roles in connective tissue remodeling and the subsequent development of fibrosis in various organs, including the lung. Myofibroblasts can be differentiated from several cell types, including locally residing mesenchymal cells (e.g., fibroblasts and smooth muscle cells), other local sources such as epithelial and endothelial cells, and BM-derived fibrocytes 53 . Studies on BM chimeric mice have revealed the crucial role of radiosensitive BM-derived CX3CR1-expressing cells in BLM-induced collagen deposition in the lung. Moreover, injection of BLM into chimeric mice with BM derived from GFP Tg mice resulted in cellular expression of GFP + and Col-I + in the lung. Therefore, fibrocytes may be a major source of collagen-producing myofibroblasts in BLM-induced PF.
Fibrocytes are defined as BM-derived cells that are intermediate between hematopoietic and mesenchymal cells, as they simultaneously express leukocyte markers such as CD45 and mesenchymal cell markers such as Col-I. Fibrocytes, once generated in the BM, circulate in the bloodstream, and migrate to organs in response to tissue-specific cues [29][30][31][54][55][56][57][58][59][60][61][62] . BLM was found to increase fibrocyte numbers in the BM, peripheral blood, and lungs, which is consistent with our previous report 26 . Moreover, fibrocytes were hypothesized to express several chemokine receptors, including CXCR4, CCR3, CCR5, and CCR7 31,63,64 , which allow them to migrate to fibrotic organs in response to their cognate ligands. It was found that inhibition of CXCR4 function 31 , or deficiencies in Ccr3 65 or Ccr5 26 attenuated BLM-induced pathologies, together with reduced intrapulmonary fibrocyte numbers. Our present observations suggested that CX3CR1 is one of the chemokine receptors involved in migration of fibrocytes into fibrotic lungs.
We have previously demonstrated that Ccl3 −/− and Ccr5 −/− mice exhibit reduction in BLM-induced fibrosis and the number of CD45 + Col-I + CCR5 + fibrocytes in the lungs, together with suppressed CXCL12 expression 26 . These observations indicated that the CCL3-CCR5 axis can mediate BLM-induced fibrocyte migration into the lungs, partly via its interactions with the CXCL12-CXCR4 axis. BLM-induced increase in intrapulmonary fibrocyte numbers was not completely nullified in Cx3cr1 −/− mice, suggesting the contribution of other chemokines such as CXCL12 and CCL3 to fibrocyte infiltration. The decreased BLM-induced fibrosis and fibrocyte infiltration into the lungs of Cx3cr1 −/− mice, however, were not accompanied with changes in intrapulmonary expression of CXCL12 and ligands for CCR5, CCL3, CCL4, and CCL5. Thus, the CX3CL1-CX3CR1 axis can regulate BLM-induced fibrocyte migration and subsequent fibrosis development independently of other chemokines. CX3CL1 has distinct structural and functional features from other chemokines. Most chemokines are produced as a secretory molecule, and bind to proteoglycan and glycosaminoglycan in the ECM through their carboxy-terminal α-helix portion 66 ; this binding is reversible and unstable. On the contrary, CX3CL1 is expressed as a membrane-bound molecule with a mucin-like stalk 2,3 . As compared with other chemokines, CX3CL1 can be firmly retained on the cell surface, and can consequently affect its target cells in a more stable manner. Moreover, in addition to chemotactic activities, CX3CL1-CX3CR1 interactions can promote the survival and/or proliferation of non-hematopoietic cells, such as neuronal cells 67 , and hematopoietic cells, such as monocytes 68 and helper T cells 9 . If CX3CL1 is able to support survival and/or proliferation of fibrocytes as well as non-hematopoietic cells, it can attract CX3CR1-expressing fibrocytes into the lung. This can eventually promote the survival and/or proliferation of fibrocytes, thereby inducing fibrotic changes.

BLM-induced lung Injury.
Mice were put under deep anesthesia with an intraperitoneal injection of pentobarbital (50 µg/g weight). A cervical midline incision was made, and the trachea was exposed. Thereafter, BLM (0.075 U) in 50 µl sterile saline was intratracheally administered with a 26-guage needle. At the indicated time intervals following BLM administration, mice were sacrificed by pentobarbital overdose, and both lungs were removed for subsequent analyses.
Generation of bone BM chimeric mice. BM chimeric mice were prepared as previously described 26 .
Briefly, male Cx3cr1 −/− BM → female WT mice, male WT BM → female Cx3cr1 −/− mice, male WT BM → female WT mice, and male Cx3cr1 −/− BM → female Cx3cr1 −/− mice. BM cells were collected from femurs of donor mice by aspiration and flushing. Recipient mice were irradiated to 15 Gy with a RX-650 irradiator (Faxitron X-ray Inc., Wheeling, IL). The animals then received intravenous injections of 5 × 10 6 BM cells in 200 μl sterile PBS(−) from donor mice under anesthesia. Thereafter, the mice were housed in sterilized microisolator cages, and were fed normal chow and autoclaved hyperchlorinated water for 60 days. Similarly, BM chimeric mice were generated by transplanting GFP-Tg mouse-derived BM cells into lethally irradiated WT mice. To verify successful engraftment and reconstitution of the BM in transplanted mice, genomic DNA was isolated from peripheral blood and tail tissues of each chimeric mouse at 30 days after BMT with a NucleoSpin tissue kit (Macherey-Nagel, Duren, Germany). Polymerase chain reaction (PCR) was used to detect the Sry gene in the Y chromosome (forward primer, 5′-TTGCCTCAACAAAA-3′; reverse primer, 5′-AAACTGCTGCTTCTGCTGGT-3′). The amplified PCR products were fractionated on a 2% agarose gel, and were visualized by ethidium bromide staining. After durable BM engraftment was confirmed, mice were treated with BLM, as described above.
Histopathological and immunohistochemical analyses. Lung tissues were fixed in 10% formalin buffered with PBS (pH 7.2), and were embedded in paraffin. Tissues were sectioned (6 µm-thick) and stained with hematoxylin and eosin or Masson's trichrome to detect collagen deposition. Immunohistochemical analyses were also performed using anti-Ly6G, anti-F4/80, or anti-CD3 Abs, as previously described 26 . Cells were quantified in 10 randomly chosen visual fields at 200× magnification, and the average of the 10 selected microscopic fields was calculated. All measurements were performed by an examiner without prior knowledge of the experimental procedures.
Double-and triple-color immunofluorescence analysis. Double-color immunofluorescence analysis was conducted to identify the types of CX3CL1-or CX3CR1-expressing cells in the lung, as previously described 69 . Briefly, deparaffinized sections were incubated with PBS containing 1% normal donkey serum and 1% BSA to reduce nonspecific reactions. Tissue sections were then incubated with a combination of anti-CX3CR1 and anti-F4/80 Abs, anti-CX3CL1 and anti-F4/80 Abs, or anti-CX3CR1 and anti-α-SMA Abs at 1 µg/ml at 4 °C overnight. Similarly, triple-color immunofluorescence analysis was conducted to identify the molecules expressed on BM-derived fibrocytes (GFP + Col-I + cells), such as TGF-β1 and CX3CR1. Sections containing GFP + cells were incubated with a combination of anti-Col-I and anti-CD45 Abs, anti-Col-I and anti-TGF-β1 Abs, or anti-Col-I and anti-CX3CR1 Abs at a concentration of 1 µg/ml at 4 °C overnight. In order to identify the subtypes of macrophages recruited into the lungs, tissue sections were incubated with a combination of anti-CD68, anti-CX3CR1, and anti-CD206 Abs, or anti-CD206 and anti-CX3CR1 Abs at a concentration of 1 mg/ml at 4 °C overnight. Following incubation with fluorochrome-conjugated secondary Abs, tissues were imaged under a fluorescence microscope.
Determination of Hyp content. Lung tissues were removed at 21 days after BLM administration to determine the content of Hyp, a major component of collagen, as previously described 26 . Data were expressed as the amount (μg) of Hyp per lung.
Flow cytometric analysis of leukocytes in BALF. BALF was collected as previously described with some modifications 26 . Briefly, mice were sacrificed by abdominal aorta dissection. The trachea was cannulated, and the airway lumen was washed two times with 1 ml ice-cold PBS. Single cell suspensions were incubated with 25 µg/ml Fc block (BD Biosciences, Piscataway NJ) for 15 min on ice to prevent nonspecific binding. Cells were stained with anti-CD68, anti-CD86, CD-206, anti-CD3, anti-CD4, anti-CD8, anti-NK1.1, anti-Gr-1, and anti-CD11b. Analysis was performed on a FACScan flow cytometer (BD Biosciences) using the FlowJo software (Tommy Digital Biology, Tokyo, Japan).
Quantitative RT-PCR analysis. Total RNA was extracted from lung tissue using ISOGEN (Nippon Gene, Toyama, Japan), according to the manufacturer's instructions. Total RNA (3 µg) was reverse transcribed to cDNA using PrimeScript ™ Reverse Transcriptase (Takara Bio, Shiga, Japan) with Oligo(dT) 15 primers. Quantitative PCR was carried out with the resultant cDNA as templates and SYBR ® Premix Ex Taq ™ II (Takara Bio) with specific primer sets, as previously described 69 (Supplemental Table 1). Primers were purchased from Takara Bio. Amplification and detection of mRNA were conducted using the Thermal Cycler Dice ® Real Time System (Takara Bio, TP800), according to the manufacturer's instructions. Expression levels of target genes were analyzed using the ΔΔCt comparative threshold method. The β-actin gene was used as internal control.
SCIenTIFIC REpoRtS | 7: 16833 | DOI:10.1038/s41598-017-17007-8 Determination of fibrocytes. Mononuclear cells were isolated from the lungs, the BM, and peripheral blood, as previously described 25 . Nonspecific binding was blocked by incubation with 25 µg/ml Fc block (BD Biosciences Pharmingen, Piscataway, NJ) for 15 min at 4 °C. Cells were stained with PerCP-labeled anti-CD45 and FITC-labeled anti-CD34. After extensive washing with PBS(−), cells were permeabilized using cytofix/cytoperm (BD Biosciences Pharmingen), stained with anti-Col-I pAbs, and incubated with PE-conjugated goat anti-rabbit Ig (Life Technologies, Grand Island, NY), as previously described 26 . After two PBS(−) washes, cells were fixed in 2% paraformaldehyde. Isotype-matched control immunoglobulins were used to detect non-specific binding. Stained cells were analyzed on the FACS Calibur flow cytometer. In parallel, BM cells from BLM-treated WT mice were stained with anti-Col-I, anti-CD45, and anti-CX3CR1 to determine the proportion of CX3CR1 + fibrocytes. Flow cytometry was performed on a FACS Calibur flow cytometer, and data were analyzed using the FlowJo software.
ELISA. Lung samples from WT mice were homogenized with PBS containing complete protease inhibitor cocktail (Roch Diagnostics, Mannheim, Germany). Homogenates were centrifuged at 10,000 × g for 10 min. Supernatants were used to quantify the active form of TGF-β1 with a commercial ELISA kit (R&D Systems), according to manufacturer's instructions. The detection limits was 7 pg/ml. Total protein in the supernatant was measured with a commercial kit (BCA protein assay kit, Pierce, Rockford, IL). Data were expressed as TGF-β1 (pg) per total protein (µg) for each sample.
Statistical analysis. The means and SEMs were calculated for all investigated parameters in the study.
Statistical significance was evaluated by ANOVA or the Mann-Whitney's U-test. P < 0.05 was accepted as being statistically significant.

Study approval.
All animal experiments were approved by the Committee on Animal Care and Use at the Wakayama Medical University. All methods were performed in accordance with relevant guidelines and regulations.
Data availability. Data that support the findings of this study are available from the corresponding author upon reasonable request.