Wound-Healing Process

(1)

of February 23, 2013.

This information is current as

Wound-Healing Process

in the Skin

β

and

TGF-γ

between

IFN-The Essential Involvement of Cross-Talk

Yoichiro Iwakura and Naofumi Mukaida

Yuko Ishida, Toshikazu Kondo, Tatsunori Takayasu,

http://www.jimmunol.org/content/172/3/1848

2004; 172:1848-1855; ;

J Immunol

References

http://www.jimmunol.org/content/172/3/1848.full#ref-list-1

, 19 of which you can access for free at:

cites 53 articles

This article

Subscriptions

http://jimmunol.org/subscriptions

is online at:

The Journal of Immunology

Information about subscribing to

Permissions

http://www.aai.org/ji/copyright.html

Submit copyright permission requests at:

Email Alerts

http://jimmunol.org/cgi/alerts/etoc

Receive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606.

Immunologists All rights reserved.

Copyright © 2004 by The American Association of

9650 Rockville Pike, Bethesda, MD 20814-3994.

The American Association of Immunologists, Inc.,

is published twice each month by

The Journal of Immunology

at Pennsylvania State Univ on February 23, 2013

http://jimmunol.org/

(2)

The Essential Involvement of Cross-Talk between IFN-

␥ and

TGF-

␤ in the Skin Wound-Healing Process

1 Yuko Ishida,*

†

Toshikazu Kondo,

†

Tatsunori Takayasu,

‡

Yoichiro Iwakura,

§

and

Naofumi Mukaida

2

*

Several lines of in vitro evidence suggest the potential role of IFN-␥ in angiogenesis and collagen deposition, two crucial steps in

the wound healing process. In this report, we examined the role of IFN-␥ in the skin wound healing process utilizing WT and

IFN-␥ KO mice. In WT mice, excisional wounding induced IFN-␥ mRNA and protein expression by infiltrating macrophages and

T cells, with a concomitant enhancement of IL-12 and IL-18 gene expression. Compared with WT mice, IFN-␥ KO mice exhibited

an accelerated wound healing as evidenced by rapid wound closure and granulation tissue formation. Moreover, IFN-␥ KO mice

exhibited enhanced angiogenesis with augmented vascular endothelial growth factor mRNA expression in wound sites, compared

with WT mice, despite a reduction in the infiltrating neutrophils, macrophages, and T cells. IFN-␥ KO mice also exhibited

accelerated collagen deposition with enhanced production of TGF-␤1 protein in wound sites, compared with WT mice.

Further-more, the absence of IFN-␥ augmented the TGF-␤1-mediated signaling pathway, as evidenced by increases in the levels of total

and phosphorylated Smad2 and a reciprocal decrease in the levels of Smad7. These results demonstrate that there is crosstalk

between the IFN-␥/Stat1 and TGF-␤1/Smad signaling pathways in the wound healing process. The Journal of Immunology, 2004,

172: 1848 –1855.

S

kin wound healing starts immediately after an injury and consists of three phases; inflammation, proliferation, and maturation. These phases proceed with a complicated but well organized interaction among various types of tissues and cells (1, 2). During the inflammatory phase, platelet aggregation at the injury site is followed by infiltration of leukocytes, including neu-trophils and macrophages, into the wound site. In the proliferative phase, re-epithelialization and newly formed granulation tissue be-gin to cover the wound area to repair tissue destruction. Moreover, collagen deposition is indispensable for granulation tissue forma-tion and accumulating evidence implicates TGF-␤1 as one of the essential factors that can regulate collagen deposition (3– 6). How-ever, the mechanisms regulating the production and activity of TGF-␤1 in vivo remain elusive.

IFN-␥ is mainly produced by NK cells and CD4⫹Th1 cells and has multiple effects on macrophages, NK cells, and T lymphocytes (7). Moreover, IFN-␥ can inhibit collagen synthesis by fibroblasts in vitro (8 –12). In line with these observations, the administration of exogenous IFN-␥ impaired collagen accumulation and disrupted wound strength, suggesting that IFN-␥ was deleterious to skin wound healing (13–16). However, the role of endogenous IFN-␥ in the skin wound healing process remains to be investigated.

After binding its specific receptor on the cell surface, IFN-␥ activates receptor-associated Janus kinases, leading to the phos-phorylation of specific tyrosine residues of Stat1 (17, 18). Stat1 mediates the biological activity of IFN-␥ by inducing the tran-scription of the target genes (19, 20). Accumulating evidence sug-gests that the IFN-␥/Stat1 system can modulate TGF-␤1 activity in vitro by interfering with its signaling molecules, the Smad proteins (21, 22). However, it remains to be investigated whether or not there is crosstalk between the IFN-␥/Stat1 and TGF-␤1/Smad sig-naling pathways in vivo, particularly in pathological conditions. In this report, we investigated the role of endogenous IFN-␥ in the skin wound healing process, particularly focusing on its interaction with the TGF-␤1/Smad system. In this study, we provided the first definitive evidence to indicate that endogenous IFN-␥ can nega-tively regulate the TGF-␤1 signaling pathway at wound sites in vivo, resulting in a retardation of the wound healing process.

Materials and Methods

Abs and reagents

The following mAb or polyclonal Abs (pAbs)3_{and recombinant protein}

were used in this study; rat anti-mouse F4/80 mAb and rat anti-mouse CD3 mAb (Dainippon Pharmaceutical Company, Osaka, Japan), rat anti-mouse IFN-␥ mAb (clone XMG 1.2) and rat anti-mouse CD31 mAb (BD Phar-Mingen, San Diego, CA), rabbit anti-myeloperoxidase (MPO) pAb (Neo-markers, Fremont, CA, USA), goat mouse Smad2 pAb, rabbit anti-Smad3 pAb, goat anti-mouse Smad7 pAb (Santa Cruz, CA, USA), rabbit phosphorylated Smad 2 (p-Smad 2) pAb (Upstate, USA), mouse anti-Stat-1 mAb, and mouse anti-p-anti-Stat-1 mAb (Transduction Laboratories, Burlington, CA, USA), mouse anti-␣-smooth muscle actin (SMA) mAb (clone asm-1, Boehringer Mannheim GmbH, Mannheim, Germany), rat anti-mouse IFN-␥ neutralizing mAb (a kind gift of Dr. H. Fujiwara, Osaka University), and recombinant murine IFN-␥ (PeproTech, London, U.K.).

*Division of Molecular Bioregulation, Cancer Research Institute, Kanazawa Univer-sity, Kanazawa, Japan;†_{Department of Legal Medicine, Wakayama Medical}

Univer-sity, Wakayama, Japan;‡_{Department of Forensic and Social Environmental Medicine,}

Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan; and

§_{Laboratory Animal Research Center, Institute of Medical Science, University of}

Tokyo, Tokyo, Japan

Received for publication July 17, 2003. Accepted for publication November 13, 2003. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1_{This work was supported in part by Grants-in-Aids from the Ministry of Education,}

Culture, Sports, Science, and Technology of the Japanese Government.

2_{Address correspondence and reprint requests to Dr. Naofumi Mukaida, Division of}

Molecular Bioregulation, Cancer Research Institute, Kanazawa University, Ka-nazawa, Japan, 13-1 Takara-machi, Kanazawa 920-0934, Japan. E-mail address: [email protected]

3_{Abbreviations used in this paper: pAb, polyclonal Ab; COL1A1, collagen 1A1; HP,}

hydroxyproline; MPO, myeloperoxidase; p-Smad2, phosphorylated Smad2; p-Stat-1, phosphorylated Stat-1;␣-SMA, ␣-smooth muscle actin; VEGF, vascular endothelial growth factor.

(3)

Mice

Pathogen-free eight- to 12-wk old male BALB/c mice were obtained from Sankyo Laboratories (Tokyo, Japan) and designated as WT mice in the following experiments. IFN-␥ KO mice, backcrossed to BALB/c mice for more than eight generations, were used in the experiments (23, 24). Age-matched male BALB/c-SCID mice were purchased from CLEA Japan, Inc (Tokyo, Japan). All of the mice were used for the experiments complied with the standards set out in the Guidelines for the Care and Laboratory Animals at the Takara-machi Campus of Kanazawa University and housed individually in cages under specific pathogen-free conditions during the whole course of the study.

Excisional wound preparation and macroscopic examination

Mice were anesthetized with i.p. administration of pentobarbital (50␮g/g weight), and full-thickness skin wounds were made in the dorsal skin under sterile conditions as described previously (25). Briefly, after shaving and cleaning with 70% ethanol, excisional full-thickness skin wounds were made in the dorsal skin by picking up a fold skin at the midline and punch-ing through two layers of skin with a sterile disposable biopsy punch (ameter of 4 mm, Kai Industries, Tokyo, Japan). Two wounds with a di-ameter of 4 mm were made at the same time, one wound on each side of midline. The same procedure was repeated on the same animals three times, generating six wounds, with three wounds at each side. Each wound site was digitally photographed at the indicated time intervals, and wound areas were determined on photographs using PhotoShop (Version 7.0 Adobe Systems, Tokyo, Japan) without a prior knowledge of the experi-mental procedures. Changes in wound areas were expressed as the per-centage of the initial wound areas. In another series of experiments, WT mice received i.p. injection of neutralizing anti-IFN-␥ mAb or control IgG (250␮g/mouse) once a day from Day 0 to 3, starting immediately after the wound preparation. In some experiments, wounds and their surrounding areas, including the scab and epithelial margins, were cut with a sterile disposable biopsy punch (diameter 8 mm, Kai Industries, Tokyo, Japan) at the indicated time intervals.

Histopathological analyses of wound sites

At the indicated intervals after the injury, wound specimens were removed and fixed in 4% formaldehyde buffered with PBS (pH 7.2) and then em-bedded with paraffin. Six-␮m thick sections were stained with hematoxylin and eosin for histological analysis. Immunohistochemical analyses were performed for the evaluation of leukocyte infiltration, angiogenesis, and IFN-␥ expression as described previously (25). A double-color immuno-fluorescence analysis was also conducted to identify the types of IFN- ␥-expressing cells and p-Smad2-positive cells, as described previously (26). In some experiments, the anti-IFN-␥ mAb was incubated with the indicated concentration of recombinant mouse IFN-␥ at 4°C overnight before use.

MPO assay

Myeloperoxidase activity was measured to evaluate neutrophil recruitment (25). Briefly, the excised wound samples were washed in PBS and homog-enized in 1 ml of 50 mM potassium phosphate buffer solution with 0.5% hexadecyl trimethyl ammonium bromide (Sigma-Aldrich, St. Louis, MO)

and 5 mM EDTA. The samples were sonicated for 20 s, freeze-thawed three times, and centrifuged at 12,000 rpm at 4°C. MPO activities in the supernatants were assayed using the SUMILON peroxidase assay kit (Sumitomo Bekuraito, Tokyo, Japan), according to the manufacturer’s in-structions. The data were expressed as absorbance divided by total dry weight (mg).

Measurement of hydroxyproline (HP) contents at wound sites

At the indicated time intervals after the injury, skin wound sites were removed using a sterile disposable biopsy punch (diameter 8 mm) and were dried for 16 h at 120 °C. As HP is a major constituent of and found almost exclusively in collagen, HP contents were measured as the index of col-lagen accumulation at the wound sites, as described previously (27). HP content was calculated by comparison to standards and expressed as the amount (␮g) per wound.

Extraction of total RNAs and RT-PCR

Total RNAs were extracted from uninjured and injured skin samples using ISOGENE (Nippon Gene, Toyama, Japan) according to the manufacturer’s instructions. Five␮g of total RNA was reverse-transcribed at 42°C for 1 h in 20␮l reaction mixture containing mouse Moloney leukemia virus re-verse transcriptase (Toyobo, Osaka, Japan) with oligo(dT) primers (Am-ersham-Pharmacia Biotech Japan, Tokyo, Japan). The resultant cDNAs were amplified together with Taq polymerase (Nippon Gene) using specific sets of primers for IFN-␥, IL-12p35, IL-12p40, IL-18, vascular endothelial growth factor (VEGF), COL1A1, and␤-actin (Table I). PCR amplification of each gene was conducted with the optimal cycles consisting of 94°C for 1 min, optimal annealing temperature shown in Table I for 1 min, and 72°C for 1 min, followed by incubation at 72°C for 3 min. The amplified PCR products were fractionated on a 2% agarose gel and visualized by ethidium bromide staining. The band intensities were measured using Image Anal-ysis software (version 1.61; National Institutes of Health, Bethesda, MD) and the ratios to␤-actin were calculated (23, 25).

Western blotting

At the indicating time intervals after the injury, wound samples were ho-mogenized with a lysis buffer (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1% Triton, 1 mM EDTA) containing Complete Protease Inhibitor Cocktail (Roche, Tokyo, Japan), and Phosphatase Inhibitor Cocktails for serine/ threonine protein phosphatases and tyrosine protein phosphatases (P2850 and P5726; Sigma-Aldrich) and centrifuged to obtain lysates. The lysates (30␮g) were electrophoresed in a 10% SDS-polyacrylamide gel and trans-ferred onto a nylon membrane. The membrane was then incubated with Abs to TGF-␤1, Stat1, p-Stat1, Smad2, p-Smad2, Smad3, or Smad7 diluted at 1: 1,000. After the incubation of HRP-conjugated secondary Abs, the immune complexes were visualized using ECL® System (Amersham, Ja-pan) according to the manufacturer’s instructions.

Statistical analysis

The means and SEMs were calculated for all parameters determined in this study. Statistical significance was evaluated by using ANOVA or Mann-Whitney’s U test. p⬍ 0.05 was accepted as statistically significant.

Table I. Sequences of the primers used for RT-PCRa

Transcript Sequence Annealing Temperature (°C) Cycle Product Size (bp) IFN-␥ (F) 5⬘-AGCGGCTGACTGAACTCAGATTGTAG-3⬘ 60 30 247 (R) 5⬘-GTCACAGTTTTCAGCTGTATAGGG-3⬘ IL-12p35 (F) 5⬘-AACAAGAGGGAGCTGCCTGCC-3⬘ 60 36 300 (R) 5⬘-CGGGTGCTGAAGGCGTGAAGC-3⬘ IL-12p40 (F) 5⬘-CGTGCTCATGGCTGGTGCAAAG-3⬘ 55 36 576 (R) 5⬘-GAACACATGCCCACTTGCTG-3⬘ IL-18 (F) 5⬘-CGTGCTCATGGCTGGTGCAAAG-3⬘ 55 32 434 (R) 5⬘-GAACACATGCCCACTTGCTG-3⬘ VEGF (F) 5⬘-TGAACTTTCTGCTCTCTTGG-3⬘ 60 32 457 (R) 5⬘-AACAAATGCTTTCTCCGCTC-3⬘ COLIAI (F) 5⬘-GCCAAGAAGACATCCCTGAAG-3⬘ 60 34 138 (R) 5⬘-TCATTGCATTGCACGTCATC-3⬘ ␤-Actin (F) 5⬘-TTCTACAATGAGCTGCGTGTGGC-3⬘ 62 26 456 (R) 5⬘-CTCATAGCTCTTCTCCAGGGAGGA-3⬘

a_{(F), Forward primer; (R), reverse primer.}

1849 The Journal of Immunology

(4)

Results

IFN-␥, IL-12, and IL-18 expression during wound healing

In our initial experiments, we examined IFN-␥ expression at the skin excisional wound sites. In uninjured skin of WT mice, IFN-␥ mRNA was only weakly expressed (Fig. 1a). IFN-␥ mRNA ex-pression increased significantly 3 days after the injury and re-mained elevated until 6 days after the injury (Fig. 1b). Moreover, there were no significant differences in IFN-␥ gene expression be-tween WT and SCID mice, suggesting that non-lymphoid cells were a major cellular source of IFN-␥ at the skin wound sites (Fig. 1, a and b). Immunohistochemical analysis revealed that IFN-␥ protein was very low in wound sites of WT mice 1 day after the injury (Fig. 2a). In contrast, a large number of cells were positive for IFN-␥ at wound sites 3 and 6 days after injury (Fig. 2, b and c). Preadsorption of the Ab with an excess amount of IFN-␥ abolished the positive signals (Fig. 2d), indicating the specificity of the re-action. A double-color immunofluorescence analysis demonstrated that IFN-␥-positive cells were also positive for F4/80 at 3 days and 6 days after injury (Fig. 3a). IFN-␥-positive and F4/80-negative cells were judged as resident fibroblasts based on their morphol-ogy. At 6 days after the injury, a few CD3-positive cells were also

positive for IFN-␥ (Fig. 3b). Considering that IFN-␥ gene was expressed to a similar extent at the wound sites of WT and SCID mice, these observations suggest that non-lymphoid cells were a major cellular source of IFN-␥ in skin wound healing. We also analyzed IL-12 and IL-18 gene expression, which when combined can induce IFN-␥ production in macrophages (28–30). The ex-pression of both genes was enhanced to similar levels at skin wound sites in WT and SCID mice (Fig. 1, a and c– e). These results indicate that F4/80-positive macrophages and to a lesser degree, T cells, might produce IFN-␥ 3 days after injury, under the combined effects of IL-12 and IL-18.

FIGURE 3. A double-color immunofluorescence analysis of wound sites. Wound sites were obtained from WT mice at 3 (a) or 6 (b) days after the injury. The samples were immunostained with anti-F4/80 (a-i, Cy3), anti-CD3 (b-i, Cy3), or anti-IFN-␥ mAb (a-ii and b-ii, FITC) as described in Materials and Methods and observed under a fluorescence microscopy (original magnification,⫻100). Signals in i and ii were digitally merged in

panels iii. Representative results from three independent experiments are

shown.

FIGURE 1. The analysis of the gene expression of IFN-␥, IL-12p35, IL-12p40, and IL-18 at excisional skin wound sites of WT and SCID mice.

a, RT-PCR analysis for gene expression of these cytokines. RT-PCR was

performed as described in Materials and Methods and representative re-sults from six independent experiments are shown in a. Under the condi-tions used, mRNA of all cytokines was weakly detected in the uninjured skin. The ratios of IFN-␥ (b), IL-12p35 (c), IL-12p40 (d), and IL-18 (e) to

␤-actin at the wound sites of WT (䡺) and SCID mice (f) were determined

by RT-PCR at 1, 3, and 6 days after the injury. Each value represents mean⫾ SEM (n ⫽ 6). ⴱ, p ⬍ 0.05⬘ vs uninjured skin of BALB/c; #, p ⬍ 0.05⬘ vs uninjured skin of SCID.

FIGURE 2. Immunohistochemical analysis of IFN-␥ protein expression in skin wound sites. Skin wound samples were obtained from WT at days 1 (a), 3 (b and d), and 6 (c) after the wound preparation. Samples were immunostained with either untreated anti-IFN-␥ mAb (a–c) or that pread-sorbed with an excess amount of rIFN-␥ (d) as described in Materials and

Methods. Representative results from three independent experiments are

shown here. Original magnification,⫻100.

(5)

Macroscopic wound closure in IFN-␥ KO and WT mice

To evaluate the pathophysiological role of locally produced IFN-␥ in the wound healing process, we made excisional skin wounds in IFN-␥ KO and WT mice. In IFN-␥ KO mice, the wound areas were reduced to 40% at 3 days after injury. In contrast, wound areas in WT mice still remained at 50% even 6 days after injury (Fig. 4). Furthermore, the administration of a neutralizing anti-IFN-␥ mAb also increased the wound closure rates in WT mice (Fig. 5). These observations indicate that wound closure and sub-sequent wound healing were accelerated in the absence of IFN-␥.

Leukocyte infiltration at the wound sites in IFN-␥ KO and WT mice

We next examined the effects of IFN-␥ deficiency on leukocyte infiltration at the excisional wound sites. Consistent with our pre-vious observations, neutrophil infiltration in WT mice was maxi-mal at 1 day after injury. In contrast, macrophages and CD3-pos-itive cells started to accumulate at 1 day after injury and reached maximal levels at 6 days after injury. The infiltration of these cells was remarkably attenuated in IFN-␥ KO mice, compared with WT mice, at every time interval examined. Only macrophage infiltration 1 day after injury was similar in WT and IFN-␥ KO mice (Fig. 6).

Angiogenesis and VEGF gene expression at the wound sites in IFN-␥ KO and WT mice

We next examined the effects of IFN-␥ on the angiogenic process; one of the important events in the proliferative phase of wound healing (Fig. 7). No significant difference was observed in the ves-sel density of the uninjured skin when WT and IFN-␥ KO mice were compared (2.2 ⫾ 0.4% vs 2.7 ⫾ 0.4%) as measured by CD31-positive areas. Six days after the injury, the vessel density within the wound bed was increased in both the WT and IFN-␥ KO mice, and the vessel density of IFN-␥ KO mice was significantly higher than WT mice (Fig. 7, a– e). VEGF mRNA was weak but similar in uninjured skin of both WT and IFN-␥ KO mice. VEGF mRNA expression was enhanced at the wound sites in both mice 3 days after injury but the enhancement was significantly greater in IFN-␥ KO than WT mice (Fig. 7, f and g). These observations imply that the lack of IFN-␥ may augment angiogenesis in skin wound sites, partly by enhancing VEGF expression.

Granulation tissue formation at the wound sites in IFN-␥ KO and WT mice

We next explored the effects of IFN-␥ on collagen content in the extracellular matrix, another crucial molecule for the wound heal-ing process. Histopathologically, at 3 days and after, granulation tissue was evident at wound sites in WT mice (Fig. 8a) and the granulation tissue formation was more prominent at the wound sites in IFN-␥ KO mice (Fig. 8b). In uninjured skin, there was no significant difference in terms of HP content and COL1A1 mRNA expression between WT and IFN-␥ KO mice (Fig. 8, c–e). In WT mice, HP content and COL1A1 mRNA expression at the wound sites started to increase progressively 3 days after injury. However, the increases in HP content and COL1A1 mRNA expression at the wound sites were consistently and significantly higher in IFN-␥ KO mice than WT mice (Fig. 8, c– e). These observations indicate that the absence of IFN-␥ augmented collagen gene expression and eventually collagen production at the wound sites.

The effects of IFN-␥ deficiency on the TGF-␤1-mediated signaling pathway

As TGF-␤1 has been considered to be a major regulator of colla-gen biosynthesis (3– 6), we examined the changes in the TGF-␤1 signaling pathway at the wound sites (Fig. 9). The amounts of both total and phosphorylated Stat1 were increased at the wound sites of WT mice 1 day after injury. In contrast, the amount of total and phosphorylated Stat1 was not significantly changed at the wound sites of IFN-␥ KO mice, due to the absence of IFN-␥-mediated signals. The amount of TGF-␤1 protein was increased at the wound sites although the increase was more marked in IFN-␥ KO mice than WT mice. Moreover, although total Smad7 levels were increased in the wound sites of WT mice, a corresponding increase in Smad7 levels was not observed in IFN-␥ KO mice. Although wound injury increased the amount of Smad3 to similar extents in the wound sites of both WT and IFN-␥ KO mice, the amounts of total and phosphorylated Smad2 remained at similar levels at the wound sites in WT mice. In contrast, the amount of total and phosphorylated Smad2 was strongly increased in the wound sites of IFN-␥ KO mice. A double color immunofluorescence analysis demonstrated that phosphorylated Smad2 was detected mainly in

FIGURE 5. Macroscopic appearance of wound healing process in WT mice administered with a control or an anti-IFN-␥ mAb. a, The wound sites were photographed at the time indicated. Day 0 picture was taken imme-diately after the injury. Representative results from 12 individual animals in each group are shown. b, Changes in percentage of wound area at each time point in comparison to the original wound area in WT mice admin-istered with a control or an anti-IFN-␥ mAb. Values represent mean ⫾ SEM.䡺, WT treated with control Ig G (IgG); f, WT treated with anti-IFN-␥ mAb (n ⫽ 12 animals). ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01, anti-IFN-␥ mAb compared with control.

FIGURE 4. Macroscopic changes in skin excisional wound sites. a, The wound sites were photographed at the time indicated. Day 0 picture was taken immediately after the injury. Representative results from 12 individ-ual animals in each group are shown here. b, Changes in percentage of wound area at each time point in comparison to the original wound area. Values represent mean⫾ SEM. 䡺, WT; f, IFN-␥ KO (n ⫽ 12 animals).

ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01, IFN-␥ KO compared with WT.

(6)

␣-SMA-positive fibroblasts, which are presumed to express IFN-␥

receptors (Fig. 10). Thus, IFN-␥ may negatively regulate TGF-␤1 signaling pathway by down-regulating the expression of TGF-␤1 and its downstream intracellular molecules at the wound sites.

Discussion

IFN-␥ has pleiotropic actions on various types of immune cells and has been implicated as one of main regulatory factors for CD4⫹

Th1 polarization (7). In addition to its immunoregulatory actions, IFN-␥ exerts multiple effects on non-immune cells, particularly fibroblasts. Several lines of evidence have demonstrated that IFN-␥ can inhibit collagen synthesis by fibroblasts (8–12) and accumulating evidence suggests that administration of exogenous IFN-␥ impairs skin wound healing (13–16). In this report, we ex-amined the roles of endogenous IFN-␥ in wound healing process, by using IFN-␥ KO mice. We found that the lack of endogenous

FIGURE 7. a– d, Immunohistochemical analyses on excisional skin wound sites of WT (a and c) and IFN-␥ KO mice (b and d) at 6 days after injury. The

sections were stained with a mAb for the endothelium (CD31) (a and b,⫻10; c and d, ⫻100). Representative results from six independent animals in each group are shown. e, Vascular areas were determined as CD31-positive areas in IFN-␥ KO (f) and WT mice (䡺) mice with the help of PhotoShop. All values represent the mean⫾ SEM (n ⫽ 6 animals). ⴱ, p ⬍ 0.05, IFN-␥ KO compared with WT. f and g, RT-PCR analysis of VEGF gene expression at wound sites in WT and IFN-␥ KO mice. Representative results from 10 independent animals are shown in f. Under the conditions used, VEGF mRNA was faintly detected in uninjured skin samples of WT and IFN-␥ KO mice. The ratios of VEGF to ␤-actin of WT (䡺) and IFN-␥ KO mice (f) were determined by RT-PCR and are shown in

g. Each value represents mean⫾ SEM (n ⫽ 10 animals). ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01, IFN-␥ KO compared with WT.

FIGURE 6. Immunohistochemical analyses on leukocyte recruitment in skin excisional wound sites. a–f, Immunohistochemical analysis was performed using anti-MPO at day 1 (a and d), anti-F4/80 at day 6 (b and e), or anti-CD3 Abs at day 6 (c and f) in skin wound samples from WT (a– c) and IFN-␥ KO (d–f) mice (⫻200). Representative results from three independent experiments are shown here. g, MPO activity at the wound site of IFN-␥ KO (f) and WT (䡺) was determined to evaluate neutrophil accumulation. The numbers of macrophages (h) or those of T cells (i) per a high-power microscopic field (original magnification,⫻200) were counted. All values represent the mean ⫾ SEM (n ⫽ 6 animals). ⴱ, p ⬍ 0.05, IFN-␥ KO compared with WT.

(7)

IFN-␥ significantly accelerated wound healing process as evi-denced by rapid wound closure and enhanced granulation tissue formation.

In WT mice, we observed enhanced IFN-␥ mRNA and protein expression at the skin excisional wound sites. A double-color im-munofluorescence analysis detected IFN-␥ protein in F4/80-posi-tive macrophages recruited to the wound sites. Several lines of evidence have demonstrated that the combined stimulation of IL-12 and IL-18 induces bioactive IFN-␥ protein in macrophages (28 –30). Consistent with this observation, IL-12 and IL-18 gene expression was cooperatively enhanced at the wound sites of WT mice after injury. Moreover, in SCID mice, IFN-␥ protein could also be detected in F4/80-positive macrophages at the wound sites with a concomitant enhancement of IL-12 and IL-18 gene expres-sion. These observations indicate that, in addition to T cells, mac-rophages are a cellular source of IFN-␥ during skin wound healing. Immediately after skin wounding, neutrophils infiltrate to the wound site, followed by macrophages. The infiltration of these inflammatory cells is regulated by coordinate expression of che-mokines and ICAM-1. IFN-␥ can up-regulate ICAM-1 expression, which is important for cell adhesion (7). In line with this obser-vation, we previously reported that leukocyte infiltration was markedly attenuated in the liver of IFN-␥ KO mice treated with acetaminophen, with a concomitant reduction in chemokine and ICAM-1 gene expression, compared with WT mice (23). Also in this skin wound model, leukocyte infiltration was remarkably at-tenuated in IFN-␥-deficient mice, with a concomitant reduction in chemokine and ICAM-1 gene expression (data not shown), result-ing in reduced neutrophil infiltration durresult-ing skin wound healresult-ing (31). In contrast, cardiac allografts in IFN-␥ KO recipient mice exhibited a massive neutrophil infiltration with accelerated tissue

necrosis, compared with allografts in WT recipients (32). How-ever, CD8⫹ lymphocytes control the cardiac allograft process, whereas T lymphocytes have little, if any, role in the skin wound healing process, as evidenced by no apparent morphological dif-ferences between WT and SCID mice (data not shown). Moreover, reduced CD8⫹lymphocyte infiltration induced aberrant chemo-kine gene expression and eventually augmented neutrophil infil-tration in cardiac allograft in IFN-␥ KO mice. Thus, in a specific context, IFN-␥ may have seemingly contradictory effects on neu-trophil infiltration.

Macrophages, which infiltrate into the wound sites, have been presumed to promote wound healing by producing various types of bioactive substances (33, 34). However, several recent reports raised questions regarding the validity of this hypothesis. Secretory leukocyte protease inhibitor-deficient mice exhibited impaired wound healing despite or because of exaggerated leukocyte infil-tration (35). Moreover, skin wound healing was accelerated de-spite reduced leukocyte infiltration in mice deficient in the TNF receptor p55 (25). In line with the latter observations, IFN-␥ KO mice exhibited accelerated wound healing with a concomitant re-duction in leukocyte infiltration. Moreover, two indispensable steps for wound healing, angiogenesis and collagen deposition, were enhanced at the wound sites of IFN-␥ KO mice, as similarly observed in TNF receptor p55 KO mice (25). Thus, under the specific pathogen free conditions, angiogenesis and collagen dep-osition can proceed in wound sites, independent of leukocyte in-filtration in this skin excisional wound model.

Accumulating evidence indicates that IFN-␥ has a negative ef-fect on collagen deposition, one of the most crucial events for wound healing (13–16), although the precise molecular mecha-nisms involved in this inhibition remain elusive. We also observed that the absence of IFN-␥ resulted in enhanced collagen deposition in wound sites as evidenced by increased COL1A1 mRNA expres-sion and HP contents. As TGF-␤1 has been implicated as a key mediator of collagen synthesis, we examined the TGF-␤1 signal-ing pathway in the wound sites of IFN-␥ KO mice. We observed that mature TGF-␤1 protein was significantly increased in the

FIGURE 8. a and b, Histopathological analyses on skin wound sites of

WT (a) and IFN-␥ KO (b) mice at 6 days after injury. Granulation tissue formation was more evident in IFN-␥ KO mice than in WT mice. c, HP contents in the excisional wound sites in WT (䡺) and IFN-␥ KO (f) mice. HP contents were determined as an indicator of collagen contents. All values represent the mean⫾ SEM (n ⫽ 6 animals). ⴱ, p ⬍ 0.05, IFN-␥ KO compared with WT. d and e, RT-PCR analysis of collagen gene expression in the wound sites in WT and IFN-␥ KO mice. Under the conditions used, RT-PCR analysis did not detect the mRNA of COL1A1 in uninjured skin samples of WT and IFN-␥ KO mice. Representative results from six ani-mals in each group are shown in d. The ratios of COL1A1 to␤-actin of WT (䡺) and IFN-␥ KO (f) were determined by RT-PCR at 1, 3 and 6 days after injury (e). Each value represents mean⫾ SEM (n ⫽ 6 animals). ⴱ,

p⬍ 0.05; ⴱⴱ, p ⬍ 0.01, IFN-␥ KO compared with WT.

FIGURE 9. Western blotting analysis on the expression of TGF-␤, Stat1, phosphorylated Stat1, Smad2, phosphorylated Smad2, Smad3, and Smad7, at the wound sites. Under the conditions used, these molecules were faintly detected in uninjured skin sites of WT and IFN-␥ KO mice. Western blotting analysis using anti-␣-tubulin Ab confirmed that an equal amount of protein was loaded onto each lane. Representative results from six individual animals in each group are shown.

(8)

wound sites of IFN-␥ KO mice, compared with WT mice, consis-tent with the previous in vitro observations that IFN-␥ inhibited TGF-␤1 protein synthesis at a posttranslational level (36, 37).

TGF-␤1 mediates its signals mainly by phosphorylating stimu-latory Smads, Smad2 and 3, whereas another Smad, Smad7, an-tagonizes its signaling pathways (38 – 44). This intracellular sig-naling machinery also plays a role in fibrotic changes in bleomycin-induced pulmonary fibrosis as in vivo gene transfer of Smad7 reduced collagen expression at the mRNA and protein lev-els by reducing the phosphorylation of Smad2 and eventually at-tenuated pulmonary fibrosis (45). By using different types of cell lines, independent groups reported that in vitro, IFN-␥/Stat1 sig-nals can increase the amount of an inhibitory Smad, Smad7 and prevent the phosphorylation of Smad2 and 3, thereby inhibiting the actions of TGF-␤1 (21, 22). We observed increases in the amount of total and phosphorylated Smad2 and a reciprocal decrease in the amount of an inhibitory Smad, Smad7, at the wound sites in IFN-␥ KO mice, compared with WT mice, with a concomitant reduction in the amount of total and phosphorylated Stat1 (Fig. 9). Moreover, immunohistochemical analyses detected IFN-␥ protein in various types of cells including macrophages and fibroblasts, whereas phosphorylated Smad2 was detected predominantly in fibroblasts. Thus, crosstalk between IFN-␥/Stat1 and TGF-␤1/Smad systems appears to operate in the skin wound healing processes in an au-tocrine and/or paracrine manner.

Enhanced TGF-␤1 production may account for augmented TGF-␤1 signaling in the skin wound site of IFN-␥ KO mice. Al-though it has been reported that TGF-␤1 can rapidly and massively induce Smad7 in several types of cells (42), in the wound sites of IFN-␥ KO mice, Smad7 protein levels were not significantly in-creased despite inin-creased levels of TGF-␤1. Thus, it is more likely that TGF-␤1-mediated signaling pathways were mainly aug-mented by the absence of IFN-␥/Stat1 signaling but not increased TGF-␤1 production.

Angiogenesis is another indispensable event for granulation tis-sue formation and subsequent wound healing. IFN-␥ can inhibit capillary growth and development in vitro (46, 47) and can induce the expression of a chemokine, IP-10, which exhibits potent anti-angiogenic activity (48). Thus, angiogenesis may be augmented by the absence of a negative regulator, i.e., IFN-␥. Although the ef-fects of IFN-␥ on the expression of a master regulator of angio-genesis, VEGF, are still controversial (49, 50), several lines of evidence suggest that IFN-␥ inhibits VEGF expression (51), con-sistent with our present observations. Furthermore, TGF-␤1 can augment VEGF transcription in various cell types (52–56). Thus, the enhancement in TGF-␤1 expression and its signaling path-ways, may be responsible for enhanced VEGF expression and sub-sequent enhanced angiogenesis in IFN-␥ KO mice.

Our present observations suggest that the absence of IFN-␥ may augment the expression and phosphorylation of a stimulatory Smad, Smad2, and thus accelerate excisional skin wound healing. However, mice deficient in another stimulatory Smad, Smad3, ex-hibited enhanced re-epithelialization, and eventually accelerated incisional skin wound healing (57). In the healing process of in-cisional wounds, re-epithelialization is presumed to be the most crucial phenomenon. As TGF-␤-mediated signals inhibit re-epi-thelialization, the lack of Smad3 might accelerate incisional skin wound healing (57). In contrast, collagen deposition might have a more important role in the healing process of an excisional skin wound. As collagen deposition was markedly attenuated through a reduction in Smad2 phosphorylation induced by Smad7 in bleo-mycin-induced pulmonary fibrosis in mice, Smad2 may be more important with respect to collagen deposition (45). This hypothesis is supported by our present observations that the amount of phos-phorylated Smad2 and total Smad2 was significantly increased in IFN-␥ KO mice compared with WT mice, despite a marginal dif-ference in total Smad3 amount.

Our observations suggest that IFN-␥ can negatively modulate the wound healing process by suppressing the production and functional activity of TGF-␤1. As TGF-␤1 can inhibit IFN-␥ pro-duction and its receptor expression, both cytokines can antagonize one another. Thus, the blockade of the IFN-␥ signal transduction pathway may enhance TGF-␤1 production and TGF-␤1 signaling in a positive feedback manner and may be an important strategy to accelerate the healing process of skin wounds.

Acknowledgments

We express our sincere gratitude to Dr. Howard A. Young (National Can-cer Institute, Frederick, MD) for his invaluable comments on the manu-script. We thank Ryoichi Mori for his technical assistance with the deter-mination of HP content, and we are grateful to Dr. Yasuhiko Yamamoto (Department of Biochemistry and Molecular Vascular Biology, Kanazawa University) for his instructive advice about Western blotting.

References

1. Singer, A. J., and R. A. Clark. 1999. Cutaneous wound healing. N. Engl. J. Med.

341:738.

2. Martin, P. 1997. Wound healing-aiming for perfect skin regeneration. Science.

276:75.

3. Sporn, M. B., and A. B. Roberts. 1993. A major advance in the use of growth factors to enhance wound healing. J. Clin. Invest. 92:2565.

4. Quaglino, D., Jr., L. B. Nanney, J. A. Ditesheim, and J. M. Davidson. 1991. Transforming growth factor-␤ stimulates wound healing and modulates extracel-lular matrix gene expression in pig skin: incisional wound model. J. Invest.

Der-matol. 97:34.

5. Mustoe, T. A., G. F. Pierce, A. Thomason, P. Gramates, M. B. Sporn, and T. F. Deuel. 1987. Accelerated healing of incisional wounds in rats induced by transforming growth factor-␤. Science 237:1333.

6. Beck, L. S., L. DeGuzman, W. P. Lee, Y. Xu, M. W. Siegel, and E. P. Amento. 1993. One systemic administration of transforming growth factor-␤1 reverses age- or glucocorticoid-impaired wound healing. J. Clin. Invest. 92:2841. FIGURE 10. A double-color immunofluorescence analysis of phosphorylated Smad2-positive cells in skin excisional wound site at 6 days after the injury in WT mice. A double-color immunofluorescence analysis was performed using anti-␣-SMA (i, Cy3) and anti-pSmad2 Abs (ii, FITC) and observed under a fluorescence microscope (⫻400). Signals were digitally merged in panel iii (derived from i and ii). Representative results from six individual animals are shown.

(9)

7. Farrar, M. A., and R. D. Schreiber. 1993. The molecular cell biology of inter-feron-gamma and its receptor. Annu. Rev. Immunol. 11:571.

8. Amento, E. P., A. K. Bhan, K. G. McCullagh, and S. M. Krane. 1985. Influences of gamma interferon on synovial fibroblast-like cells. Ia induction and inhibition of collagen synthesis. J. Clin. Invest. 76:837.

9. Duncan, M. R., and B. Berman. 1985. Gamma interferon is the lymphokine and beta interferon the monokine responsible for inhibition of fibroblast collagen production and late but not early fibroblast proliferation. J. Exp. Med. 162:516. 10. Granstein, R. D., T. J. Flotte, and E. P. Amento. 1990. Interferons and collagen

production. J. Invest. Dermatol. 95(Suppl):75S.

11. Harrop, A. R., A. Ghahary, P. G. Scott, N. Forsyth, A. Uji-Friedland, and E. E. Tredget. 1995. Regulation of collagen synthesis and mRNA expression in normal and hypertrophic scar fibroblasts in vitro by interferon-gamma. J. Surg.

Res. 58:471.

12. Yufit, T., V. Vining, L. Wang, R. R. Brown, and J. Varga. 1995. Inhibition of type I collagen mRNA expression independent of tryptophan depletion in inter-feron-␥-treated human dermal fibroblasts. J. Invest. Dermatol. 105:388. 13. Granstein, R. D., M. R. Deak, S. L. Jacques, R. J. Margolis, T. J. Flotte,

D. Whitaker, F. H. Long, and E. P. Amento. 1989. The systemic administration of gamma interferon inhibits collagen synthesis and acute inflammation in a murine skin wounding model. J. Invest. Dermatol. 93:18.

14. Laato, M., J. Heino, B. Gerdin, V. M. Kahari, and J. Niinikoski. 2001. Interferon-gamma-induced inhibition of wound healing in vivo and in vitro. Ann. Chir.

Gynaecol. 90(Suppl). 215:19.

15. Cornelissen, A. M., J. C. Maltha, J. W. Von den Hoff, and A. M. Kuijpers-Jagtman. 2000. Local injection of IFN-gamma reduces the number of myofibroblasts and the collagen content in palatal wounds. J. Dent. Res. 79:1782.

16. Miles, R. H., T. P. Paxton, D. Zacheis, D. J. Dries, and R. L. Gamelli. 1994. Systemic administration of interferon-gamma impairs wound healing. J. Surg.

Res. 56:288.

17. Darnell, J. E. Jr. 1997. STATs and gene regulation. Science. 277:1630. 18. Stark, G. R., I. M. Kerr, B. R. Williams, R. H. Silverman, and R. D. Schreiber.

1998. How cells respond to interferons. Annu. Rev. Biochem. 67:227. 19. Meraz, M. A., J. M. White, K. C. Sheehan, E. A. Bach, S. J. Rodig, A. S. Dighe,

D. H. Kaplan, J. K. Riley, A. C. Greenlund, D. Campbell, K. Carver-Moore, R. N. DuBois, R. Clark, M. Aguet, and R. D. Schreiber. 1996. Targeted disrup-tion of the Stat1 gene in mice reveals unexpected physiologic specificity in the JAK-STAT signaling pathway. Cell. 84:431.

20. Durbin, J. E., R. Hackenmiller, M. C. Simon, and D. E. Levy. 1996. Targeted disruption of the mouse Stat1 gene results in compromised innate immunity to viral disease. Cell. 84:443.

21. Ulloa, L., J. Doody, and J. Massague. 1999. Inhibition of transforming growth factor-beta/SMAD signalling by the interferon-gamma/STAT pathway. Nature.

397:710.

22. Ghosh, A. K., W. Yuan, Y. Mori, S. J. Chen, and J. Varga. 2001. Antagonistic regulation of type I collagen gene expression by interferon-gamma and trans-forming growth factor-beta. Integration at the level of p300/CBP transcriptional coactivators. J. Biol. Chem. 276:11041.

23. Ishida, Y., T. Kondo, T. Ohshima, H. Fujiwara, Y. Iwakura, and N. Mukaida. 2002. A pivotal involvement of IFN-gamma in the pathogenesis of acetamino-phen-induced acute liver injury. FASEB J. 16:1227.

24. Tsuji, H., N. Mukaida, A. Harada, S. Kaneko, E. Matsushita, Y. Nakanuma, H. Tsutsui, H. Okamura, K. Nakanishi, Y. Tagawa, Y. Iwakura, K. Kobayashi, and K. Matsushima. 1999. Alleviation of lipopolysaccharide-induced acute liver injury in Propionibacterium acnes-primed IFN-gamma-deficient mice by a con-comitant reduction of TNF-alpha, IL-12, and IL-18 production. J. Immunol.

162:1049.

25. Mori, R., T. Kondo, T. Ohshima, Y. Ishida, and N. Mukaida. 2002. Accelerated wound healing in tumor necrosis factor receptor p55-deficient mice with reduced leukocyte infiltration. FASEB J. 16:963.

26. Ozawa, K., T. Kondo, O. Hori, K. Kitao, D. M. Stern, W. Eisenmenger, S. Ogawa, and T. Ohshima. 2001. Expression of the oxygen-regulated protein ORP150 accelerates wound healing by modulating intracellular VEGF transport.

J. Clin. Invest. 108:41.

27. Swift, M. E., H. K. Kleinman, and L. A. DiPietro. 1999. Impaired wound repair and delayed angiogenesis in aged mice. Lab Invest. 79:1479.

28. Schindler, H., M. B. Lutz, M. Ro¨llinghoff, and C. Bogdan. 2001. The production of IFN-␥ by IL-12/IL-18-activated macrophages requires STAT4 signaling and is inhibited by IL-4. J. Immunol. 166:3075.

29. Puddu, P., L. Fantuzzi, P. Borghi, B. Varano, G. Rainaldi, E. Guillemard, W. Malorni, P. Nicaise, S. F. Wolf, F. Belardelli, and S. Gessani. 1997. IL-12 induces IFN-␥ expression and secretion in mouse peritoneal macrophages. J.

Im-munol. 159:3490.

30. Munder, M., M. Mallo, K. Eichmann, and M. Modolell. 1998. Murine macro-phages secrete interferon␥amma upon combined stimulation with interleukin (IL)-12 and IL-18: a novel pathway of autocrine macrophage activation. J. Exp.

Med. 187:2103.

31. Nagaoka, T., Y. Kaburagi, Y. Hamaguchi, M. Hasegawa, K. Takehara, D. A. Steeber, T. F. Tedder, and S. Sato. 2000. Delayed wound healing in the absence of intercellular adhesion molecule-1 or L-selectin expression.

Am. J. Pathol. 157:237.

32. Miura, M., T. El-Sawy, and R. L. Fairchild. 2003. Neutrophils mediate paren-chymal tissue necrosis and accelerate the rejection of complete major histocom-patibility complex-disparate cardiac allografts in the absence of interferon-gamma. Am. J. Pathol. 162:509.

33. Leibovich, S. J., and R. Ross. 1975. The role of the macrophage in wound repair. A study with hydrocortisone and antimacrophage serum. Am. J. Pathol. 78:71. 34. Danon, D., M. A. Kowatch, and G. S. Roth. 1989. Promotion of wound repair in

old mice by local injection of macrophages. Proc. Natl. Acad. Sci. USA. 86:2018. 35. Ashcroft, G. S., K. Lei, W. Jin, G. Longenecker, A. B. Kulkarni, T. Greenwell-Wild, H. Hale-Donze, G. McGrady, X. Y. Song, and S. M. Wahl. 2000. Secretory leukocyte protease inhibitor mediates non-redundant functions necessary for normal wound healing. Nat. Med. 6:1147.

36. Tredget, E. E., R. Wang, Q. Shen, P. G. Scott, and A. Ghahary,. 2000. Trans-forming growth factor-beta mRNA and protein in hypertrophic scar tissues and fibroblasts: antagonism by IFN-alpha and IFN-gamma in vitro and in vivo. J.

In-terferon Cytokine Res. 20:143.

37. Gurujeyalakshmi, G. and S. N. Giri. 1995. Molecular mechanisms of antifibrotic effect of interferon gamma in bleomycin-mouse model of lung fibrosis: down-regulation of TGF-beta and procollagen I and III gene expression. Exp. Lung.

Res. 21:791.

38. Massague, J. 1997. TGF-beta signal transduction. Annu. Rev. Biochem. 67:753. 39. Heldin, C. H., K. Miyazono, and P. ten Dijke. 1997. TGF-beta signalling from

cell membrane to nucleus through SMAD proteins. Nature. 390:465. 40. Nakao, A., T. Imamura, S. Souchelnytskyi, M. Kawabata, A. Ishisaki, E. Oeda,

K. Tamaki, J. Hanai, C. H. Heldin, K. Miyazono, and P. ten Dijke. 1997. TGF-beta receptor-mediated signalling through Smad2, Smad3 and Smad4. EMBO J.

16:5353.

41. Macias-Silva, M., S. Abdollah, P. A. Hoodless, R. Pirone, L. Attisano, and J. L. Wrana. 1996. MADR2 is a substrate of the TGFbeta receptor and its phos-phorylation is required for nuclear accumulation and signaling. Cell. 87:1215. 42. Nakao, A., M. Afrakhte, A. Moren, T. Nakayama, J. L. Christian, R. Heuchel,

S. Itoh, M. Kawabata, N. E. Heldin, C. H. Heldin, and P. ten Dijke. 1997. Iden-tification of Smad7, a TGFbeta-inducible antagonist of TGF-beta signalling.

Na-ture. 389:631.

43. Hayashi, H., S. Abdollah, Y. Qiu, J. Cai, Y. Y. Xu, B. W. Grinnell, M. A. Richardson, J. N. Topper, M. A. Gimbrone, J. L. Wrana, and D. Falb. 1997. The MAD-related protein Smad7 associates with the TGFbeta receptor and func-tions as an antagonist of TGFbeta signaling. Cell. 897:1165.

44. Topper, J. N., J. Cai, Y. Qiu, K. R. Anderson, Y. Y. Xu, J. D. Deeds, R. Feeley, C. J. Gimeno, E. A. Woolf, O. Tayber, G. G. Mays, B. A. Sampson, F. J. Schoen, M. A. Gimbrone, and D. Falb. 1997. Vascular MADs: two novel MAD-related genes selectively inducible by flow in human vascular endothelium. Proc. Natl.

Acad. Sci. USA. 94:9314.

45. Nakao, A., M. Fujii, R. Matsumura, K. Kumano, Y. Saito, K. Miyazono, and I. Iwamoto. 1999. Transient gene transfer and expression of Smad7 prevents bleomycin-induced lung fibrosis in mice. J. Clin. Invest. 104:5.

46. Sato, N., H. Nariuchi, N. Tsuruoka, T. Nishihara, J. G. Beitz, P. Calabresi, and A. R. Frackelton. Jr. 1990. Actions of TNF and IFN-gamma on angiogenesis in vitro. J. Invest. Dermatol. 95(6 Suppl):85S.

47. Maheshwari, R. K., V. Srikantan, D. Bhartiya, H. K. Kleinman, and D. S. Grant. 1991. Differential effects of interferon gamma and alpha on in vitro model of angiogenesis. J. Cell. Physiol. 146:164.

48. Angiolillo, A. L., C. Sgadari, D. D. Taub, F. Liao, J. M. Farber, S. Maheshwari, H. K. Kleinman, G. H. Reaman, and G. Tosato. 1995. Human interferon-induc-ible protein 10 is a potent inhibitor of angiogenesis in vivo. J. Exp. Med. 182:155. 49. Frank, S., B. Stallmeyer, H. Kampfer, N. Kolb, and J. Pfeilschifter. 1999. Nitric oxide triggers enhanced induction of vascular endothelial growth factor expres-sion in cultured keratinocytes (HaCaT) and during cutaneous wound repair.

FASEB J. 13:2002.

50. Bo¨lling, B., J. Fandrey, P. J. Frosch, and H. Acker. 2000. VEGF production, cell proliferation and apoptosis of human IGR 1 melanoma cells under nIFN-alpha/ beta and rIFN-gamma treatment. Exp. Dermatol. 9:327.

51. Kawano, Y., N. Matsui, S. Kamihigashi, H. Narahara, and I. Miyakawa. 2000. Effects of interferon-gamma on secretion of vascular endothelial growth factor by endometrial stromal cells. Am. J. Reprod. Immunol. 43:47.

52. Pertovaara, L., A. Kaipainen, T. Mustonen, A. Orpana, N. Ferrara, O. Saksela, and K. Alitalo. 1994. Vascular endothelial growth factor is induced in response to transforming growth factor-␤ in fibroblastic and epithelial cells. J. Biol. Chem.

269:6271.

53. Benckert, C., S. Jonas, T. Cramer, Z. Von Marschall, G. Schafer, M. Peters, K. Wagner, C. Radke, B. Wiedenmann, P. Neuhaus, M. Hocker, and S. Rosewicz. 2003. Transforming growth factor beta 1 stimulates vascular endothelial growth factor gene transcription in human cholangiocellular carcinoma cells. Cancer

Res. 63:1083.

54. Gary Lee, Y. C., D. Melkerneker, P. J. Thompson, R. W. Light, and K. B. Lane. 2002. Transforming growth factor beta induces vascular endothelial growth fac-tor elaboration from pleural mesothelial cells in vivo and in vitro. Am. J. Respir.

Crit. Care. Med. 165:88.

55. Yamamoto. T, O. Kozawa, K. Tanabe, S. Akamatsu, H. Matsuno, S. Dohi, and T. Uematsu. 1997. Involvement of p38 MAP kinase in TGF-beta-stimulated VEGF synthesis in aortic smooth muscle cells. J. Cell. Biochem. 82:591. 56. Pintavorn, P. and B. J. Ballermann. 1997. TGF-beta and the endothelium during

immune injury. Kidney Int. 51:1401.

57. Ashcroft, G. S., X. Yang, A. B. Glick, M. Weinstein, J. L. Letterio, D. E. Mizel, M. Anzano, T. Greenwell-Wild, S. M. Wahl, C. Deng, and A. B. Roberts. 1999. Mice lacking Smad3 show accelerated wound healing and an impaired local inflammatory response. Nat. Cell. Biol. 1:260.