Transgenic mice reveal novel activities of growth hormone · 8 Thorey et al. Grove, PA, USA and Molecular Probes, Eugene, OR). Cell nuclei were stained with DAPI (Sigma, Munich, Germany)

1 Thorey et al.

Transgenic mice reveal novel activities of growth hormone

in wound repair, angiogenesis and myofibroblast

differentiation

Irmgard S. Thorey1,2, Boris Hinz3, Andreas Hoeflich2, Susanne

Kaesler1, Philippe Bugnon1, Martin Elmlinger4, Rüdiger Wanke5,

Eckhard Wolf2 and Sabine Werner1

1Institute of Cell Biology, Department of Biology, ETH Zürich, Hönggerberg,

CH-8093 Zürich, Switzerland 2Institute of Molecular Animal Breeding/Gene Center, Ludwig-Maximilians-

University, Feodor-Lynen-Str. 25, D-81377 Munich, Germany 3Laboratory of Cell Biophysics, Swiss Federal Institute of Technology (EPFL),

Bâtiment SG - AA-B 143, CH-1015 Lausanne, Switzerland 4Pediatric Endocrinology, Children’s Hospital, University of Tübingen,

D-72076 Tübingen, Germany 5Institute of Veterinary Pathology, Ludwig-Maximilians-University, Munich,

Germany

Address for correspondence:

Dr. Sabine Werner Institute of Cell Biology

ETH Zürich, Hönggerberg, CH-8093 Zürich, Switzerland

Phone: 41-1-633-3941 Fax: 41-1-633-1174

E-mail: [email protected]

Running title: Growth hormone and wound healing

JBC Papers in Press. Published on April 7, 2004 as Manuscript M311467200

Copyright 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on August 17, 2020

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

2 Thorey et al.

Abstract

An increasing number of patients is being treated with growth hormone (GH)

for the enhancement of body growth, but also as an anti-aging strategy.

However, the side effects of GH have been poorly defined. In this study we

determined the effect of GH on wound repair and its mechanisms of action at

the wound site. For this purpose, we performed wound healing studies in

transgenic mice overexpressing GH. Full-thickness incisional and excisional

wounds of transgenic animals developed an extensive, highly vascularized

granulation tissue. However, wound bursting strength was not increased.

Wound closure was strongly delayed as a result of enhanced granulation

tissue formation and impaired wound contraction. The latter effect is most

likely due to a significantly reduced number of myofibroblasts at the wound

site. Using in vitro studies with stressed collagen lattices, we identified GH as

an inhibitor of transforming growth factor β-induced myofibroblast

differentiation, resulting in a reduction in fibroblast contractile activity. These

results reveal novel roles of growth hormone in angiogenesis and

myofibroblast differentiation, which are most likely not mediated via insulin-

like growth factors at the wound site. Furthermore, our data suggest that

systemic GH treatment is detrimental for wound healing in non-healing

impaired individuals.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

3 Thorey et al.

Introduction

Injury to the skin initiates a series of events, which finally lead to at least

partial reconstruction of the wounded tissue (1). The important role of locally

acting cytokines and growth factors in wound repair is well documented (2).

However, crucial functions of endocrine acting hormones have only recently

been recognized (3-5). One of the hormones, which are thought to influence

the repair process, is growth hormone (GH). This hormone is successfully

used for the treatment of growth defects in children. In addition, an increasing

number of healthy adult patients receive GH therapy as an anti-aging strategy

(6). However, the side effects of GH treatment have been poorly defined and

it is as yet unclear if GH affects the wound healing process in healthy

individuals.

Treatment with recombinant GH has been used in patients with severe burn

injuries, and in most studies enhanced rates of wound healing and patient

survival were observed (7). In addition, GH was shown to stimulate

granulation tissue formation and biomechanical wound strength in animal

models of impaired healing (8-11), although a lack of efficacy of GH has also

been reported (12, 13). Thus, the role of GH in wound repair remains to be

determined. In particular, the mechanisms of GH action in this process have

not yet been elucidated.

Many effects of GH are mediated via the insulin-like growth factor (IGF)

system (14), suggesting that this might also be a mechanism of GH action in

the skin. IGF-I is a mitogen for keratinocytes (15), and it stimulates collagen,

glycosaminoglycane and proteoglycane synthesis by dermal fibroblasts (16,

17). It exerts its effects in concert with a family of six IGF-binding proteins

(IGFBPs) which control the availability of IGF for signaling through its receptor

in a complex manner: While IGFBP-3 and IGFBP-4 have a predominantly

negative effect on IGF activity, IGFBP-5 can act to enhance IGF signaling, but

only when associated with the surface extracellular matrix of target cells (16,


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

4 Thorey et al.

18). In addition, IGFBPs have also IGF-independent effects on cell growth

(19).

Interestingly, imbalances in the IGF/IGFBP system are associated with wound

healing abnormalities. Thus, up-regulation of IGF or IGFBPs was detected in

hypertrophic scars, while decreased IGF levels were found in diabetic skin

and foot ulcers, and in the wounds of healing-impaired diabetic or

glucocorticoid-treated rats or mice (2). Based on these findings, the effect of

members of the IGF system on dermal tissue repair has been analyzed: Local

IGF gene therapy or protein treatment improved the wound healing process in

different animal models (20-23). In contrast, raising the serum levels of IGF-I

by systemic application did not improve wound healing, whereas GH was

effective in the same experimental set-up, possibly via induction of IGF-I at

the wound site (24).

These findings, together with the obvious importance of the GH/IGF system in

the homeostasis of the skin, prompted us to analyze the wound healing

process in transgenic mice overexpressing bovine GH (bGH) under a

phosphoenolpyruvate carboxykinase (PEPCK) promoter. These mice have

greatly elevated postnatal serum GH levels of ca. 2000 ng/ml (compared to

less than 10 ng/ml endogenous hormone in control mice), and serum IGF-I

levels which are 2-3 times higher than those of control littermates. In contrast

to growth hormone treatment models, these animals allow long-term studies

without immunologic complications (25). Interestingly, the skin of these

animals has a significantly increased absolute (males and females) and

relative weight (males only) (26, 27). GH transgenic males develop a

phenotype, which has been likened to an “oversized coat”. While skin

alterations in females are moderate, males develop a pronounced dermal

hypertrophy with dermal fibrosis and decreased subcutaneous adipose


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

5 Thorey et al.

tissue. This male-specific phenotype is dependent on androgen production

(26). Here we show that both transgenic males and females suffer from

severe wound healing abnormalities.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

6 Thorey et al.

Experimental Procedures

Animals, wounding and preparation of wound tissue. Hemizygous

PEPCK-bGH transgenic mice in an NMRI outbred background were bred with

C57BL/6 wild-type mice. Transgenic and non-transgenic littermates were

anaesthetised by intraperitoneal (i.p.) injection of ketamine (10 g/l)/xylazine (8

g/l) solution (250µl/25g body weight). Four full-thickness excisional wounds, 5

mm diameter each, were made on either side of the dorsal midline by

excising skin and panniculus carnosus. Wounds were left uncovered and

harvested 5 or 13 days after injury. Mice were housed individually during the

healing period. For expression analyses, two complete wounds from each

animal including 2 mm of the epithelial margins were excised and immediately

frozen in liquid nitrogen. Non-wounded back skin served as a control. For

histological analysis the complete wounds were isolated, bisected, fixed

overnight in 4% paraformaldehyde in PBS (for histology) or 95% ethanol/1%

acetic acid (for immunofluorescence, bromodeoxyuridine (BrdU) staining and

morphometry) and embedded in paraffin. Sections (7 µm) from the middle of

the wound were stained using the Masson Trichrome procedure.

Morphometrical measurements of area and length were performed on BrdU-

stained sections using the OpenLab software (Improvision Ltd., Basel,

Switzerland) and are represented as single data points and median.

Alternatively, all values ranging from 25% below and above the median were

represented as boxes (25th tp 75th percentile). In this case the highest and

lowest values were also indicated as bars. Statistical significance of

differences was determined by the Mann-Whitney U test for non-Gaussian

distribution. All experiments with animals were carried out with permission

from the local veterinary authorities.

Wound bursting strength. Mice were anesthetized as described above. The


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

7 Thorey et al.

dorsal region was shaved and treated with a depilatory agent (Pilca Perfect,

Stafford-Miller Continental, Oevel, Belgium). Two full-thickness incisions (1

cm) were made at one anterior and one posterior, dorsal site, and the skin

margins were closed with strips of a wound plaster (Fixomull stretch,

Beiersdorf, Hamburg, Germany). Mice were sacrificed on day 5 post-

wounding, and bursting strength of the wounds was determined in situ using

the BTC-2000 system (SRLI Technologies, Nashville, TN) according to the

manufacturer`s protocol for the non-human disruptive linear incision analysis.

Data are shown as median, 25th to 75th percentile, and highest and lowest

value. Statistical significance of differences in tensile strength between the

groups was determined by the Mann-Whitney U test.

Labelling with bromodeoxyuridine (BrdU). BrdU labelling was performed

as described (28). Sections (7 µm) from the middle of the wound were

incubated with a peroxidase-conjugated monoclonal antibody directed against

BrdU (Roche Diagnostics Ltd, Rotkreuz, Switzerland) and stained with 0.05%

diaminobenzidine/0.01% H2O2.

Immunofluorescence and immunohistochemistry. Acetic acid/ethanol-

fixed paraffin sections (7 µm) from the middle of the wound or immortalized

mouse embryonic fibroblasts were incubated with antibodies directed against

α-smooth muscle actin (SMA) (anti-αSM-1, IgG2a mAb (28)), ED-A

fibronectin (IST-9, IgG1 mAb, kind gift from Dr. L. Zardi, National Institute for

Cancer Research, Laboratory of Cell Biology, Genova, Italy (29)), desmin

(D33, IgG1 mAb, Dako, Copenhagen, Denmark), PECAM-1 (BD Biosciences,

Heidelberg, Germany), Ly-6G (BD Biosciences), F4/80 (Serotec, Oxford, UK),

cytokeratins 6, 14 (Babco, Richmond, CA) and 10 (Dako), followed by Cy2-,

Cy3-, Cy5-, FITC- or TRITC-, Alexa647- or Alexa568-conjugated conjugated

secondary antibodies (Jackson Immunoresearch Laboratory Inc., West


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

8 Thorey et al.

Grove, PA, USA and Molecular Probes, Eugene, OR). Cell nuclei were

stained with DAPI (Sigma, Munich, Germany) and F-actin with phallodin-

Alexa488 (Molecular Probes). Subsequently, slides were washed in PBS,

prior to mounting in Tris/glycerol containing n-propyl-gallate and analyzed

using a Leica DM IRBE fluorescence microscope equipped with a TCSNT

confocal laser system or a Leica DM RXA2 upright microscope, equipped with

a laser scanning confocal head (TCS SP2 AOBS). Conventional

immunofluorescence images were analyzed on a Zeiss Axiovert 100

microscope.

Quantification of α-SMA expression by fluorescence image analysis.

Expression levels of α-SMA in sections of wound granulation tissue were

quantified as described (30). Briefly, sections were triple-immunostained for

α-SMA, desmin and cell nuclei, and confocal digital images were taken using

a Plan-Neofluar 40x/1.3 objective (Zeiss) mounted on a Zeiss Axiovert 200M,

equipped with a spinning disk Nipkow Confocal head (Yokogawa CSU10,

Visitron Systems, Puchheim, Germany), CoolSNAP-HQ monochrome digital

CCD camera (Photometrics, Roper Scientific, Inc., Trenton, NJ) and image

acquisition software (MetaMorph, Universal Imaging Corp. Downington, PA).

Per image a region of interest was manually selected, including the

granulation tissue and all further automatic processing was restricted to this

region. First, desmin-staining, specific for vascular smooth-muscle cells, was

subtracted from α-SMA- and DAPI-images in order to exclude vessels from

α-SMA quantification. Second, α-SMA expression was analyzed by

calculating the mean pixel intensity of the whole region of interest after

background correction and related to the number of cell nuclei obtained from

the desmin-corrected DAPI-image. Tissue was collected from 3 animals per

experimental condition, 3 sections were prepared per animal and 4 regions of

interest were analyzed per section.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

9 Thorey et al.

Immunohistochemistry and quantification of blood vessels. For

immunohistochemisty, sections were deparaffinized and treated for 40 min at

room temperature with 1.5% H2O2 in methanol to block endogenous

peroxidases. After a 30 min incubation with 10% rabbit serum, sections were

incubated overnight at 4oC with the PECAM-1 antibody, followed by a 1h

incubation with biotinylated donkey anti-rat IgG (Dianova, Hamburg,

Germany). Detection was performed with the peroxidase substrate kit (Vector

Laboratories, Burlingame, USA), as described by the manufacturer. Sections

were counterstained with hematoxylin, mounted and analyzed using a Zeiss

Axioskop 2 equipped with a Zeiss AxioCam HRc camera. Stained vessels

within an area of 0.2 mm x 0.5 mm below the hyperproliferative epithelium

were counted.

Relaxed and stressed collagen lattice contraction. To determine the effect

of GH on the contractile activity of immortalized embryonic murine fibroblasts

cells were grown in stressed collagen lattice (31) at 1.75x105 cells/ml for 5d

(with our without 1 ng/ml TGF-β1). Attached lattices were then released and

diameter reduction after 30 min was normalized to the initial gel diameter. In

one experimental series, mouse sera and GH were added from the beginning

of the 5d (long-term) culture and in a second series only 5h prior to gel

release (short-term). Tractional cell forces were assessed using relaxed, free-

floating collagen lattices as described (32). 9 x 104 cells/ml gel were seeded

into 6-cm Petri dishes, and collagen gel areas were measured 12 h later. A

minimum of 3 lattices was assayed and experiments were performed in

duplicates.

Acetic acid extraction of proteins from skin. Skin samples were

homogenized in liquid nitrogen and extracted in 1 M acetic acid for 2 h at 4°C.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

10 Thorey et al.

Samples were then centrifuged for 5 min at 11.000 g at 4°C and the

supernatant was frozen. Pellets were re-extracted in 1 M acetic acid as

described above. Supernatants were combined and stored at -20°C

overnight. Before determination of protein concentration the samples were

again centrifuged and the supernatants were transferred to fresh tubes, while

the newly formed precipitate was discarded.

Westernblot analysis. 15 µg total protein from lysates of 5-day wounds of

wild-type and transgenic mice were separated by SDS polyacrylamide gel

electrophoresis and transferred to nitrocellulose membranes. Membranes

were probed with primary antibodies against α-SMA and ED-A fibronectin,

stripped and re-probed for total actin using antibody rbAb (Sigma). Detection

was performed with secondary antibodies conjugated to horseradish

peroxidase (Jackson) and the ECL chemiluminescence detection system

(Amersham, Rahn AG, Zürich, Switzerland). Bands were digitized with a

scanner and the ratio between all band densities of one blot was calculated

by commercial computer software (ImageQuant V3.3, Molecular Dynamics,

Sunnyvale, CA). Relative protein expression was normalized to the respective

values for total actin. Samples from 4 animals per experimental condition

were used to calculate mean protein/total actin ratio.

Western ligand blotting. Ligand blot analysis of IGFBPs present in the skin

was performed according to (33), with modifications (34). Briefly, protein

samples were diluted 1:2 with sample buffer (62.5 mM Tris-HCl (pH 6.8), 2%

(w/v) sodium dodecyl sulfate (SDS), 10% (w/v) sucrose), boiled (5 min) and

electrophoresed on a 5% stacking/12% separating SDS-polyacrylamide gel.

Separated proteins were transferred to PVDF membrane (Millipore,

Eschborn, Germany). Blots were blocked with 1% fish gelatin and incubated

with [125I]IGF-II (106 c.p.m. per blot). Binding proteins were visualized and


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

11 Thorey et al.

quantified using the Phosphor-Imager and the ImageQuant software

(Amersham Biosciences, Freiburg, Germany).

Radioimmunoassays for IGF-I, IGF-II and IGFBP-2. IGF-I, IGF-II and

IGFBP-2 were measured using radioimmunoassays described before (35,

36). For all assays, dilution curves of mouse samples were linear and

paralleled those of human standards.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

12 Thorey et al.

Results

bGH transgenic mice show excessive granulation tissue formation and

reduced wound closure at day 5 after wounding. We generated full-

thickness excisional wounds in bGH transgenic mice and control littermates,

and sampled the excised wounds for histological analysis at day 5 after injury

(14 wounds from 7 transgenic males, 12 wounds from 6 wild-type males, 13

wounds from 7 transgenic females and 14 wounds from 7 wild-type females).

At this time point, the inflammatory response is still strong on the one hand,

and on the other hand reepithelialization is already well under way.

Furthermore, part of the clot has been replaced by granulation tissue, which

consists of inflammatory cells, endothelial cells and fibroblasts, which deposit

large amounts of new extracellular matrix (1,2). Fig.1 shows representative

Masson Trichrome stained sections of wounds from transgenic males and

females in comparison to their wild-type littermates. Wounds were much

wider in the transgenic animals of both sexes and filled with a large and cell-

rich granulation tissue (Fig.1A,C,E,G; arrowheads). A large fraction of the

wound was not yet covered by epidermis (Fig.1A,C,E,G; arrows). In the wild-

type animals, on the other hand, the granulation tissue was much less

extended and, therefore, a significant part of the wound was already covered

by newly formed epidermis (Fig.1B,D,F,H; arrows). Finally, the outer margins

of the hyperproliferative epithelium (HE) had moved towards each other due

to contraction of the wounds (Fig.1B,D,F,H; arrowheads).

In addition to the size, the quality of the granulation tissue was also very

different between transgenic mice and control animals: Cells had invaded into

most parts of the original clot and a significant amount of matrix had been

deposited in bGH overexpressing mice (see blue collagen staining). By

contrast, the loose, spongy tissue characteristic for the blood clot was only

partially replaced by granulation tissue in control mice (Fig.1B,D,F). In a few


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

13 Thorey et al.

wild-type mice, where most of the clot was replaced by granulation tissue at

day 5 after injury (Fig.1H), the granulation tissue was much less extended

compared to transgenic mice.

The oval structures that are detectable in the granulation tissue of some

transgenic animals were identified as hypertrophic nerve fibres after staining

of the sections with an antibody against neurofilament M (data not shown).

They were also detected in most control wounds and to a lesser extent in

wild-type mice of other mouse strains (data not shown). Therefore, they

appear not to be a result of bGH overexpression.

To quantify the histological data, one set of sections from one wounding

experiment (including 5 wounds from 3 transgenic males, 4 wounds from 2

wild-type males, 4 wounds from 2 transgenic females and 5 wounds from 3

wild-type females) was used to measure the cross-sectional area of the

granulation tissue (Fig.2A, encircled area) and the progress of wound closure

(% of distance covered by epidermis between the wound edges which are

indicated by arrowheads in Fig. 1A-H and Fig.2A). Differences in wound

closure (p = 0.0159) and granulation tissue area (p = 0.0317) were significant

in both sexes (Fig.2B,C).

To determine the rate of cell proliferation within the hyperproliferative

epithelium, the animals were injected with bromodeoxyuridine (BrdU) 2 h prior

to sacrifice, and sections from the center of the wounds were stained with an

antibody directed against BrdU. The percentage of stained, proliferating cells

was similar in wild-type and transgenic animals (data not shown). Consistent

with this finding, the area of the HE, as well as its length along the border

between epidermis and dermis, which reflects the rate of keratinocyte

migration, were not significantly different between bGH transgenic and wild-

type mice (data not shown). These results demonstrate that the reduced

wound closure is not the result of delayed reepithelialization, but rather of


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

14 Thorey et al.

enhanced granulation tissue formation and reduced wound contraction.

Excessive formation of granulation tissue does not lead to a higher

wound bursting strength in bGH transgenic mice. To determine whether

the formation of a larger and denser granulation tissue in the bGH transgenic

animals leads to a functional improvement, we generated incisional wounds

on the back of bGH transgenic and wild-type males and females. Since

incisional wounds are fully healed at day 5 after injury, we sacrificed the

animals at this time point and measured the negative pressure at which the

wounds ruptured. No significant improvement of bursting strength in

transgenic versus wild-type animals was found; rather, the median bursting

strength was slightly higher in both wild-type males and females, compared to

bGH transgenic males and females, respectively (Fig.2D). Similar to the

results from excisional wounds, histology of incisional wounds showed an,

albeit smaller, increase in the density and volume of the granulation tissue

(data not shown). Thus, the excessively dense and thick granulation tissue in

bGH transgenic animals does not function to confer a higher bursting

strength.

Altered neovascularization in 5-day wounds of bGH transgenic animals.

Analysis of Masson trichrome stained sections suggested that the number of

blood vessels is enhanced in bGH transgenic animals. To verify this

impression, we performed immunofluorescence staining on paraffin sections

of 5-day wounds using an antibody against CD31/PECAM (green) combined

with confocal laser scanning microscopy (Fig.3). This staining revealed an

increase in the number and size of newly formed blood vessels (Fig.3A,C),

compared to wounds of wild-type mice (Fig.3B,D). Similar alterations were

observed in transgenic versus wild-type males. To quantify the difference, we

counted the number of blood vessels below the hyperproliferative wound


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

15 Thorey et al.

epidermis in immunohistochemically stained sections. The difference was

statistically significant (p = 0.0373) in female, but not in male transgenic mice

(p = 0.0703) (Fig.3E). By contrast, we observed a decreased number of blood

vessels within the uninjured skin of transgenic males (data not shown),

demonstrating that bGH alone does not enhance dermal vascularization but

rather requires the presence of additional factors at the wound site to

enhance this process.

In contrast to angiogenesis, no obvious difference in the number of

macrophages was detected between transgenic and wild-type mice as

determined by staining of wound sections with an antibody against the F4/80

antigen (data not shown).

Staining of the keratinocyte-specific markers cytokeratin 14 (Fig.3A-D, red),

cytokeratin 10 (Fig.3A, B; blue-purple) and cytokeratin 6 (data not shown) did

not reveal obvious alterations in keratinocyte differentiation in transgenic as

compared to wild-type animals.

bGH transgenic males show a marked delay in healing at day 13 after

wounding. Fig.4 shows representative Masson Trichrome-stained sections

from 13-day wounds (10 wounds from 6 transgenic males, and 9 wounds

each from 5 wild-type males, 5 transgenic females and 5 wild-type females

were analyzed). At this time point, reepithelialization was completed in wild-

type animals, leaving behind a hyperproliferative epithelium and a granulation

tissue, which both had been greatly reduced in size during the second week

after injury. By contrast, bGH transgenic males (Fig.4A,C) still had an

extensive hyperthickened epithelium, and most notably, 2 of the 10 wounds

analyzed were not yet completely reepithelialized. In wild-type males, on the

other hand, contraction of the wounds had brought the margins of the HE

closely towards each other (arrowheads in Fig.4B,D). Furthermore, in

transgenic males a granulation tissue several times the size of that in wild-


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

16 Thorey et al.

type animals in height and width was revealed. All transgenic females had

reepithelialized their wounds after 13 days. Very few wounds taken from

transgenic females still had an enlarged wound epithelium and granulation

tissue (Fig.4E); in the majority of these wounds (Fig.4G), the sizes of these

compartments were similar to those found in wild-type females (Fig.4F,H).

The cross-sectional areas of HE and granulation tissue measured in 13-day

wounds are summarized in Fig.4J and K, respectively. For each gender, three

independent wounding experiments were performed, from which data were

compiled and sections through the middle of the wounds were selected (9

wounds from 5 transgenic males, 8 wounds from 5 wild-type males, 5 wounds

from 3 transgenic females and 7 wounds from 5 wild-type females). The

areas were significantly different (HE; p = 0.0003, granulation tissue; p =

0.0025) between transgenic and wild-type males, but not between transgenic

and wild-type females. In accordance with these observations, the percentage

of BrdU-positive proliferating cells, as well as total cell number within the

wound epidermis were still significantly higher in the transgenic males in

comparison to wild-type males (% BrdU positive cells; p = 0.0091, total

number of cells; p = 0.0001), while these parameters were not significantly

different between transgenic and wild-type females. The increased area and

length of the wound epidermis in male animals is most likely due to the

extended granulation tissue that had to be covered, leading to prolonged

keratinocyte hyperproliferation. Consistent with this finding, we found a

prolonged expression of the differentiation-specific keratin 6 in transgenic

males (data not shown).

Taken together, these data indicate that wound closure is delayed in male

and female bGH transgenic mice within the first five days after injury and

even longer in the male animals.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

17 Thorey et al.

Altered myofibroblast differentiation and organization of α-smooth

muscle actin (α-SMA) in 5-day wounds of bGH transgenic animals. Of

particular importance for wound contraction and closure is a subpopulation of

fibroblasts that differentiate into myofibroblasts during the course of wound

healing (37). To investigate whether the obvious defect in wound contraction

in the bGH transgenic animals is due to alterations in myofibroblast

differentiation, sections were stained for α-SMA, which is expressed by

myofibroblasts. In wounds from wild-type females, a highly organized

structure of α-SMA-containing stress fibers underlying the hyperproliferative

epithelium was revealed (Fig.5B,D; green), with filaments extending along the

hyperproliferative wound epithelium and also underneath the uninjured

dermal tissue at the edge of the wound towards the subcutaneous muscle,

the panniculus carnosus. The latter can be identified on the basis of strong

desmin staining (red). In the granulation tissue, α-SMA expressing cells could

unmistakably be identified as myofibroblasts (green), since they were

negative for desmin in contrast to smooth muscle cells of small vessels, which

co-express α-SMA and desmin (Fig.5C,D, yellow, indicated by arrows).

However, in wounds from transgenic females, fewer myofibroblasts residing

underneath the keratinocyte tongues at the wound edges expressed much

less α-SMA, most of which was not organized into parallel filaments but rather

extended in all directions (Fig.5A,C; green). A similar difference was also

seen in male mice (data not shown). Quantification of myofibroblast

appearance in granulation tissue by immunofluorescence intensity analysis of

5d wound sections demonstrated approximately 3-times higher α-SMA

expression in female wild-type compared to transgenic mice. This was

confirmed by Westernblot analysis of total lysates of 5-day wound tissue

(Fig.6A,B), revealing a 3-fold and 2.5-fold higher expression of α-SMA in wild-

type male and female mice compared to their respective transgenic

littermates. A second marker for myofibroblasts, ED-A fibronectin (38) was


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

18 Thorey et al.

expressed at significantly higher levels in wild-type (4-fold) compared to

transgenic female mice, but exhibited similar expression levels in male

transgenic and control mice. These results suggest that impaired wound

contraction in bGH transgenic animals correlates with a decreased number of

myofibroblasts and an impairment of stress fiber organization in wound

myofibroblasts.

Effect of bGH on collagen contraction and myofibroblast differentiation.

As a next step we determined whether the impaired wound contraction, which

we observed in bGH-overexpressing transgenic mice, is a direct effect of GH

on the generation of contractile and tractional forces by fibroblasts. For this

purpose we analyzed the effect of recombinant bGH on the ability of murine

embryonic fibroblasts to contract attached (stressed) three-dimensional

collagen lattices, which are a model of the organized matrix in advanced

granulation tissue (37), and to retract freely floating (relaxed) gels, in which

loose collagen fibrils are reorganized by migrating fibroblasts (39),

comparable to early dermal wound healing. The GH concentrations that we

used ranged from 0.1% - 10% of the serum concentrations of bGH transgenic

mice, and therefore likely reflect the concentrations that reach the fibroblasts

within the dense granulation tissue. When GH was added for 5h to attached

(Fig.7A, control) and 12h to free-floating gels (data not shown) in control

medium, we did not observe any inhibitory effect of GH on collagen area

reduction, but rather a tendency of this hormone to enhance

contraction/retraction in a dose-dependent manner. As a next step, we

determined if growth hormone affects myofibroblast differentiation and the

contractile ability of these cells. For thus purpose, fibroblasts were cultured

for 5 days in the presence of transforming growth factor β1 (TGF-β1). This

factor was chosen, since TGF-βs are the most potent known inducers of

myofibroblast differentiation (31, 40), and since TGF-β1 is the predominant


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

19 Thorey et al.

TGF-β isoform in murine skin wounds (41). When fibroblasts were grown for

5d in stressed gels in the presence of TGF-β1, i.e. when myofibroblast

differentiation has been induced, short-term addition of GH moderately

decreased myofibroblast contractile activity (Fig.7A, TGF-β1). Interestingly,

addition of GH for the whole culture period of 5d had no effect on stressed

collagen contraction in control medium (Fig.7C, control), but significantly

reduced the contractile activity of (myo)fibroblasts grown in the presence of

TGF-β1 (Fig.7C, TGF-β1).

Next, we tested the possibility that the serum of bGH transgenic mice

contains an inhibitory activity, which counteracts the positive short-term

influence of GH on fibroblast collagen gel reduction. In control medium,

replacement of 5% FCS with pooled sera of transgenic or wild-type mice

resulted in an improvement of gel retraction (data not shown) and contraction

when added for 5h (Fig.7B, control) and 5d (Fig.7D, control); this effect is

presumably due to the greater efficiency of murine over bovine growth

factors. At lower concentrations of mouse sera the degree of collagen gel

retraction was intermediate between the control and the 5% mouse serum

values, but no significant difference was seen between sera from transgenic

or wild-type mice at any concentration tested (data not shown). When added

to TGF-β1-treated (myo)fibroblasts, serum from transgenic mice slightly

reduced stressed collagen contraction compared to 10% FCS and wild-type

mouse serum (Fig.7B, D, TGF-β1).

To assess the level of myofibroblast differentiation, murine embryonic

fibroblasts were treated for 5d with TGF-β1 in the presence or absence of

bGH or mouse sera. TGF-β1 strongly induced expression of α-SMA (Fig.7E),

compared to almost α-SMA-negative cells grown in control medium (data not

shown). This α-SMA-inducing effect of TGF-β1 was completely abolished by

simultaneously adding GH (Fig.7F) and moderately reduced by transgenic

mouse serum (Fig.7H); wild-type mouse serum did not change α-SMA


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

20 Thorey et al.

expression of TGF-β1-treated fibroblasts (Fig.7G). When added to fibroblasts

in control medium, GH and mouse sera had no significant effect on α-SMA

expression (data not shown).

Thus in long-term experiments, GH appears to inhibit TGF-β1-induced

myofibroblast differentiation and α-SMA expression, which is most likely

responsible for reduced collagen (31) and wound contraction. Moreover, in

short-term treatment, GH seems to inhibit the contractile activity of

differentiated myofibroblasts but not the retraction forces exerted by α-SMA-

negative fibroblasts.

Elevated levels of IGFBPs, but not of IGF-I, in 5-day wounds of bGH

transgenic mice. Since many effects of GH are mediated via IGFs, we

compared the levels of IGF and IGF binding activities in skin and wound

lysates derived from bGH transgenic mice and wild-type mice. IGF-I levels in

protein lysates prepared from individual mice (n=4) were determined by

quantitative radioimmunoassay (Fig.8A). As reported previously for serum

IGF-I levels (27), they were elevated by a factor of 2-3 in transgenic skin of

both males and females, compared to wild-type (males; p = 0.0006, females;

p = 0.0008). In the females, a 6-fold upregulation of IGF-I protein in

comparison to non-wounded skin was observed in both wild-type and

transgenic animals at day 5 after injury; however, no difference was observed

between transgenic and wild-type animals. Much less upregulation of IGF-I

was observed in the males; levels in 5-day-wounds of both wild-type and

transgenic males were similar to those in the uninjured skin of transgenic

males. The difference between males and females was highly significant for

both wild-type (p = 0.0037) and transgenic (p = 0.0025) mice. Thus, first, IGF-

I levels seem to be generally lower in healing wounds of male mice,

compared to females; and second, in bGH transgenic mice of both genders,

the higher abundance of IGF-I in intact skin of transgenic mice compared to


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

21 Thorey et al.

wild-type mice is not maintained after injury . When levels of IGF-II and

IGFBP-2 within the same samples were determined, no significant differences

were revealed between wild-type and bGH transgenic mice (data not shown).

IGF binding activities in skin and 5-day wounds were analyzed by western

ligand blotting of protein lysates from individual mice (n = 3 for each gender

and genotype) with a radiolabelled IGF-II tracer (Fig.8B). Three labelled

bands visualized on the blots corresponded to IGFBP-3 (35-45 kDa), IGFBP-

5 (25-35 kDa) and IGFBP-4 (18-25 kDa), the three IGF binding proteins,

which are expressed mainly by dermal fibroblasts (18, 42). An increase in IGF

binding to all three proteins could be observed in skin lysates from both

transgenic males and females, compared to wild-type animals; this increase

was significant for IGFBP-3 in both genders (males; p = 0.0085, females; p =

0.0013) and for IGFBP-4 in the females (p = 0.017), and not significant in the

other skin samples, due to larger variations between the individuals.

Following injury, IGF binding activity was upregulated (2-4-fold in the wild-type

animals and 1.5-3-fold in bGH transgenic mice). In 5-day wound lysates, a

significant difference was observed between wild-type and bGH transgenic

mice in binding to both IGFBP-3 (males; p = 0.0052, females; p= 0.025) and

IGFBP-4 (males; p = 0.0014, females; p = 0.0008), while no difference was

apparent in binding to IGFBP-5. Thus, overall in the wounds of bGH

transgenic mice the balance between IGFs and IGFBPs is shifted towards an

under-abundance of IGF-I and a greater relative activity of the predominantly

inhibitory IGF binding proteins, IGFBP-3 and IGFBP-4.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

22 Thorey et al.

Discussion

Growth hormone: beneficial or detrimental for wound healing? We have

studied the wound healing process in transgenic mice with highly elevated

levels of serum bGH. The most striking phenotype was the strong increase in

granulation tissue formation in female and male transgenic animals. This

result is consistent with data obtained by other investigators who found that

local or systemic application of GH stimulates granulation tissue formation in

wounds of normal and malnourished rats (8, 43). However, in the latter and in

other studies (9), the increased granulation tissue correlated with an improved

biomechanical strength of the wounds. By contrast, we could not detect an

increase in wound bursting strength, demonstrating that matrix maturation is

not improved by the elevated levels of bGH. These discrepancies between

our results and previous studies might result from the use of healing-impaired

animals in most treatment studies. These animals suffer from reduced

mechanical wound strength, which is likely to be improved by application of

GH. However, we show that GH is not able to further increase the normal

bursting strength of a non-compromised wound and can even delay the

wound healing process. Thus, GH treatment of healthy individuals appears to

have adverse effects on skin wound healing. On the other hand, GH was

found to be helpful in severely burned animals or patients, who are

characterized by the induction of a catabolic state with decreased GH and

IGF levels. In this case the action of exogenously administered GH may lie

more in the restoration of the anabolic metabolism, rather than in the

stimulation of tissue repair at the cellular level (7).

Enhanced wound angiogenesis in bGH transgenic mice. A remarkable

feature of the granulation tissue of bGH transgenic mice was the increased

number of blood vessels. This may be a direct effect of GH, since the latter

was shown to stimulate angiogenesis (44, 45). However, GH alone is not


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

23 Thorey et al.

sufficient to stimulate this process in the skin, since the number of blood

vessels in unwounded skin of bGH transgenic males was even reduced.

Thus, additional factors present at the wound site are likely to contribute to

the increased angiogenesis. Possible candidates are VEGF, IGF-I and TGF-

β1, angiogenic factors which are overexpressed in skin wounds (2). These

growth factors might act synergistically with GH to stimulate angiogenesis. By

contrast, they are probably not directly responsible for the increase in vessel

number and size, since their expression was not enhanced in wounds of bGH

transgenic mice compared to control animals (data not shown). Furthermore,

IGF-I activity in wounded skin of bGH mice appears even reduced due to the

enhanced expression of inhibitory IGFBPs at the wound site. Therefore, the

stimulatory effect of GH on granulation tissue formation is most likely not

mediated via IGF-I.

Abnormal wound contraction and myofibroblast differentiation in bGH

transgenic mice. The most striking abnormality in the bGH transgenic mice

was the inefficient wound contraction, leading to a delay in wound closure,

which is more pronounced in the males than in the females. It has been

suggested that the initial closure of the wound is due to tractional forces that

are developed by migrating fibroblasts (46). This process can be mimicked in

vitro using free-floating collagen lattices, where the reduction in collagen

lattice diameter is entirely due to tractional forces exerted by fibroblasts (39).

However, collagen gel retraction in vitro was even enhanced by recombinant

bGH and by serum from bGH transgenic mice. The positive effect may result

from the induction of IGF-I expression in fibroblasts by bGH, since IGF-I is

known to promote fibroblast-mediated collagen gel contraction (47, 48).

At later stages of wound healing, a subpopulation of fibroblasts differentiates

into myofibroblasts, which are thought to be responsible for contraction of the

granulation tissue (37). Remarkably, a strongly reduced number of


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

24 Thorey et al.

myofibroblasts was observed in bGH transgenic mice, and the existing

myofibroblasts showed a decrease in the abundance of parallel α-SMA

containing stress fibers. This finding suggests that the contractile capability of

the existing myofibroblasts is impaired. Of particular importance for

myofibroblast differentiation is TGF-β1 (40). The normal expression of this

growth and differentiation factor in bGH transgenic mice suggests that

alterations in the levels of TGF-β1 are not responsible for the defect in

myofibroblast differentiation. By contrast, our in vitro studies with stressed

three-dimensional collagen lattices revealed that GH reduces the contractile

activity of (myo)fibroblasts grown in the presence of TGF-β1. Consistent with

this observation, TGF-β1-induced myofibroblast differentiation in vitro was

strongly inhibited by long-term treatment with bGH. This result revealed a

novel role of GH as an inhibitor of TGF-β-induced myofibroblast differentiation

and suggests that this hormone also impairs TGF-β-induced myofibroblast

differentiation and the contractile capacity of these cells in wounded skin.

Interaction of growth hormone with androgens at the wound site?

Interestingly, the enlarged, highly vascularized granulation tissue and the

myofibroblast abnormalities were observed in both male and female

transgenic mice. This finding demonstrates that the abnormalities seen in

early wounds are not a result of the highly fibrotic skin of male transgenic

animals, which could impair wound contraction. Rather, bGH appears to

affect granulation tissue formation in mice of both genders in a similar

manner. However, the phenotype was stronger in the male animals, possible

due to the influence of sexual hormones as already observed in non-injured

skin of these animals (26). In this study, the severe skin fibrosis seen in male

animals was prevented when the animals were castrated, demonstrating that

androgens synergize with GH and possibly locally acting growth factors to

induce skin fibrosis. In analogy, the continuous production of matrix at the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

25 Thorey et al.

wound site and the extension of the inflammatory phase (data not shown)

could also be influenced by androgens. This hypothesis is supported by

findings from Ashcroft and Mills (5), who found that testosterone is a negative

regulator of wound healing and enhances the inflammatory response by up-

regulation of pro-inflammatory cytokines.

In summary, we have shown striking effects of enhanced serum GH levels on

the wound healing process. In addition to its known effect on granulation

tissue formation, our results revealed a novel role of GH in wound

angiogenesis. Most interestingly, GH was identified as an inhibitor of TGF-β-

induced myofibroblast differentiation. Finally, our data strongly suggest that

systemic GH treatment is detrimental for wound healing in non-healing

impaired individuals, in particular in males.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

26 Thorey et al.

Acknowledgements

We thank C. Born-Berclaz and J. Smith-Clerc for excellent technical

assistance and Dr. Luciano Zardi (National Institute for Cancer Research,

Laboratory of Cell Biology, Genova, Italiy) for providing the IST-9 antibody.

Dr. J.-J. Meister is gratefully acknowledged for his constant support of the

work of B.H. and for providing lab facilities at the EPFL. This work was

supported by grants from the Swiss National Science Foundation (No. 31-

61358.00 to S.W. and No.3100A0-102150/1 to B.H) and the German Ministry

for Education and Research (BMBF) (to S.W. and E.W.).


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

27 Thorey et al.

References

1. Martin, P. (1997) Science 276, 75-81

2. Werner, S., and Grose, R. (2003) Phys. Rev. 83, 835-870

3. Ashcroft, G.S., Dodsworth, J., van Boxtel, E., Tarnuzzer, R.W., Horan,

M.A., Schultz, G.S., and Ferguson, M.W. (1997) Nat. Med. 3, 1209-

1215

4. Grose, R., Werner, S., Kessler, D., Tuckermann, J., Huggel, K., Durka,

S., Reichardt, H.M., and Werner, S. (2002) EMBO reports 3: 575-582.

5. Ashcroft, G.S., and Mills, S.J. 2002. J. Clin. Invest. 110, 615-624

6. Cummings, D.E., and Merriam. G.R. (2003) Annu Rev Med. 54, 513-

533.

7. Lal, S.O., Wolf, S.E., and Herndon, D.N. (2000) Growth Horm. IGF

Res. 10, Suppl. B, S39-S43

8. Steenfos, H.H. and Jansson, J.O. (1992) J. Endocrinol. 132, 293-298

9. Seyer-Hansen, M., Andreassen, T.T., Jorgensen, P.H., and Oxlund, H.

(1993) Eur. Surg. Res. 25, 162-168

10. Rasmussen, L.H., Garbarsch, C., Schuppan, D., Moe, D., Horslev-

Pedersen, K., Gottrup, F., and Steenfos, H. (1995) Eur. J. Surg. 161,

157-162

11. Ghofrani, A., Holler, D., Schuhmann, D., Saldern, S., and Messmer,

B.J. (1999) Plast. Reconstr. Surg. 104, 470-475.

12. Seyer-Hansen, M., Andreasen, T.T., and Oxlund, H. (1999) Growth

Horm. IGF Res. 9, 254-261

13. Welsh, K.M., Lamit, M., and Morhenn, V.B. (1991) J. Invest. Surg.

Oncol. 17, 942-945

14. Green, H., Morikawa, M., and Nixon, T. (1985) Differentiation 29, 195-

198

15. Eming, S.A, Snow, R.G., Yarmush, M.L., and Morgan, J.R. (1996) J.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

28 Thorey et al.

Invest. Dermatol. 107, 113-120

16. Martin, J.L., and Baxter, R.C. (1999) in Contemporary Endocrinology;

The IGF System. (Rosenfeld, R., and Roberts, C., eds) pp. 227-254,

Humana Press, Inc. Totowa, NJ

17. Kuroda, K., Utani, A., Hamasaki, Y. and Shinkai, H. (2001) J.

Dermatol. Sci. 26, 156-160

18. Clemmons, D.R. (1999) in Contemporary Endocrinology; The IGF

System. (Rosenfeld, R., and Roberts C. jr., eds) pp. 273-279, Humana

Press, Inc. Totowa, NJ

19. Schneider, M.R., Wolf, E., Hoeflich, A., and Lahm. H. (2002) J.

Endocrinol. 172, 423-440

20. Lynch S,E., Colvin, R.B. and Antoniades, H.N. (1989) J. Clin. Invest.

84, 640-646

21. Jyung, R.W., Mustoe, J.A., Busby, W.H., and Clemmons, D.R. (1994)

Surgery 115, 233-239.

22. Bitar, M.S. (1997) Horm. Metab. Res. 29, 383-386

23. Jeschke, M.G., Barrow, R.E., Hawkins, H.K., Yang, K., Hayes, R.L.,

Lichtenbelt, B.J., Perez-Polo, J.R., and Herndon, D.N. (1999) Gene

Ther. 6, 1015-1020

24. Dunaiski, V., and Belford, D.A. (2002) Growth Hormone IGF Res. 12,

381-387

25. Wolf, E., and Wanke, R. (1996) in Welfare of Transgenic Animals.

Zutphen, L.F.M., and van der Meer, M., eds) pp. 26-47, Springer

Verlag. Heidelberg, Germany

26. Wanke, R., Milz, S., Rieger, N., Ogiolda, L., Renner-Muller, I., Brem,

G., Hermanns, W., and Wolf, E. (1999) J. Invest. Dermatol. 113, 967-

971.

27. Hoeflich, A., Nedbal, S., Blum, W.F., Erhard, M., Lahm, H., Brem, G.,

Kolb, H.J., Wanke, R., and Wolf, E. (2001) Endocrinol. 142, 1889-1898


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

29 Thorey et al.

28. Skalli, O., Ropraz, P., Trzeciak, A., Benzonana, G., Gillessen, D., and

Gabbiani, G. (1986). J. Cell Biol. 103, 2787-2796.

29. Borsi, L., Carnemolla, B., Castellani, P., Rosellini, C., Vecchio, D.,

Allemanni, G., Chang, S.E., Taylor-Papadimitriou, J., Pande, H., and

Zardi, L. (1987). J. Cell Biol. 104, 595-600.

30. Hinz, B., Mastrangelo, D., Iselin, C.E., Chaponnier, C., and Gabbiani,

G. (2001) Am. J. Pathol. 159, 1009-1020.

31. Hinz, B., Celetta, G., Tomasek, J.J., Gabbiani, G., and Chaponnier, C

(2001a). Mol. Biol. Cell 12, 2730-2741

32. Kessler, D., Dethlefsen, S., Haase, I., Plomann, M., Hirche, F., Krieg,

T., and Eckes, B. (2001) J. Biol. Chem. 276, 36575-36585

33. Hossenlopp, P., Seurin, D., Segovia Quinson, B., Hardouin, S., and

Binoux, M. (1986) Anal. Biochem. 154, 138-143.

34. Wolf, E., Kramer, R., Blum, W. F., Foll, J., and Brem, G. (1994)

Endocrinol. 135, 1877-1886

35. Bang, P., Baxter, R. C., Blum, W. F., Breier, B. H., Clemmons, D. R.,

Hall, K., Hintz, R. L., Holly, J. M., Rosenfeld, R. G., and Zapf, J. (1994)

J.Endocrinol. 143, C1-2

36. Hoeflich, A., Wu, M., Mohan, S., Föll, J., Wanke, R., Froehlich, T.,

Arnold, G., Lahm, H., Kolb, H. J., and Wolf, E. (1999) Endocrinol. 140,

5488-5496

37. Tomasek, J.J., Gabbiani, G., Hinz, B., Chaponnier, C. and Brown, R.A.

(2002) Nat. Rev. Mol. Cell Biol. 3, 349-363

38. Serini, G., Bochaton-Piallat, M.L., Ropraz, P., Geinoz, A., Borsi, L.,

Zardi, L., and Gabbiani, G. (1998). J. Cell Biol. 142, 873-881

39. Bell, E., Ivarsson, B., and Merrill, C. (1979) Proc. Natl. Acad. Sci. USA

76, 1274-1278

40. Desmouliere, A., Geinoz, A., Gabbiani, F., and Gabbiani, G. (1993) J.

Cell. Biol. 122, 103-111


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

30 Thorey et al.

41. Frank, S., Madlener, M., and Werner, S. (1996). J. Biol. Chem. 271,

10188-10193

42. Giannini, S., Mohan, S., Kasuya, J., Galli, G., Rotella, C.M., LeBon,

T.R., and Fujita-Yamaguichi, Y. (1994) J. Clin. Endocrinol. Metabol. 79,

1824-1830

43. Zaizen, Z., Ford, E.G., Shimada, H., Kosi, M., Costin, G., and Atkinson,

J.B. (1995) Eur. J. Pediatr. Surg. 5, 226-230

44. Gould, J., Aramburo, C., Capdevielle, M., and Scanes, C.G. (1995) Life

Sci. 56, 587-594

45. Struman, I., Bentzien, F., Lee, H., Mainfroid, V., D'Angelo, G., Goffin,

V., Weiner, R.I., and Martial, J.A. (1999) Proc. Natl. Acad. Sci. USA

96, 1246-12451

46. Ehrlich, H.P., and Rajaratnam, J.B. (1990) Tissue Cell 22, 407-417

47. Lee, Y.R., Oshita, Y. and Tsuboi, R. (1996) Endocrinol. 137, 5378-

5283

48. Kanekar, S., Borg, T.K., Terracio, and L., Carver, W. (2000) Cell

Adhes. Commun. 7, 513-523.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

31 Thorey et al.

Figure legends

Fig. 1. Impaired wound healing in bGH overexpressing transgenic mice.

Masson Trichrome stained paraffin sections from the center of 5-day wounds

of bGH overexpressing transgenic males (A,C) and females (E,G) and their

male and female wild-type littermates (B,D; and F,H; respectively). Shown are

low-magnification views of representative wounds. Arrows indicate the

foremost tips of the epithelial tongues. Arrowheads indicate the wound edges.

D, dermis; E, epidermis; Es, eschar; G, granulation tissue; HE,

hyperproliferative epithelium; M, muscle (Panniculus carnosus). Note the

extensive skin fibrosis in transgenic males, and the increased granulation

tissue in all transgenic animals. Bars represent 1 mm.

Fig. 2. Delayed wound closure, enlarged granulation tissue, but no

improvement of wound bursting strength in wounds of bGH transgenic

mice. A set of representative sections from full-thickness excisional wounds

(Fig.1) was subjected to morphometric analysis as illustrated in A. The area of

granulation tissue, which was measured, is encircled by a black line. Wound

closure was determined as percentage of distance covered by epidermis

between the wound edges which are indicated by arrowheads in Fig. 1A-H,

Fig. 2A). B,C: Both the percentage of wound closure (J; p=0.159) and the

area of granulation tissue/fibrin clot (K; p=0.317) were significantly different

between transgenic and wild-type males as well as between transgenic and

wild-type females (open circles). Shown are single data points for each

wound and median. *p<0.05. D: Incisional wounds were created on the back

of bGH transgenic and wild-type mice 5 days prior to post-mortem

measurement of bursting strength using a suction device. Shown are median

values, 25th to 75th percentile (box) and range (vertical bars) of males (19

transgenic; 50 wild type) and females (20 transgenic; 26 wild type).


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

32 Thorey et al.

Differences between the four groups are not significant. Note the extended

upper range of wild-type wound bursting strength, presumably reflecting

completed reepithelialization of some wounds.

Fig. 3. Enhanced angiogenesis in bGH transgenic mice. A-D: Paraffin

sections prepared from 5-day wounds of bGH transgenic mice or their wild-

type siblings were triple-stained with a polyclonal antiserum against

cytokeratin 14, followed by Cy-3 conjugated anti-rabbit IgG (red); a rat

monoclonal antibody against CD 31/PECAM, followed by FITC-conjugated

anti-rat-IgG (green), and a mouse monoclonal antibody against cytokeratin

10, followed by Cy-5 conjugated anti-mouse IgG (blue; signal throughout the

fibrin clot/granulation tissue probably represents binding to mouse serum

IgGs). Shown are low-magnification views of representative wound edges

(A,B, bars represent 100 µm) and higher magnification views of areas of the

fibrin clot/granulation tissue underlying the hyperproliferative epithelium (C,D,

bars represent 20 µm). D, dermis; E, epidermis; G, granulation tissue; HE,

hyperproliferative epithelium. E: Quantification of blood vessels below the

hyperproliferative epithelium in sections from female transgenic mice (10

wound halves) and control animals (5 wound halves) and from male

transgenic mice (8 wound halves) and control animals (5 wound halves)

revealed a significantly higher number of vessels in transgenic females

compared to controls (p= 0.0373), whereas the difference was not significant

in males (p = 0.0703). Shown are median values, 25th to 75th percentile (box)

and range (vertical bars).

Fig. 4. Severely delayed wound healing in bGH overexpressing males.

Masson-Trichrome stained paraffin sections from the center of 13-day

wounds from bGH transgenic males (A,C) and females (E, G) and their male

and female wild-type siblings (B,D; and F,H; respectively). Shown are low-


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

33 Thorey et al.

magnification views of typical wounds. Panel A shows a wound with

incomplete closure; arrows indicate the foremost tips of the epithelial tongues.

Arrowheads in all panels indicate the wound edges. D, dermis; E, epidermis;

G, granulation tissue; HE, hyperproliferative epithelium; M, muscle

(Panniculus carnosus). Note the enlarged granulation tissue of the transgenic

males and of one transgenic female. Bars represent 1 mm.

J,K: A set of representative sections was subjected to morphometric analysis

which revealed highly significant differences in the areas of the

hyperproliferative epithelium (J; p=0.0003) and the granulation tissue (K;

p=0.0025) between transgenic males (filled squares) and wild-type males

(open squares), while differences between transgenic females (filled circles)

and wild-type females (open circles) were not significant. Shown are single

data points for each wound, and median. **p<0.01; ***p<0.001

Fig.5. Altered myofibroblast differentiation in bGH transgenic mice.

Paraffin sections prepared from 5-day wounds of bGH transgenic mice or

their wild-type siblings were triple-stained with a polyclonal antiserum

against cytokeratin 14 followed by Alexa647-conjugated anti-rabbit IgG (A-

D, blue), a mouse monoclonal IgG2a antibody against α-SMA (green)

followed by FITC-conjugated anti-mouse IgG2a and a mouse monoclonal

IgG1 antibody against desmin (red) followed by TRITC-conjugated anti-

mouse IgG1. Sections were viewed by laser confocal microscopy and

overlays were prepared from the extended-view images from all three

channels. Shown are low-magnification views of representative wound

edges (A,B,: bars represent 100 µm) and higher magnification views of

areas of the fibrin clot/granulation tissue underlying the hyperproliferative

epithelium (C,D; bars represent 20 µm). D, dermis; E, epidermis; G,


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

34 Thorey et al.

granulation tissue; HE, hyperproliferative epithelium. M in panels A, B

indicates the position of the subcutaneous muscle (panniculus carnosus)

and arrowheads (C, D) point to small vessels.

Fig.6. Expression of myofibroblast marker proteins is down-regulated

in bGH transgenic mice. (A) Expression of total actin and the

myofibroblast markers ED-A fibronectin and α-SMA was determined by

Western blotting on total lysates, prepared from 5-day wounds of bGH

transgenic (tg) male and female mice and their wild-type littermates (wt).

Protein loads have been equilibrated for vimentin expression levels. (B)

Protein bands were scanned, quantified by image analysis and normalized

to total actin. Results are presented as mean values ± standard deviation.

Significance was tested by means of a two-taled heteroscedastic

Student`s t-test. Differences were considered to be statistically significant

at values of p<0.05. In contrast to largely unchanged levels of total actin,

expression of α-SMA is significantly reduced in transgenic compared to

wild-type mice. ED-A fibronectin (ED-A FN) is significantly reduced in

female but not in male transgenic mice compared to their respective wild-

type littermates (**p<0.01).

Fig. 7. Effects of bGH and transgenic mouse serum on stressed

collagen gel contraction and myofibroblast differentiation.

Immortalized mouse fibroblasts were seeded into mechanically stressed,

attached collagen gels and grown for 5d in normal medium (control) and

medium, containing 1 ng/ml TGF-β1 (TGF-β1) in order to induce


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

35 Thorey et al.

myofibroblast differentiation. Gels were additionally treated with different

concentrations of bGH (12.3 ng/ml, 61.3 ng/ml, or 245 ng/ml) (A, C) or 5%

FCS + 5% pooled sera from wild-type mice; or from their bGH transgenic

siblings (bGH levels in pooled sera from transgenic mice were 1.6 µg/ml)

(B, D). Treatment was either short-termed (5h, A, B) or performed during

the whole culture period of 5d (C, D). After 5d, attached gels were

released and area reduction after 30 min was normalized to control

contraction (10% FCS). Indicated in column charts: *p ≤ 0.05, **p ≤ 0.01,

***p ≤ 0.001 as tested by means of a two-tailed heteroscedastic Student's

t-test within one experimental condition (control, TGF-β1) against the

respective 0-value (A,C) or between wt and tg contraction (B, D); n=3.

After 5d culture in medium containing 1 ng/ml TGF-β1 (E) and additionally

245 ng/ml bGH (F), 5% serum from wild-type mice + 5% FCS (G) or 5%

serum from transgenic mice + 5% FCS (H), cells were stained for α-SMA

(red), F-actin (phalloidin, green) and nuclei (DAPI, blue) to determine the

level of myofibroblast differentiation. Note that bGH significantly reduces

the level of α-SMA expression. Bar: 50 µm.

Fig. 8. Altered balance between IGF-I and IGFBPs in wounds of bGH

transgenic mice. Levels of IGF-I in total protein lysates prepared from

uninjured skin and 5-day wounds (5d W) of bGH transgenic males and

females and their wild-type littermates were determined by quantitative

radioimmunoassay, and means and SEM were calculated (A). Significant

differences were revealed between transgenic and wild-type skin of both

males (p =0.0006) and females (p = 0.0008).

Aliquots of total protein lysates from uninjured skin and 5-day wounds of bGH

transgenic males and females and their wild-type littermates were subjected


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

36 Thorey et al.

to Western ligand blot analysis using radiolabelled IGF-II tracer (B,C). Binding

activity was determined separately for 3 different samples per gender,

genotype and condition in one experiment, and means and SEM were

calculated (B). IGFBP-3 binding was significantly different between transgenic

and wild-type animals in the skin of males (p = 0.0085) and females (p =

0.0013), as well as in 5-day wounds of males (p = 0.0052) and females (p =

0.025). IGFBP-4 binding was significantly different between transgenic and

wild-type animals in the skin of the females only (p = 0.017), and in 5-day

wounds of both males (p = 0.0014) and females (p = 0.0008). As an example,

the image of a ligand blot prepared in a different experiment with 5-day

wound samples of females is also shown (C). 3, 4, 5 indicates the positions of

the bands corresponding to IGFBP-3, IGFBP-5, and IGFBP-4. Note that the

medium-size band in the serum positive control (S), which corresponds to

IGFBP-2, does not co-migrate with IGFBP-5. *p<0.05, **p<0.01; ***p<0.001.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

Martin Elmlinger, Ruediger Wanke, Eckhard Wolf and Sabine WernerIrmgard S. Thorey, Boris Hinz, Andreas Hoeflich, Susanne Kaesler, Philippe Bugnon,

angiogenesis and myofibroblast differentiationTransgenic mice reveal novel activities of growth hormone in wound repair,

published online April 7, 2004J. Biol. Chem.

10.1074/jbc.M311467200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M311467200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;M311467200v1&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2004/04/07/jbc.M311467200.citation

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=early/2004/04/07/jbc&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/early/2004/04/07/jbc.M311467200.citation

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/

Documents

Transgenic mice reveal novel activities of growth hormone · 8 Thorey et al. Grove, PA, USA and Molecular Probes, Eugene, OR). Cell nuclei were stained with DAPI (Sigma, Munich, Germany)