Proc. Natl. Acad. Sci. USA
Vol. 92, pp. 2572-2576, March 1995
Hepatic expression of mature transforming growth factor (81 in
transgenic mice results in multiple tissue lesions
NANCY SANDERSON*, VALENTINA FACTOR*, PETER NAGY*, JEFFREY KOppt, PATURU KONDALAHt,
LALAGE WAKEFIELD*, ANITA B. ROBERTSt, MICHAEL B. SPORNt, AND SNORRI S. THORGEIRSSON*§
Laboratories of *Experimental Carcinogenesis and tChemoprevention, Division of Cancer Etiology, National Cancer Institute, and tLaboratory of Oral Medicine,
National Institute of Dental Research, National Institutes ofHealth, Bethesda, MD 20892
Communicated by Roscoe 0. Brady, National Institutes ofHealth, Bethesda, MD, December 6, 1994 (receivedfor review November 4, 1994)
factor 131 (TGF-f81) has been implicated in a number of
disease processes, particularly those involving fibrotic and
inflammatory lesions. To determine the in vivo effects of
overexpression of TGF-f81 on the function and structure of
hepatic as well as extrahepatic tissues, transgenic mice were
generated containing a fusion gene (Alb/TGF-,81) consisting
of modified porcine TGF-j81 cDNA under the control of the
regulatory elements of the mouse albumin gene. Five trans-
genic lines were developed, all of which expressed the
Alb/TGF-j81 transgene selectively in hepatocytes. The trans-
genic line 25 expressing the highest level of the transgene in
the liver also had high (>10-fold over control) plasma levels
of TGF-181. Hepatic fibrosis and apoptotic death of hepato-
cytes developed in all the transgenic lines but was more
pronounced in line 25. The fibrotic process was characterized
by deposition of collagen around individual hepatocytes and
within the space ofDisse in a radiating linear pattern. Several
extrahepatic lesions developed in line 25, including glomeru-
lonephritis and renal failure, arteritis and myocarditis, as well
as atrophic changes in pancreas and testis. The results from
this transgenic model strongly support the proposed etiolog-
ical role for TGF-181 in a variety offibrotic and inflammatory
disorders. The transgenic model may also provide an appro-
priate paradigm for testing therapeutic interventions aimed at
neutralizing the detrimental effects of this important cyto-
Aberrant expression oftransforming growth
The association of aberrant expression of peptide growth
factors with chronic diseases, particularly those involving
inflammation and tissue damage resulting in fibrotic lesions, is
well established (1). Of the growth factors involved in these
disease processes, transforming growth factor /31 (TGF-,B1)
appears to have a particularly important role in the fibrotic
process (1, 2). Three isoforms, TGF-/31, TGF-/32, andTGF-033,
have been cloned from cDNA libraries derived from mam-
malian cells (3). TGF-31 is the best characterized of the three
isoforms and generally regarded as the prototype oftheTGF-f3
family. TGF-,31 is biologically active as a 25-kDa homodimer
linked by disulfide bonds and is a highly conserved molecule
whose specific high-affinity receptors are present on essen-
tially all cells (3-5). TGF-,31 is present in most tissues, with the
highest concentrations found in bone and blood platelets (6).
TGF-/31 is stored in and secreted from a variety of cell types
as a biologically inactive, latent high molecular weight complex
incapable of binding to its receptors. Latent TGF-,B1 remains
inactive under neutral conditions but can be activated in vitro
by acid, alkali, urea, heat, or proteases (7). The latent complex
is activated extracellularly in vivo by a mechanism that is not
fully characterized (8). The conversion of the latent complex
to biologically active forms may constitute an important reg-
ulatory mechanism by which TGF-31 activity is controlled in
During carbon tetrachloride-induced rat liver fibrosis, in-
creased expression of TGF-31 occurs early and is maintained
throughout the fibrotic process (9). The concentration of
TGF-31 is also increased in other models ofexperimental liver
fibrosis, such as murine schistosomiasis and following hepatic
irradiation, and in cirrhotic human livers (10-12). These
results suggest thatTGF-p1 acts in vivo as a powerful stimulus
for collagen formation during chronic tissue injury. We there-
fore postulated that the presence ofmatureTGF-p1in the liver
may have initiated and maintained increased proliferation and
collagen synthesis in the mesenchymal cells by an autocrine
mechanism throughout the fibrotic process.
To specifically address the question of the potential etio-
logical role ofTGF-p1 in hepatic fibrosis, we have generated
a transgenic mouse model in which mature TGF-fl is over-
expressed in the liver. Our results demonstrate that overex-
pression of mature TGF-p1 in the liver causes multiple tissue
lesions, including hepatic fibrosis and extensive glomerulone-
MATERIALS AND METHODS
Fusion Gene Construction and Production of Transgenic
Mice. A 4.7-kbp Not I-Sal I vector-free DNA fragment (Fig.
1) was microinjected into the nuclei ofone-cell mouse embryos
obtained from mating of hybrid (C57BL/6J x CBA)F1 mice.
A 1.5-kbp modified porcine TGF-31 cDNA was inserted into
aKpn I site located downstream of a 5' flanking region of the
mouse albumin gene that contains the enhancer (located
between kbp -10.4 and -8.5) and the proximal promoter (13).
Prior to insertion, Cys223 and Cys225 codons in the porcine
TGF-,B1 cDNA were replaced with serine codons, resulting in
preferential secretion ofthe mature form ofTGF-/31 (14). The
transgene also contained 625bp ofthe 3' flanking region ofthe
human growth hormone gene, harboring the 3' untranslated
region of exon 5 and the polyadenylylation signal (15). The
transgenic mice were generated as described (16) and identi-
fied by Southern analysis of tail DNA (16) using a nick-
translated Sac I probe derived from the construct (Fig. 1).
RNA Preparation and Analysis. Total RNA was isolated
from mouse tissues by guanidinium isothiocyanate extraction
and cesium chloride centrifugation (17). Poly(A)+ RNA was
isolated by oligo(dT)-cellulose chromatography (18). For
Northern blot analysis, 5 ,ug of Poly(A)+ RNA was electro-
phoresed through a 0.9% agarose gel containing 12% form-
aldehyde, transferred to a filter, probed with nick-translated
porcine TGF-/31 cDNA (Sac I fragment in Fig. 1) and mouse
TGF-31 cDNA, and washed (19). Other probes were a 321-bp
cDNA from the last exon of the mouse a(I) collagen gene
Abbreviation: TGF-,3, transforming growth factor ,3.
§To whom reprint requests should be addressed at: National Cancer
Institute, Building 37, Room 3C28, 37 Convent Drive MSC4255,
Bethesda, MD 20892-4255.
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked "advertisement" in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
Proc. Natl. Acad. Sci. USA 92 (1995)
Not I (Xb/Nh)
free DNA fragment that was microinjected into the nuclei of one-cell
mouse embryos obtained from mating of hybrid (C57BL/6J x CBA/
J)F,mice. Xb, Xba I; Nh, Nhe I; Alb, albumin; Pr, promoter; hGH, 3'
flanking region of the human growth hormone gene.
Schematic representation of the 4.7-kbp Not I-Sal I vector-
(kindly provided by B. de Crombrugghe, Texas Medical Cen-
ter, Houston) and a probe for the 18S rRNA (American Type
Culture Collection), used as internal standard. Typically,
ethidium bromide was added to the RNA samples prior to
electrophoresis to enhance staining and to assess the loading
and the transfer efficiency of each RNA sample.
Histology and Immunohistochemistry. Tissues were fixed in
buffered 10% formaldehyde, embedded in paraffin, sectioned,
and stained with hematoxylin and eosin, Gomori's silver
reticulin, and Masson's trichrome by standard methods. Im-
munohistochemical staining for TGF-,31 was done on 5-pum
deparaffinized sections with an indirect immunoperoxidase
antiserum detection protocol (Elite kit; Vector Laboratories)
(9). TGF-j31 was detected by the rabbit polyclonal LC(1-30)
Proliferation of hepatocytes was assessed by incorporation
of 5-bromo-2'-deoxyuridine (BrdUrd, 100 mg/kg given i.p. 2
hr before sacrifice) with a diaminobenzidine staining proce-
dure (Amersham cell proliferation kit).
Measurement ofPlasmaTGF-.81.Plasma TGF-31 was mea-
sured with an ELISA kit (Genzyme). The procedure measures
both the mature and latent forms of TGF-,B1.
genic lines (2, 18, 34, and 25) expressed high levels of the
Alb/TGF-f31 transcripts whereas line 53 displayed a low level
(Fig. 2A). Expression of the transgene was restricted to the
liver (shown for line 25 in Fig. 2B). Although some hetero-
geneity in the expression of theTGF-,l31 protein was observed
in the liver of the transgenic mice, the protein was abundantly
expressed in virtually all the hepatocytes (Fig. 3A and B). The
expression ofthe Alb/TGF-f31 transgene was observed in both
male and female offspring of all five transgenic lines with the
exception of line 25, in which only the male progeny expressed
Expression of the Alb/TGF-181 Transgene and Hepatic
Fibrosis. The local consequences of expressing the Alb/
TGF-,B1 transgene in the liver were qualitatively the same in
all the mouse lines: hepatic fibrosis and increased mitotic
activity of the hepatocytes (Fig. 4; Fig. 3 C-F). The hepatic
fibrosis was accompanied by increased expression of type I
collagen mRNA (Fig. 4) and increased reticulin staining (Fig.
3 C andD). The most advanced hepatic fibrosis was seen in line
25, and there was an apparent association between the level of
transgene expression and the expression of type I collagen
(Fig. 4) as well as the extent of hepatic fibrosis. There was,
however, an age-related decrease in the expression ofboth the
transgene and type I collagen (Fig. 4). During the first 4-6
weeks of life all the males in line 25 expressed high levels of
transgene mRNA. This is also the period when "20% of the
animals died ofrenal failure (see below). Also, during this time
period the most advanced fibrotic lesions were observed in the
liver. At late time points (see weeks 9 and 12 in Fig. 4) the
expression of the transgene was significantly reduced but
considerable heterogeneity in transgene expression was seen.
The fibrotic process was characterized by deposition of
collagen around individual hepatocytes andwithin the space of
Disse in a radiating linear pattern similar to that seen in
alcohol-induced fibrosis (22). However, a fully developed cir-
Generation of Alb/TGF-B1l Transgenic Mice and Expres-
sion of the Transgene. To achieve high and sustained local
concentration of mature TGF-31 in the liver, we wished to (i)
target the expression of the TGF-,B1 transgene to the hepa-
tocytes, thereby providing the needed cell mass to produce the
required level of the molecule, and (ii) bypass the complex
activation of the latent form of the TGF-31 by generating
mature TGF-f31 directly from the transgene. These two ob-
jectives were achieved by generating transgenic mice bearing
a fusion gene consisting of the mouse albumin enhancer/
promoter and porcine TGF-,1l cDNA (Alb/TGF-,B1) in which
the Cys223 and Cys225 codons had been replaced with serine
codons (Fig. 1). Site-directed mutagenesis ofthe two cysteines,
located in the proregion of the pre-pro form of TGF-131,
generates a [Ser223, Ser225] TGF-j31 molecule that is mostly in
the active form when secreted (21). Mature TGF-31 was se-
creted when the Alb/TGF-,B1 construct was transfected into
human hepatoma cell line Hep G2, which supports the albumin
enhancer/promoter (data not shown).
We anticipated that successful expression of the Alb/
TGF-31 transgene in the liverwould increase plasma TGF-,31.
A further objective was therefore to determine whether in-
creased plasma levels of mature TGF-,1 resulted in extrahe-
patic tissue lesions. Founder mice were generated by injecting
one-cell embryos with the construct (Fig. 1) and used to es-
tablish five lines ofAlb/TGF-f31 transgene mice (nos. 2, 18,25,
34, and 53). Screening ofthe transgenic mice was done by both
Southern blots of DNA isolated from mouse tissues and by
PCR (data not shown). Northern blot analysis of hepatic
expression of the TGF-,B1 transgene in seven lines of the
Alb/TGF-f31 transgenic mice showed that four of the trans-
" e c
mouse lines. Northern blot analysis used liver poly(A)+ RNA (5 ,ug)
and a nick-translated mouse TGF-(31 cDNA. RNA from Hep G2 cells
transfected with the Alb/TGF-f31 construct was also analyzed. Neg.
Cont., negative control (mRNA from control liver). (B) Organ
distribution oftheAlb/TGF-,B1 transgene expression invarious mouse
tissues (line 25). Northern blot analysiswas performed as inA. The blot
was probed first with nick-translated mouse TGF-31 cDNA (Upper)
and then with a probe (Fig. 1) spanning the 3' end of the porcine
TGF-,B1 cDNA and the hGH part of the transgene (Lower).
(A) Transgene expression in seven of the Alb/TGF-131
CellBiology:Sanderson et al.
Cell Biology: Sanderson et al.
genic mouse livers. (A) Control mouse liver (age, 8weeks) stainedwith
LC(1-30) antibody. (B) Mature TGF-f31 in transgenic line 25 (age, 6
weeks) detected by staining with LC(1-30) antibody. (C and D)
Reticulin staining of control and transgenic line 25 mouse liver (age,
6 weeks), respectively. (E and F) BrdUrd-labeled hepatocytes in liver
from 8-week-old) control mouse and 8-week-old transgenic mouse of
line 25, respectively. (X45.)
Histological and immunohistochemical analyses of trans-
rhosis was rarely seen in mice, possibly due to the limited
lifespan of line 25 as a consequence of renal failure (see
below). The other prominent characteristic seen in the livers
of the Alb/TGF-,B1 mice was increased mitotic activity (Fig. 3
E and F). Also, apoptotic hepatocytes in the transgenic mice
were frequently seen (data not shown), consistent with find-
ings byOberhammer etal. (23) that intravenous administration
of mature TGF-31 in rats can induce apoptotic lesions in the
Plasma Levels ofTGF-j3 in Alb/TGF-131 Transgenic Mice.
We had anticipated that successful expression of the Alb/
TGF-f31 transgene in the liver would lead to increased plasma
levels ofthe mature form and, to a lesser extent, the latent form
of TGF-31. This was of considerable interest, since intrave-
nous administration of mature TGF-,l31 had been shown to
induce a variety of tissue lesions in rats and rabbits (24). The
mean plasma level ofTGF-,31 in the transgenic lines at 3 weeks
ofage ranged from 40 ng/ml in line 25 to 9 ng/ml in line 2 (Fig.
SA), compared with S ng/ml in aged-matched controls. The
postnatal time course ofTGF-f31 plasma levels in control mice
and in line 25 is shown in Fig. SB. The finding of high plasma
levels of TGF-,B1 (37 ng/ml) in 2-week-old control mice was
unanticipated. Apart from the 2-week time point, however, the
plasma levels of TGF-,1l in the control mice remained steady
at 5 ng/ml over the 12-week observation period (Fig. SB). As
expected the TGF-31 plasma levels in line 25 were consistently
higher than those seen in the age-matched controls (Fig. SB).
The difference in the TGF-,31 plasma levels between line 25
and controls was most pronounced during the first 6 weeks of
postnatal life, after which time the plasma levels of TGF-,B1
were only 2-3 times higher in the transgenic mice. The method
used to assess plasma TGF-f31 does not differentiate between
mature and latent forms of the molecule.
Extrahepatic Lesions in Alb/TGF-131 Transgenic Line 25.
The most dramatic phenotype observed in the Alb/TGF-,1
transgenic mice was seen in line 25 and consisted ofboth acute
and chronic renal failure. Approximately 20% of the male
mice in line 25 died of acute renal failure before the age of 12
weeks. The mice destined to enter into the acute phase of renal
failure were smaller than both the nontransgenic and the other
transgenic littermates. These mice also showed early signs of
ascites and proteinuria (25). Microscopically, the kidney lesion
(Fig. 6 A-C) was characterized by features of end-stage
glomerulonephritis with uniform involvement of all glomeruli;
there was accumulation of glomerular extracellular matrix and
thickening of the glomerular basement membrane as well as
prominent capsular epithelial proliferation with formation of
capsular crescents (Fig. 6C). All line 25 male mice examined
after 3 weeks of age manifested these lesions in varying degree
of severity, and by the age of 1 year all the male mice had
.Z .w . l ..
transgene and al(I) collagen expression in Alb/TGF-,B1 transgenic
mouse livers (line 25). Northern blot analysis for transgenic TGF-31
was performed as described in Materials andMethods and in the legend
to Fig. 2. The al(I) collagen probe was a 321-bp sequence from the
last exon of the mouse procollagen al(I) gene. Poly(A)+ RNA from
1-week-old mice was isolated from pooled livers (four animals). For
the other time points each lane represents Poly(A)+ RNA from a
single animal. C, control [poly(A)+ RNA from 8-week-old control
Postnatal time course (1, 2, 3, 9, and 12 weeks) of TGF-l31
3-week-old control mice and transgenic mice of lines 2, 18, 34, and 25.
(B) Postnal time course of TGF-f31 plasma levels (mean ± SE, n =
5-12) in control mice (0) and in transgenic line 25 (0).
(A) TGF-,Bl plasma levels (mean
± SE, n
= 7) in
Proc. Natl. Acad Sci. USA 92(1995)
Proc. Natl. Acad. Sci. USA 92 (1995)
6-week-old mouse kidney. (B) Kidney section from 6-week-old trans-
genic mouse showing extensive glomerulonephritis. (C) Higher mag-
nification of the diseased kidney shown in B illustrates advanced
glomerulonephritis characterized by hypocellularity of the glomerulus
and crescent formation. (D) Extensive arteritis in the heart of 5-week-
old transgenic mouse. (E) Myocardium from control 5-week-old
mouse. (F) Myocardium from 5-week-old transgenic mouse showing
myocarditis characterized by extensive infiltration of inflammatory
cells and fibrosis. (G) Pancreas from 3-week-old control mouse. (H)
Pancreas from 3-week-old transgenic mouse showing mild fibrosis and
atrophic islets of Langerhans. (I) Testis from 4-week-old control
mouse. (J) Testis from 4-week-old transgenic mouse showing atrophic
seminiferous tubules. (Hematoxylin and eosin staining of formalin-
fixed sections; x 180 in C, x45 in all other panels).
Extrahepatic lesions in transgenic line 25 mice. (A) Control
developed chronic glomerulonephritis. Among the other Alb/
TGF-131 transgenic lines only line 18 showed a mild form of
renal lesions (data not shown).
Other extrahepatic lesions were observed in line 25. These
included extensive arteritis affecting small and medium-size
arteries in both the heart and the kidney and involving all layers
of the vessel in the inflammatory reaction (Fig. 6D). Also,
extensive myocarditis and cardiac fibrosis were seen (Fig. 6 E
and F). Atrophic changes were observed in the pancreas and
testis in those mice in line 25 that had high plasma TGF-,B1 and
as a rule died of renal failure before the age of 6 weeks (Fig.
6 G-J). The pancreas in line 25 mice was characterized by mild
fibrosis and small islets ofLangerhans (Fig. 6H). However, the
small and atrophic islets of Langerhans in the transgenic mice
were capable of expressing insulin, glucagon, and somatostatin
(data not shown). Also, blood glucose levels were within
normal limits in the transgenic mice (data not shown). The
atrophic changes in testis were illustrated by thickened tubular
basement membranes; the seminiferous tubules were small
and farther apart, and the interstitial cells ofLeydig were more
prominent (Fig. 6 I and J) than in age-matched control mice.
In addition, line 25 mice had hypoplastic bone marrow and
mild anemia (data not shown).
The possible etiological role of TGF-,B1 in numerous disease
processes-in particular, the chronic inflammatory fibrotic
disorders-is well recognized (1, 2). We have generated a
transgenic mouse model inwhich mature TGF-,1 is selectively
expressed in the hepatocytes, in order to critically examine the
role of TGF-,B1 in hepatic fibrosis as well as the extrahepatic
effect of this cytokine. The transgene was designed to express
mature TGF-,B1 in hepatocytes by employing the albumin pro-
moter/enhancer and by converting Cys223 and Cys225 in the
TGF-,1 propeptide to serines (13, 14). All four transgenic
lines (lines 2, 18, 34, and 25; Fig. 2) expressing high levels of
the transgene showed increased hepatic fibrosis. Line 25 had
both the highest expression ofthe transgene in the liver and the
highest TGF-f31 plasma levels (Figs. 2 and 4). There were two
distinct patterns of hepatic fibrosis in line 25. The most
extensive hepatic fibrosiswas seen duringthe first 10-12weeks
of life, at which time high expression of the transgene was
accompanied by a high level of type I collagen expression in
the liver (Fig. 3). The fibrotic process was characterized by
deposition of collagen around individual hepatocytes and
within the space of Disse in a radiating linear pattern similar
to that seen in alcohol-induced fibrosis (22). Fully developed
cirrhosiswas rarely seen in the mice, possibly due to the limited
lifespan of line 25 as a consequence of renal failure. The
second pattern of hepatic fibrosis was seen in the mice that
survived beyond 12 weeks of age and was characterized by
chronic low-grade fibrosis that correlated well with the lower
level of transgene expression. However, the mice displaying
this chronic low-grade fibrosis were highly sensitive to chem-
ical induction of liver cirrhosis (V.F. and S.S.T., unpublished
data). The other characteristic effect on hepatic cell biology
excreted by theAlb/TGF-031transgene was increased mitotic
activity and the appearance of apoptotic cells (Fig. 3). This is
not unexpected, since intravenous administration of mature
TGF-,B1 in rats can induce apoptotic lesions in the liver (23).
Several extrahepatic lesions were observed in the Alb/
TGF-,B1 mice and, similar to the liver, the most pronounced
lesions were observed in line 25. Increased plasma TGF-131,
particularly during the first 6 weeks of life, was strongly
associated with the appearance of extrahepatic lesions. The
most striking extrahepatic lesion, observed in line 25 and to a
lesser extent in line 18, involved the kidney. Chronic glomer-
ulonephritis was seen in all line 25 male mice by .3 weeks of
age, and 20% of these mice died of end-stage renal disease by
12 weeks of age. The kidney lesions involved primarily the
glomeruli, but interstitial fibrosis and vacuolation ofproximal
tubular epithelial cells were also seen (for detailed description
ofthe renal lesion in the Alb/TGF-31 transgenic mice, see ref.
25). The critical role of TGF-j31 in the development of
glomerulonephritis induced experimentally by immunological
mechanisms hasbeen demonstrated (1, 26). Our results further
highlight the critical role ofTGF-,B1 in glomerular inflamma-
tory diseases by demonstrating that external production with
elevated plasma levels of mature TGF-f31 can precipitate
massive glomerulonephritis and renal failure.
Other inflammatorylesions seen in theAlb/TGF-31 line 25
mice involved the heart and small to medium-size arteries in
CellBiology:Sanderson et al.
2576 Download full-text
Cell Biology: Sanderson et al.
both the kidney and heart (Fig. 6 D-F). The inflammatory
reaction in the vascular lesions involved all layers of the vessel,
similar to that seen in polyarteritis nodosa (27). Since clinical
manifestations of polyarteritis nodosa frequently feature both
myocarditis and glomerulonephritis, we hypothesize that ma-
ture TGF-f31 may be a common etiological agent in this com-
Like the other extrahepatic lesions, the atrophic changes in
testis and pancreas were seen only in line 25 and were strongly
associated with high plasma levels of TGF-,B1. Intravenous
administration of recombinant mature human TGF-31, 100
,tg/kg per day for 5 days, has been shown to cause a zone of
degeneration and vacuolation of beta cells at the periphery of
the islets and diffuse atrophy of acinar cells in rats (24). It
therefore seems likely that the atrophic changes observed in
both pancreas and testis are due to direct effects of TGF-p1.
The appearance of extrahepatic tissue lesions in the Alb/
TGF-,B1 transgenic mice was observed only in line 25 (and to
a lesser extent in line 18), in which high plasma TGF-,B1 was
found during the first 6 weeks of life. The plasma TGF-,B1 was
determined with an ELISA (see Materials and Methods) that
measures both mature and latent forms of cytokine. It is of
interest that the plasma levels of TGF-,31 at 2 weeks of age in
the nontransgenic littermates (for line 25 these are female
mice, since the transgene is expressed only in the male pro-
geny) are high but drop to control levels at the time ofweaning
(3 weeks; Fig. SB). In contrast to the transgenic littermates,
which already have both hepatic and extrahepatic lesions at 2
weeks, the nontransgenic progeny are entirely normal. These
results strongly indicate that most if not all of the TGF-,B1 in-
the plasma of the 2-week-old nontransgenic mice is in the
biologically inactive latent form, whereas the biologically
active mature TGF-,B1 constitutes the major portion of the
plasma TGF-g31 in the transgenic mice. The notion that the
mature transgenic TGF-,31 generated in the liver is responsible
for both the hepatic and extrahepatic lesions is further sup-
ported by data obtained in rats after acute and chronic
intravenous administration of recombinant human TGF-l31
(24): high doses of recombinant mature TGF-31 produced a
spectrum oflesions in multiple target tissues including all those
seen in the present study.
In summary, we have developed a transgenic mouse model
in which expression of mature TGF-f31 is targeted to the liver.
The phenotypes generated by this model include both fibrotic
and inflammatory lesions characteristic for a number of im-
portant human diseases. Since mature TGF-f1 is the sole
etiological agent responsible for the observed disease pheno-
types, this animal model aids in elucidating the mechanism by
which TGF-31 initiates these disease processes and in testing
therapeutic modalities to prevent the detrimental effects of
this important cytokine.
Border, W.A. & Nobel, N.A. (1994) N. Engi. J. Med. 331,
Roberts, A. B. & Sporn, M. B. (1993) Growth Factors 8, 1-9.
Massague, J. (1990) Annu. Rev. Cell Biol. 6, 597-641.
Moses, H. L., Yang, E. Y. & Pietenpol, J. A. (1990) Cell 63,
Lin, H. Y., Wang, X. F., Ng, E. E., Weinberg, R. A. & Lodish,
H. F. (1992) Cell 68, 775-785.
Assoian, R. K., Frolik, C. A., Roberts, A. B., Miller, D. M. &
Sporn, M. B. (1983) J. Biol. Chem. 258, 7155-7160.
Lyons, R. M. & Moses, H. L. (1990) Eur. J. Biochem. 187,
Lyons, R. M., Gentry, L. E., Purchio, A. F. & Moses, H. L. (1990)
J. Cell Biol. 110, 1361-1367.
Nakatsukasa, H., Nagy, P., Evarts, R. P., Hsia, C.-C., Marsden, E.
& Thorgeirsson, S. S. (1990) J. Clin. Invest. 85, 1833-1843.
Anscher, M. S., Crocker, I. R. & Jirtle, R. L. (1990) Radiat. Res.
Czaja, M. J., Weiner, F. R., Flanders, K. C., Giambrone, M.-A.,
Wind, R., Biempica, L. & Zern, M. A. (1989) J. Cell Biol. 108,
Castilla, A., Prieto, J. & Fausto, N. (1991) N. Engl. J. Med. 324,
Pinkert, C. A., Ornitz, D. M., Brinster, R. L. & Palmiter, R. D.
(1987) Genes Dev. 1, 268-276.
Samuel, S. K., Hurta, R. A. R., Kondaiah, P., Khalil, N., Turley,
E. A., Wright, J. A. & Greenberg, A. H. (1992) EMBO J. 11,
Low, M. J., Hammer, R. E., Goodman, R. H., Habenere, J. F.,
Palmiter, R. D. & Brinster, R. L. (1985) Cell 41, 211-219.
Murakami, H., Sanderson, N. D., Nagy, P., Marino, P. A., Mer-
lino, G. & Thorgeirsson, S. S. (1993) CancerRes. 53, 1719-1723.
Schweizer, J. & Goerttler, K. (1980)Eur. J. Biochem. 11, 243-249.
Aviv, H. & Leder, P. (1972) Proc. Natl. Acad. Sci. USA 69,
Evarts, R. P., Nakatsukasa, H., Marsden, E. R., Hu, Z. & Thor-
geirsson, S. S. (1992) Mol. Carcinog. 5, 25-31.
Flanders, K C., Thompson, N. L., Cissel, D. S., van Obberghein-
Shilling, E., Baker, C. C., Kass, M. E., Ellingsworth, L. R., Rob-
erts, A. B. & Sporn, M. B. (1989) J. Cell Biol. 108, 653-660.
Brunner, A. M., Marquardt, H., Malacko, A. R., Lioubin, M. N.
& Purchio, A. F. (1989) J. Biol. Chem. 264, 13660-13664.
Desmet, V. J. (1985) Acta Med. Scand. 703 (Suppl.), 111-126.
Oberhammer, F. A., Pavelka, M., Sharma, S., Tiefenbacher, R.,
Purchio, A. F., Bursch, W. & Schulte-Hermann, R. (1992) Proc.
Natl. Acad. Sci. USA 89, 5408-5412.
Terrell, T. G., Working, P. K, Chow, C. P. & Green, J. D. (1993)
Int. Rev. Exp. Pathol. 34, 43-66.
Kopp, J. B., Sanderson, N., Factor, V., Nagy, P., Mozes, M.,
Bottinger, E., Klotman, P. E. & Thorgeirsson, S. S. (1995) J. Clin.
Invest., in press.
Okuda, S., Languino, L. R., Ruoslahti, E. & Border, W. A. (1990)
J. Clin. Invest. 86, 453-462.
Anderson, W. A. D., ed. (1971) Pathology (Mosby, St. Louis).
Proc. Natt Acad ScL USA 92(1995)