Oncogene (1997) 15, 1115 ± 1120
 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00
SHORT REPORT
Keratinocyte growth factor expression in hormone insensitive prostate
cancer
Hing Y Leung, Pyush Mehta, Lisa B Gray, Anne T Collins, Craig N Robson and David E Neal
Department of Surgery, The Medical School, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne
NE2 4HH
Cellular interactions between stroma and epithelium are
important in the growth and proliferation of prostate
cancer. Peptide growth factors may facilitate the
progression of prostate cancer as autocrine and/or
paracrine factors. Keratinocyte Growth Factor (KGF
or FGF7) has a di€erentiative and proliferative e€ect on
the epithelium of the developing rat prostate. We
investigated if KGF may act as a paracrine agent in
human prostate cancer and examined the expression of
KGF and Fibroblast Growth Factor Receptors (FGFRs)
(IIIb and IIIc isoforms of the FGFR1 and FGFR2
genes). Sixty-®ve percent (11 out of 17 informative
cases) of prostate cancers (CaP) expressed KGF mRNA
by RT ± PCR, while KGF expression was not detected in
benign prostatic hyperplasia (BPH) (n=6). Upregulation
of KGF expression was related to hormone insensitive
tumours (P50.05). Tumour grade and stage were not
associated with KGF expression. The source of KGF
expression was further characterised using an in vitro
primary culture model, showing its restriction to the
prostatic stroma. The FGFR1IIIb isoform was expressed
in all cases of prostate cancer (n=17), and FGFR1IIIc
mRNA was not detected. In the BPH group, FGFR1IIIb
transcripts were detected in four out of six cases.
FGFR2IIIb expression was detected in ®ve of six cases
of BPH and twelve out of seventeen (71%) cases of
prostate cancer. In CaP, though not reaching statistical
signi®cance, the persistence of FGFR2IIIb expression
appeared to be associated with hormone insensitive
tumours (P=0.052). FGFR2IIIc expression was present
in eleven of seventeen tumours but was absent in all six
cases of BPH. Functional assessment of recombinant
KGF in a proliferation assay demonstrated a mitogenic
e€ect of up to 100% on cultured prostatic epithelial
cells.
Keywords: kGF; FGF; FGFR isoform; alternative
splicing; prostate cancer; paracrine loop
Cancer of the prostate (CaP) is a common malignancy
in males, ranking as the ®fth most common cancer in
the world. Each year in England and Wales, 9000 men
die of the disease. Patients with prostate cancer often
present with locally advanced and/or metastatic
disease, and their prognosis remains poor. Fifty
percent of men with locally advanced prostate cancer
have lymph node metastases on presentation and only
half will survive 3 years. Only younger men with
Correspondence: HY Leung
Received 6 December 1996; revised 7 May 1997; accepted 7 May
1997
localised disease are suitable for radical prostatectomy
and the mainstay of management of advanced disease
is androgen deprivation either by medical means or
orchiectomy. As many as 30% will not respond to
androgen ablation in the ®rst instance and, even
among the hormone sensitive tumours, an increasing
proportion will subsequently progress and develop
hormone insensitivity. At the end of the third year
following diagnosis, only a third of the locally
advanced disease and ten percent of metastatic
tumours have not progressed. Peptide growth factors
and their high anity tyrosine kinase receptors may
contribute towards to the progression of prostate
cancer (reviewed by Byrne et al., 1996).
In man, the family of Fibroblast Growth Factors
(FGFs) consists of nine cloned ligand genes, FGF1 to
FGF9 (Klagsbrun, 1989; Tanaka et al., 1992;
Miyamoto et al., 1993), and the receptor family
(FGFR) comprises four members, FGFR1 to 4
(Partanen et al., 1992). More recently, FGF10 in rat
(Yamasaki et al., 1996) and four FGF-homologous
factors in human have been cloned (Smallwood et al.,
1996). The existence of multiple ligands and receptors
allows interaction between a single FGFR and multiple
FGFs, and between di€erent FGFR monomers
through heterodimerisation following activation by
FGF (reviewed by Leung et al., 1994).
The utilisation of splice variants in FGFR expression is complex and is best documented for the FGFR1
and FRGF2 genes (Eisemann et al., 1991; Crumley et
al., 1991). Functional implications of the di€erent
splice variants are not fully understood, but FGFR
isoforms di€ering in their extracellular domains have
distinct ligand binding repertoires. For instance, usage
of two alternative splice exons in the second half of the
third immuno-globulin loop in the extracellular domain
of FGFR2 results in di€erential ligand binding, with
the IIIb isoform binding KGF at high anity but not
FGF2 (bFGF), and vice versa for the IIIc isoform
(Miki et al., 1992).
Fibroblast growth factors have been implicated in
the biology of the prostate. FGF1 (aFGF) and FGF2
(bFGF), but not KGF, are mitogenic to cultured
human prostatic stromal cells derived from BPH,
whilst cultured benign epithelial cells respond to
FGF1 and KGF, but not to FGF2 (Story et al.,
1994). Cellular interactions between stroma and
epithelium are important in the embryogenesis of the
prostate. Cunha et al. (1983) showed that under
androgen stimulation, developing epithelium derived
from the urogenital sinus, though itself lacking
androgen receptors, was able to di€erentiate and
proliferate to mature prostatic epithelium when
KGF and FGFR in human prostate cancer
HY Leung et al
1116
combined with androgen receptor expressing stromal
cells. More recently, KGF was identi®ed as an
androgen-induced paracrine factor derived from
prostatic stroma and found to have di€erentiating
and proliferative e€ects on the epithelium of the
developing rat prostate (Yan et al., 1992; Surgimura
et al., 1996). The gene product of FGFR2 was found in
rat benign prostatic epithelium; the presence of the
exon IIIb was suggested by the strong binding to KGF
and FGF1 but not to FGF2. During the development
of prostate cancer in a rat model, the FGF binding
characteristics of prostatic epithelium changed to a
higher anity for FGF2 and FGF1, but not for KGF.
This resulted from a switch of FGFR expression from
the IIIb to the IIIc isoform. Such an alteration in the
receptor responsiveness to FGF2 occurred along with a
simultaneous over-expression of FGF2 in the malignant prostate (Yan et al., 1993), hence constituting a
potential autocrine loop action. These ®ndings support
the hypothesis that FGFs and their high anity
receptors may contribute to the tumorigenesis of
prostate cancer in the rat.
We tested the hypothesis that KGF may act in a
paracrine fashion in human prostate cancer and
examined if the expression status of KGF and FGFRs
(IIIb and IIIc isoforms of the FGFR1 and FGFR2
a
1
2
4
5
6
7
8
genes) in a panel of resected prostate cancer may relate
to the clinical progression of the disease. The cell type
responsible for KGF expression were further characterized using an in vitro primary culture model.
Functional assessment of KGF was performed by
proliferation assay using cultured prostatic epithelium.
Seventeen of the 18 prostate cancer cases and all six
cases of BPH were shown to have intact RNA
following ethidium bromide stained gel electrophoresis
(Figure 1a) and on RT ± PCR for GAPDH expression,
giving a band product of 300 bp (Figure 1b). Case 3 of
the prostate cancer samples had signi®cant degradation
of RNA and was excluded from the study. Of the 17
cases of prostate cancer, 11 (65%) showed KGF
expression, with a PCR product of 685 bp (Figure
1c). Northern blot analysis con®rmed the expression of
a 2.4 kb KGF transcript (Figure 1d). All six cases of
BPH were negative for KGF mRNA expression. KGF
expression was associated with hormone insensitive
prostate cancer. Seven out of ten KGF positive
tumours failed to respond to hormone manipulation
and showed no signi®cant fall in serum PSA levels,
while only one out of seven KGF negative tumours
were hormone insensitive (Fisher's exact test: P50.05).
The grade and stage of the tumours did not correlate
with KGF expression (Tables 1 and 2).
M
9
b
2
3
4
5
6
7
▲
28S —
1
18S —
8
10
9
11
12
13
▲
10
11
12
13
14
15
16
17
18
M
28S —
14
15
16
17
–
+
18
18S —
▲
c
d
1
2
3
4
5
6
7
8
9 10 11
12
13
14 15
16 17
18
▲
Figure 1 (a) ethidium bromide stained gel, (b) GAPDH expression, (c) KGF expression, (d) Northern blotting for KGF mRNA
expression. The single step acid-phenol/chloroform extraction method was employed. Oligo-dT based reverse transcription was
performed using the `ready to go kit'TM (Pharmacia Ltd, UK). The primer sequences were: KGF forward-CAAATGGATACTGACATGGATC, reverse-GCCATAGGAAGAAAAGTGGGCTGT; GAPDH forward-AGTCAACGGATTTGGTCGTA
and reverse-AAATGAGCCCCAGCCTTCT. PCR reactions were performed in 50 ml reactions containing 50 mM KCl, 10 mM
TrisCl pH 8.3, 1.5 mM MgCl2, 0.2 mM dNTPs, 1 unit of Taq polymerase (Bioline Ltd, UK). PCR protocols were optimised as
follows: KGF ampli®cation (30 cycles: 948C, 1 min; 558C 1 min, 728C 1 min); GAPDH ampli®cation (25 cycles 948C 30 s, 628C
20 s, 728C 30 s). A 700 bp KGF cDNA probe was provided by Dr S Tronick (Santa Cruz Biotechnology, Inc., Santa Cruz,
California, USA), and was radio-labelled using random primed labelling mixture (Finch et al., 1989). Total RNA from a KGF
expressing tumour on RT ± PCR (+) and a KGF negative tumour (7) were used for Northern blotting. The arrow signi®es the
expected 2.4 kb KGF transcript
KGF and FGFR in human prostate cancer
HY Leung et al
Using in vitro cultured prostatic epithelial and
stromal cells, KGF mRNA expression detected by
RT ± PCR was shown to be restricted to prostatic
stromal cells (Table 3). Stromal cells derived from both
malignant and benign prostate glands expressed KGF
transcripts. Primary cultured prostatic epithelial cells
from both BPH and CaP showed no KGF expression.
Furthermore, established prostatic epithelial cell lines,
including LNCaP cancer cells and SV40 transformed
epithelial PNT1a cells, did not express KGF.
All 17 cases of prostate cancer were positive for
FGFR1IIIb mRNA expression (Figure 2a and b). In
the six cases of BPH, four expressed FGFR1IIIb at
low levels (Table 2). FGFR1IIIc transcripts were not
detected in either BPH or CaP. FGFR1IIIb was also
expressed in LNCaP and PNT1a cells (Table 3).
Twelve of 17 cases (71%) of prostate cancer expressed
FGFR2IIIb (Figure 2c). Five of the six cases of BPH
also expressed FGFR2IIIb isoform. FGFR2IIIc
transcripts were detected in 11 of 17 tumours, but
negative in all six cases of BPH (Figure 2d). The
expression of FGFR2IIIc isoform did not show any
association with hormone responsiveness, grade or
stage of the tumours. None of the hormone sensitive
tumours were FGFR2IIIb positive and FGFR2IIIb
expression appeared to persist in hormone insensitive
tumour. This association however did not reach
statistical signi®cance (Fisher's exact test: P=0.052).
No correlation was observed between KGF and
FGFR2IIIb expression, with FGFR2IIIb expression
in 6 of 7 KGF positive tumours and 6 of 10 KGF
negative tumours.
Human recombinant KGF at a concentration of
10710 mol/l enhanced proliferation of cultured prostatic
1
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Age
73
75
74
87
72
68
62
79
73
57
79
63
81
60
83
80
65
PSA
Grade (1 ± 3) T stage M stage (ng/ml)
3
3
3
2
1
2
2
3
3
3
2/3
3
3
2
2/3
3
2
T4
T3
T3/4
T2
T1
T1
T4
T4
T4
T4
T4
T4
T3
T1
T1
T1
T1
1
0
0
0
0
0
1
1
0
0
1
1
1
0
0
?
0
Table 2
Case
1
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
BPH/1
BPH/2
BPH/3
BPH/4
BPH/5
BPH/6
Table 3
Table 1 Clinico-pathological details
Case
epithelial cells by 50%. This mitogenic e€ect increased
further to over 100% when KGF concentrations were
1079 mol/l and over. At concentrations below 10711
mol/l, KGF did not have any demonstrable e€ect on
the growth of cultured prostatic epithelium. Human
recombinant FGF2 (bFGF) was included as a
comparative control in parallel experiments and did
380
141
87
52
4
72
78
17
50
17
302
16
11
18
4
Horm.
Status
R
S
R
S
S
S
R
R
R
R
R
R
S
S
S
S
S
Snap-frozen prostate chips were prepared following transurethral
resection of the prostate (TURP). Eighteen cases of prostate cancer
(CaP) and six cases of benign prostatic hyperplasia (BPH) were
included in this study. The mean age of patients with prostate
tumours was 72.4 years (range 57 ± 87 years). Digital rectal
examination revealed locally advanced disease in ten patients with
tumours invading through the prostatic capsule (TNM staging: T3
and T4 tumours); while seven patients had organ con®ned disease
(T1 and T2). The mean serum prostate speci®c antigen levels in these
two groups of patients with locally advanced and clinically organ
con®ned tumours were 127 ng/ml (range 17 ± 380) and 17.5 ng/ml
(range 4 ± 52 ng/ml) respectively. Grade 1, 2, 3 represent well,
moderate and poorly di€erentiated tumours respectively; M1
signi®ed bony metastasis on isotopic bone scan; hormone status as
judged by the response to hormone manipulation: R= no response,
S= signi®cant drop in serum prostate speci®c antigen levels
KGF and FGFR expression in BPH and CaP samples
FGF-7
R1-IIIb
R1-IIIc
R2-IIIb
R2-IIIc
++
7
++
++
++
++
++
+
+
++
++
++
7
7
7
7
7
++
+
+
++
+++
+
+
++
+
+
++
+
+++
+++
++
+
++
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
+++
+++
+++
+
7
+++
+++
+++
++
++
+
+
7
7
7
+
+
+
7
7
+
++
+
+
7
7
7
++
+
7
+++
+
++
+
7
7
7
7
7
7
+
7
7
+
+
+
7
7
7
7
7
7
7
+++
+
+++
++
+++
7
7
7
7
7
In vitro expression of KGF and IIIb isoform of FGFR1
and FGFR2
DU145
LNCaP
PNT1a
18 cultured epithelial cells
from BPH
18 cultured epithelial cells
from CaP
18 cultured ®broblast from
BPH
18 cultured ®broblast from
CaP
KGF
FGFR1IIIb
FGFR2IIIb
7
7
7
7
+
+
+
7
7
7
+
7
7
+
+
7
n.d.
n.d.
+
7
+
PNT1a, a SV40 transformed prostatic epithelial cell line derived from
normal prostate, was a gift from Professor N Maitland (University of
York, York, UK). The established prostatic cancer cell line LNCaP
and DU145 were obtained from the American Type Culture
Collection. Prostate tissues obtained following TURP was used for
primary culture to study the epithelial and stromal cells separately
(Collins et al, 1994). In brief, prostate chips were minced into 1 mm
cubes and incubated overnight in RPMI-1640 (Life Technologies,
Ltd, Scotland), containing 5% foetal calf serum (Life Technologies)
and collagenase type 1 at 200 IU (Lorne Laboratories, UK).
Following washings and repeated centrifugation (400 g for 20 min
each time), stromal and epithelial cells were separated to over 95%
purity (as assessed by immuno-¯orescence for cyto-keratin, vimentin
and a-actin (Collins et al, 1994), and were plated as individual
cultures for subsequent analysis. Stromal cells were maintained in
RPMI-1640 medium with 10% FCS, penicillin (100 units/ml) and
streptomycin (100 mg/ml). Epithelial cells were cultured in complete
WAJC-404 serum free medium supplemented with insulin (2.5 mg/
ml), dexamethasone (1 mM), epidermal growth factor (10 ng/ml),
bovine pituitary extract (25 mg/ml), cholera toxin (10 ng/ml), heparin
(25 mg/ml), penicillin (100 units/ml) and streptomycin (100 mg/ml).
Cultured stromal and epithelial cells were harvested for experimentation within the ®rst two passages
1117
KGF and FGFR in human prostate cancer
HY Leung et al
1118
a
1
2
b
c
4
b
c
5
b
b
c
b
6
c
b
c
b
7
c
8
b
c
9
b
c
10
b
c
b
c
▲
c
11
b
12
c
b
13
c
b
14
15
16
c
b
17
c
b
18
c
b
c
▲
b
1
2
b
c
7
b
1
10
2
11
4
12
5
13
6
14
b
8
16
b
5
c
9
c
b
14
b c
7
15
c
8
c
13
b c
c
4
b
6
c
10
c
15
b c
b
b
18
c
11
c
b
16
b c
17
b c
d
9
17
b
9
10
12
c
b
c
18
b c
1
2
3
4
5
6
7
11
12
13
14
15
16
17
8
18
Figure 2 FGFR mRNA expression (a) RT ± PCR for FGFR1IIIb and IIIc, (b) Southern analysis using FGFR1IIIb-speci®c
oligonucleotide probe on RT ± PCR products (subsequent hybridization with FGFR1IIIc-speci®c probe gave no signals, data not
shown), (c) FGFR2IIIb mRNA expression, (d) FGFR2IIIc mRNA expression. PCR primers: FGFR1IIIb forwardCATTCGGGGATTAATAGCTC, FGFR1IIIc forward-ACTGCTGGAGTTAATACCAC, FGFR1 reverse-GGAGTCAGCAGACACTGT; FGFR21IIIb forward-CACTCGGGGATAAATAGTT, FGFR2IIIc forward-ATTCTATTGGGATATCCTTT, FGFR2
reverse-ACTCGGAGACCCCTGCCA. PCR conditions were carried out as follows: FGFR1 IIIb and IIIc isoforms (30 cycles: 948C
1 min, 558C 20 s, 728C, 1 min); FGFR2 IIIb and IIIc isoforms (30 cycles: 948C 1 min, 558C 20 s, 728C, 30 specify.) Southern blots
were prepared using capillary transfer onto Hybond N+ membrane (Amersham International, Plc) and probed with a corresponding
internal oligonucleotide to the receptor isoforms for which the PCR was performed (FGFR1: IIIb-GGATGCGGAGGTGCTGACCCTGTTCAATGT and IIIc-CAAAGAGATGGAGGTGCTTCACTTAAGAAA; FGFR2: IIIb-ATATAGGGC AGGCCAACCAGTCTGCCT and IIIc-ACTCTGCATGGTTGACAGTTCTGC. Following pre-hybridization blocking, the ®lters were
hybridized at 508C for 2 h with the corresponding 32P-radiolabelled oligonucleotide probe in a bu€er containing 56SSC,
56Denhardt's solution, 0.5% sodium dodecyl sulphate, and 1 mg/ml salmon sperm DNA. The ®lters were then washed in
successive washes of 26SSC/0.1% SDS twice followed by a single wash in SSC/0.1% SDS once at 508C. Filters were exposed by
autoradiography for 3 to 4 h
KGF and FGFR in human prostate cancer
HY Leung et al
not show any mitogenic e€ect on the growth of the
cultured prostatic epithelial cells (Figure 3).
Using an in vitro primary culture model to separate
epithelial and stromal cells, prostatic stromal cells were
con®rmed to be the exclusive source of KGF
expression, in keeping ®ndings from other workers
(Stroy et al., 1994; Ittman and Mansukhani, 1997).
However, using in situ hybridization with a digoxigenin-labelled oligonucleotide probe, McGarvery and
Stearns (1995) detected KGF expression at low level
in the epithelium of BPH and at higher levels in the
high grade malignant epithelium. Despite repeated
experiments with RT ± PCR for KGF, we did not
observe any evidence of such `ectopic' KGF expression
in the prostatic epithelial cells, including primary
cultured epithelial cells from BPH and CaP, as well
as the established cell lines (LNCaP, DU145 and PC3),
and the transformed PNT1a cells (Table 3). Although
KGF expression was not detected in cDNA extracted
from BPH specimens, it was detected in cultured
stromal cells from both BPH and prostate cancer.
KGF expression in cultured stromal cells derived from
BPH may represent cellular responses to in vitro
culture conditions, similar to the upregulation of
KGF expression observed during wound healing and
active in¯ammation (Werner et al., 1992a).
KGF functions via its interaction with the tyrosine
kinase FGF receptors. We have demonstrated mRNA
expression of the IIIb isoform of both FGFR1 and
FGFR2 in the prostate. The FGFR2IIIb isoform has
been shown to bind to KGF at high anity and has been
previously referred to as KGFR (Werner et al., 1992b).
Comparing the sequence homology between the IIIb
exons of the human FGFR1 and FGFR2 genes, we and
others (Werner et al., 1992b) reasoned that FGFR1IIIb
is likely to interact with KGF to give rise to signal
transduction activities. We have therefore examined the
expression of FGFR1 and FGFR2 in our study. During
the course of this study, KGF was shown to interact with
the FGFR1IIIb isoform only extremely weakly (Ornitz
et al., 1996). Formation of heterodimers between
FGFR2IIIb and other FGFR isoforms that do not
interact with KGF directly at high anities may
nevertheless result in signal transduction activities.
FGFR1 expression in BPH, when compared to
normal prostate, has been shown to be upregulated
(Story et al., 1994 and Hamaguchi et al., 1995). In our
study, FGFR1 expression in CaP was at higher levels
than in BPH. Of the 17 informative tumour cases, all
expressed FGFR1IIIb. FGFR2IIIb expression was
detected in 12 of 17 cases of CaP and ®ve of six
cases of BPH. The co-expression of KGF and the IIIb
isoforms of FGFR1 and FGFR2 suggests a potential
paracrine loop. This was supported by our ®nding of a
mitogenic e€ect of KGF on cultured prostatic
epithelial cells. We have observed that FGFR was
expressed in cultured prostatic epithelium and KGF
1119
Figure 3 Proliferation assay using human recombinant KGF and
FGF2 on cultured prostatic epithelial cells. Using primary
cultured prostate epithelial cells from BPH, the e€ects of
exogenous recombinant human KGF (R&D) were assessed in a
thymidine incorporation experiment. Primary cultured prostatic
epithelial cells were trypsinized and plated at a density of 16104
cells per well in a 96 cell plate (Corning) with WAJC-404 serum
free medium supplemented with insulin (2.5 mg/ml), dexamethasone (1 mM), penicillin (100 units/ml) and streptomycin (100 mg/
ml) along with a range of concentrations of recombinant KGF.
All treatment conditions were performed in triplicate. Following a
period of culture for 3 to 4 days, [3H]Thymidine incorporation
was performed and DNA synthesis measured as described
previously (Collins et al., 1994). Human recombinant FGF2
(bFGF) was used in parallel as an internal control
had a signi®cant mitogenic e€ect on prostatic
epithelium. In addition, FGF2 did not have any
e€ect. This is in keeping with previous reports (Story
et al., 1994; Peehl et al., 1996) and further supports the
mutually exclusive nature of the IIIb and IIIc isoforms
at the functional level.
In summary, we have demonstrated signi®cant
upregulation of KGF expression in hormone resistant
prostate cancer and con®rmed its role as a potential
paracrine growth factor. Future work on the heterodimerisation of FGFR monomers and the consequences of their signal transduction activities will
allow a better understanding of the role of FGFs in
prostate cancer.
Abbreviations
FGF, ®broblast growth factor; FGFR, ®broblast growth
factor receptor; BPH, benign prostatic hyperplasia; CaP,
prostate cancer.
Acknowledgements
This project was funded by a grant from the Harker
Foundation, University of Newcastle upon Tyne. The
authors thank Wendy Robson for tissue collection and
Kate Gilmore for technical assistance.
References
Byrne RL, Leung HY and Neal DE. (1996). Br. J. Urol., 77,
627 ± 633.
Cunha GR. (1983). Recent. Prog. Horm. Res., 39, 559 ± 598.
Collins AT, Zhiming B, Gilmore K and Neal DE. (1994). J.
Endocrin., 143, 269 ± 277.
Crumley G, Bellot F, Kaplow JM, Schlessinger J, Jaye M and
Dionne CA. (1991). Oncogene, 6, 2255 ± 2262.
Eisemann A, Aha JA, Graziani G, Tronick SR and Ron D.
(1991). [published erratum appears in Oncogen 1991
6(12):2379]. Oncogene, 6, 1195 ± 1202.
KGF and FGFR in human prostate cancer
HY Leung et al
1120
Finch PW, Rubin JS, Miki T, Ron D and Aaronson SA.
(1989). Science, 2245, 752 ± 755.
Hamaguchi A, Tooyama I, Yoshiki T and Kimura H. (1995).
Prostate, 27, 141 ± 147.
Ittman M and Mansukhani. (1997). J. Urol., 157, 351 ± 356.
Klagsbrun M. (1989). Prog. Growth Factor Res., 1, 207 ± 235.
Leung HY, Hughes CM, Kloppel G, Williamson RCN and
Lemoine NR. (1994). Int. J. Oncol., 4, 1219 ± 1223.
McGarvey TW and Stearns ME. (1995). Exp. Mole.
Pathology, 63, 52 ± 62.
Miki T, Bottaro DP, Fleming TP, Smith CL, Burgess WH,
Chan AM-L and Aaronson SA. (1992). Proc. Natl. Acad.
Sci. USA, 89, 246 ± 250.
Miyamoto M, Naruo K-I, Seko C, Matsumoto S, Kondo T
and Kurokawa T. (1993). Mol. Cell. Biol., 13, 4251 ± 4259.
Ornitz DM, Xu J, Colvin JS, McEwen DG, MacArthur CA,
Coulier F, Gao G and Goldfarb M. (1996). J. Biol. Chem.,
271, 15292 ± 15297.
Partanen J, Vainikka S, Korhonen J, Armstrong E and
Alitalo K. (1992). Prog. Growth Factor Res., 4, 69 ± 83.
Peehl DM, Wong ST and Rubin JS. (1996). Growth
Regulation, 6, 22 ± 31.
Smallwood PM, Munoz-Saniuan I, Tong P, Macke JP,
Hendry SH, Gilbert DJ, Copeland NG, Jenkins NA and
Nathans J. (1996). Proc. Natl. Acad. Sci. USA, 93, 9850 ±
9857.
Story MT, Hopp KA, Moter M and Meirer DA. (1994).
Growth Factors, 10, 269 ± 280.
Sugimura Y, Foster BA, Hom YK, Lipschutz JH, Rubin JS,
Finch PW, Aaronson SA, Hayashi N, Kawamura J and
Cunha G. (1996). Int. J. Dev. Biol., 40, 941 ± 951.
Tanaka A, Miyamoto K, Minaminon N, Takeda M, Sato B,
Matsuo H and Matsumoto K. (1992). Proc. Natl. Acad.
Sci. USA, 89, 8928 ± 8932.
Werner S, Peters KG, Longaker MT, Fuller-Pace F, Banda
MJ and Williams LT. (1992a). Proc. Natl. Acads. Sci.
USA, 89, 6896 ± 6900.
Werner S, Duan DS, De Vries C, Peters KG, Johnson DE
and Williams LT. (1992b). Mol. Cell. Biol., 12, 82 ± 88.
Yamasaki M, Miyake A, Tagashira S and Itoh N. (1996). J.
Biol. Chem., 271, 15918 ± 15921.
Yan G, Wang F, Fukabori Y, Sussman D, Hou J and
McKeehan W. (1992). Biochem. Biophys Res. Commun.,
183, 423 ± 430.
Yan G, Fukabori Y, McBride G, Nikolaropolous S and
McKeehan WL. (1993). Mol. Cell. Biol., 13, 4513 ± 4522.