Content uploaded by li Xin
Author content
All content in this area was uploaded by li Xin on Jun 08, 2015
Content may be subject to copyright.
Expression of colony-stimulating factor 1 receptor
during prostate development and prostate
cancer progression
Hisamitsu Ide*, David B. Seligson
†
, Sanaz Memarzadeh
‡
, Li Xin*, Steve Horvath
§¶
, Purnima Dubey
‡
, Maryann B. Flick
储
,
Barry M. Kacinski
储
**
††
, Aarno Palotie
†¶
, and Owen N. Witte*
‡,‡‡
*Howard Hughes Medical Institute, Departments of †Pathology and Laboratory Medicine, ‡Microbiology, Immunology, and Molecular Genetics,
§Biostatistics, and ¶Human Genetics, University of California, Los Angeles, CA 90095; and Departments of 储Therapeutic Radiology,
**Dermatology, and ††Obstetrics and Gynecology, Yale University School of Medicine, New Haven, CT 06520-8040
Contributed by Owen N. Witte, September 4, 2002
Colony-stimulating factor-1 receptor (CSF-1R) is the major regulator of
macrophage development and is associated with epithelial cancers of
the breast and ovary. Immunohistochemistry analysis of murine
prostate development demonstrated epithelial expression of CSF-1R
during the protrusion of prostatic buds from the urogenital sinus,
during the prepubertal and androgen-driven proliferative expansion
and branching of the gland, with a decline in older animals. Models
of murine prostate cancer showed CSF-1R expression in areas of
carcinoma- and tumor-associated macrophages. Several human pros-
tate cancer cell lines and primary cultures of human prostate epithelial
cells had low but detectable levels of CSF-1R. Human prostatectomy
samples showed low or undetectable levels of receptor in normal
glands or benign prostatic hypertrophy specimens. Staining was
strongest in areas of prostatic intraepithelial neoplasia or carcinoma
of Gleason histological grade 3 or 4. The activated form of the
receptor reactive with antibodies specific for phosphotyrosine mod-
ified peptide sequences was observed in samples of metastatic
prostate cancer. Immunohistochemistry showed strong expression of
CSF-1R by macrophage lineage cells, including villous macrophages
and the syncytiotrophoblast layer of placenta, Kupper cells in the
liver, and histiocytes infiltrating near prostate cancers. These obser-
vations correlate CSF-1R expression with changes in the growth and
development of the normal and neoplastic prostate.
Perturbation of protein tyrosine kinase signaling is frequently
associated with malignant transformation (1). Tyrosine kinase
receptors and their ligands have been implicated in prostate de-
velopment and cancer, including transforming growth factor
␣
,
epidermal growth factor, insulin-like growth factor 1 (IGF-1),
fibroblast growth factors, hepatocyte growth factor (HGF), plate-
let-derived growth factor (PDGF), and nerve growth factors (re-
viewed in refs. 2 and 3). Strategies attacking EGF and PDGF
receptors are currently being evaluated in prostate cancer (4–7).
We defined the tyrosine kinase expression profile of normal
prostatic epithelial cells during a phase of rapid growth in a
relatively low androgen environment from day 10 murine prostate.
CD44 was used as a marker of early progenitor cells within prostatic
epithelium and is expressed by actively proliferating epithelia at
sites of epithelial–mesenchymal interaction (8, 9). Embryonic for-
mation of the prostate occurs through epithelial budding from the
urogenital sinus. Elongation and branching of the ducts begin
prenatally and are extensive during the first 21 days after birth.
Although ductal morphogenesis of the prostate is androgen-
dependent, the early postnatal period is marked by low levels of
circulating androgen (10, 11). Mice with loss-of-function mutations
in the homeobox genes NKX 3.1 or Hox D13 show mild defects in
prostate development (10). A more severe block in prostate de-
velopment is seen in P63⫺兾⫺mice, which do not develop a
recognizable prostate (12).
We prepared cDNA libraries from highly enriched CD44⫹
prostate cells from day 10 mice. A PCR-based strategy targeting
highly conserved tyrosine kinase catalytic domain sequences was
used (13, 14). One of the most frequently recovered tyrosine
kinases was the colony-stimulating factor-1 receptor (CSF-1R).
CSF-1R is encoded by the cellular homolog of the retroviral
oncogene v-fms (15) and is the major regulator of development
and response for all cells belonging to the mononuclear phago-
cyte lineage (16 –18).
In osteoclastogenesis, one of the critical factors produced by
bone stromal cells is CSF-1 (19, 20). The Csf1
op
兾Csf1
op
mouse
has inactivated the CSF-1 gene. These mice are osteopetrotic,
toothless, and have low fecundity and reduced macrophage
numbers (21, 22). CSF-1R null mutation (Csf1r
⫺
兾Csf1r
⫺
) mice
have a very similar but more severe phenotype (23). Prostate
histology and function have not been well characterized in either
Csf1
op
兾Csf1
op
or Csf1r
⫺
兾Csf1r
⫺
mice.
CSF-1R is expressed in testis, uterus, ovary, placenta, and
mammary glands (21, 24). Elevated expression of CSF-1R has been
seen in breast, ovarian, and uterine cancers, and the extent of
expression in these tumors correlates with high grade and poor
prognosis (24, 25). High circulating levels of CSF-1 correlate with
active disease in ovarian and endometrial cancers and with meta-
static breast and prostate cancer (24 –26). In this article, we show
that CSF-1R is expressed during the early phases of murine prostate
development and prostate cancer progression in mouse and human.
Materials and Methods
Animal and Cell Lines. Prostate cancer cell lines LNCaP (27), PC-3
(28) and DU145 (29), and breast cancer line BT-20 (30) were
obtained from American Type Culture Collection (Rockville,
MD). BT-20 cells were incubated in medium with 1
⌴dexa-
methasone (synthetic glucocorticoid, Sigma) (31); LAPC-4 was
provided by R. Reiter [University of California, Los Angeles
(UCLA); ref. 32]; and basaloid PrEC prostate cells were from
Clonetics (Walkersville, MD). Mouse prostate (C57BL兾6) was
fixed with 10% buffered formalin and embedded in paraffin wax.
Prostate tumors were from transgenic adenocarcinoma mouse
prostate (TRAMP) (33) and phosphatase and tensin homolog
deleted from chromosome 10 (PTEN) ⫹兾⫺mice (34, 35). Hong
Wu (UCLA) and Norman Greenberg (Baylor College of Med-
icine) kindly provided P TEN⫹兾⫺and TRAMP mice, respec-
tively. E. Richard Stanley and Xu-Ming Dai (Albert Einstein
College of Medicine) kindly provided tissue sections of Csfr
⫺
兾
Csf1r
⫺
mice. The M-NFS-60 murine macrophage line was used
as a control in some experiments (36). Csf1r
op
兾Csf1r
op
mice were
from The Jackson Laboratory.
Abbreviations: IGF-1, insulin-like growth factor 1; HGF, hepatocyte growth factor; PDGF,
platelet-derived growth factor; CSF-1R, colony stimulating factor-1 receptor; TRAMP,
transgenic adenocarcinoma mouse prostate; PTEN, phosphatase and tensin homolog
deleted from chromosome 10; TMA, tissue microarray analysis; IHC, immunohistochemis-
try; PrEC, prostate-derived basaloid epithelial cell populations; PIN, prostatic intraepithelial
neoplasia; UCLA, University of California, Los Angeles.
‡‡To whom correspondence should be addressed. E-mail: owenw@microbio.ucla.edu.
14404–14409
兩
PNAS
兩
October 29, 2002
兩
vol. 99
兩
no. 22 www.pnas.org兾cgi兾doi兾10.1073兾pnas.222537099
Prostate tissues were minced and digested with collagenase I
(Sigma) at a concentration of 1,500 units兾ml for 1 hr at 37°C.
Single-cell suspensions were analyzed with FACS Vantage (Bec-
ton Dickinson). For depletion of macrophages from mouse
prostate tissues, cells were incubated with mouse monoclonal
antibody CD11b for 30 min at 4°C, washed three times, and
incubated with sheep anti-mouse IgG Dynabeads M-450 (Dynal,
Oslo) for 30 min at 4°C, then separated by magnetic column. The
retained CD11b-positive and CD11b-negative cells that flow
through the column were tested.
RNA Expression Analysis. Total RNAs were extracted by the RNeasy
Mini kit (Qiagen, Chatsworth, CA). Each portion of total RNAs
from the cells was reverse-transcribed by oligo(dT) primer and
SuperScript reverse transcriptase (Life Technologies, Gaithers-
burg, MD) in a volume of 25
l after DNase I treatment. The
resulting cDNA was subjected to PCR. DNA sequences of the
primer pairs used are as follows: human CSF-1R, 5⬘-ACACTA-
AGCTCGCAATCCC-3⬘and 5⬘-GTATCGAAGGGTGAGCT-
CAAA-3⬘; mouse CSF-1R, 5⬘-GACCTGCTCCACTTCTC-
CAG-3⬘and 5⬘-GGGTTCAGACCAAGCGAGAAG-3⬘; human
CSF-1, 5⬘-GGAGTGGACACCTGCAGTCT-3⬘and 5⬘-TGTG-
CAGGGGCTGCTCACCA-3⬘; prostate-specific antigen, 5⬘-
GGTCGGCACAGCCTGTTTCA-3⬘and 5⬘-CCACGATGGT-
GTCCTTGATC-3⬘;

-actin, 5⬘-GACTACCTCATGAAGAT-
CCT-3⬘and 5⬘-GCGGATGTCCACGTCACACT-3⬘.
Immunoblot Analysis. For Western blot analyses, the cells were
resuspended in boiling sample buffer, and each sample was sepa-
rated on a 4–12% gradient SDS Tris Glycine Gel (NOVEX, San
Diego). Proteins were transferred onto a nitrocellulose membrane
(Micron Separations, Westboro, MA) and visualized with a chemi-
luminescence kit (Amersham Pharmacia). Rabbit anti-human
CSF-1R antibody (Santa Cruz Biotechnology, lot K130, 1:200
dilution) was used for Western blot analysis. Anti-ABL Western
blots were probed with 21–24 mouse monoclonal antibodies as an
internal loading control, as described (37). Goat anti-mouse IgG
and goat anti-rabbit antibody conjugated by horseradish peroxidase
(Bio-Rad) were used as secondary antibodies.
Tissue Microarray Analysis (TMA). Archival formalin-fixed paraffin-
embedded tissue samples were provided through the Depart-
ment of Pathology at the UCLA Medical Center under Institu-
tional Review Board approval. Primary radical prostatectomy
cases from 1984 to 1995 were randomly selected from the
pathology database to represent a wide spectrum of tumor
grades and primary disease stages. Three TMAs were con-
structed by the method of Kononen et al. (38), encompassing a
total of 246 individual patients. Patients treated preoperatively
with neoadjuvant hormones were then excluded from the anal-
ysis (n⫽20). An additional 11 cases were uninformative due to
missing spots or lack of target tissues. Overall, 215 prostatectomy
cases encompassing 1,109 informative tissue spots on three
TMAs were used for analysis. Included were 214 radical pros-
tatectomies and one cystoprostatectomy. The median age at
diagnosis of this cohort was 65 years. Semiquantitative assess-
ment of antibody staining on the TMAs was performed by one
pathologist (D.B.S.). Target tissue for scoring included only the
glandular elements of the prostate tissue. The maximal intensity
of diaminobenzidine brown chromogen staining was graded on
a0–2 scale (0 ⫽negative, 1 ⫽weak, and 2 ⫽strong staining).
Immunohistochemistry (IHC). Serial 4-
m-thick sections were depar-
affinized in three changes of xylene and rehydrated through a
100–70% descending series of ethanol, immersed in citrate buffer
(pH 6.0) in a 95°C water bath for 25 min, and then placed in 3%
H
2
O
2
in methanol for 10–20 min at room temperature to block
endogenous peroxidase activity. After the blocking of nonspecific
protein binding by incubation for 30 min to1hwith5%goat serum,
the whole-tissue sections were incubated with each primary anti-
body against either monoclonal mouse anti-human CD68 (DAKO),
polyclonal rabbit anti-human CSF-1R (Santa Cruz Biotechnology,
lot K130, 1:200 dilution), or polyclonal rabbit anti-human phos-
phorylated tyrosine 723 peptide (PY723) of CSF-1R (1:50 dilution)
overnight at 4°C (39). Subsequently, sections were processed for
IHC by using the EnVision
⫹
System (DAKO), StreptABComplex兾
HRP kit (DAKO), or Vector ABC Elite (Vector Laboratories),
as described (34). For PY723 of CSF-1R staining, we used the
EnVision
⫹
(DAKO) and tyramide signal amplification (NEN)
systems, according to the manufacturers’guidelines. For double
staining of CD68 and CSF-1R, we used the EnVision Doublestain
system (DAKO), according to the protocol recommended by the
manufacturer. Specificity of stain ing was confirmed by replacing the
primary antibody with mouse IgG for CD68 or rabbit IgG for other
antibodies at the same concentration and by blocking of positive
staining using excess purified peptides.
Results
CSF-1R Expression in Mouse Prostate Development and Cancer. Lim-
ited data are available on the expression of CSF-1R in rodent
prostate. One group has reported that a rat prostate cell line is
positive for CSF-1R by microarray analysis, and the level of
expression was found to increase with added androgen (40). RT-
PCR analysis for CSF-1R on a panel of mouse prostate-derived cell
lines showed a low frequency of positive lines (Fig. 7, which is
published as supporting information on the PNAS web site, www.
pnas.org). The midgestational budding of epithelium from the
urogenital sinus first defines the mouse embryonic prostate. IHC
for CSF-1R with a commercial rabbit anti-CSF-1R peptide anti-
serum (Santa Cruz Biotechnology) strongly and uniformly stains
these structures (Fig. 1). Because the gland actively proliferates
after birth, the level of CSF-1R remains high at 3 and 10 days (Fig.
1). Postpuberty at 8 weeks of age, limited cell division is occurring,
and the expression of CSF-1R is reduced to background levels
(Fig. 1). The staining on day 10 from a wild-type mouse is blocked
with competing peptide. Sections of prostate from a day 10
Csf1r
⫺
兾Csf1r
⫺
animal show a much lower level of staining. We
also examined liver sections from 8-week-old wild-type and
CSF-1R ⫺兾⫺mice by IHC. Positive strong staining was detected
in Kupper cells of wild-type mice but not in CSF-1R⫺兾⫺mice (Fig.
1). Strong staining for CSF-1R was seen in day 10 sections of
prostate harvested from Csf1r
op
兾Csf1r
op
mice (Fig. 8, which is
published as supporting information on the PNAS web site).
TRAMP is a transgenic mouse model in which the promoter
of the probasin gene controls simian virus 40 T antigen expres-
sion. These mice develop prostatic intraepithelial neoplasia by
12–18 and cancer by 19–25 weeks of age (33). Immunohisto-
chemical analysis of tissue sections from TRAMP tumors har-
vested between 23 and 30 weeks of age showed moderate to high
expression of CSF-1R in 20–25% of the cancerous glands (Fig.
2A). The staining is largely intracellular, in concert with previous
analyses of the membrane versus internal pools of CSF-1R (41).
Background staining was detected in normal prostate glands
from age-matched control mice. After preincubation with the
immunizing peptide, staining for CSF-1R is negative (Fig. 2A).
To verify the specificity of CSF-1R staining in prostate cancer
epithelial cells in TRAMP tumors, single-cell suspensions were
prepared, and CD11b-positive tissue macrophages were de-
pleted by adherence to antibody-conjugated magnetic beads
(Dynabeads; see Materials and Methods). The per-cell level of
CSF-1R was, as expected, much higher in the CD11b-positive
macrophage cell fraction by immunoblotting, as normalized to
the c-abl protein levels (Fig. 2B) or PCR analysis normalized to

-actin levels (Fig. 2C). Interestingly, the CD11b-negative cell
fraction is enriched for expression of the slower migrating
mature glycosylated form (165 kDa) of CSF-1R compared with
Ide et al. PNAS
兩
October 29, 2002
兩
vol. 99
兩
no. 22
兩
14405
MEDICAL SCIENCES
the macrophage-enriched CD11b-positive fraction, which has
roughly equal levels of the 165-kDa form and the immature
mannose-rich 140-kDa form of the receptor (42). A macrophage
cell line, NFS60, predominately shows the mature form.
CSF-1R expression was analyzed in prostate tumors from
PTEN⫹兾⫺mice from 12 to 18 months of age. Moderate diffuse
expression of CSF-1R was detected in the prostate tumors of
these mice compared with age-matched controls (Fig. 9, which
is published as supporting information on the PNAS web site).
Expression of CSF-1R in Human Prostate Cancer Cell Lines and in
Human Prostate Cancer Tissue. Several papers have noted increased
levels of CSF-1R in human breast, ovarian, and uterine cancers and
representative cell lines like BT-20 (24, 25). There are reports of the
expression of CSF-1R in human prostate cancer cell lines PC-3,
DU-145, and LNCaP by RT-PCR (43). Gene expression databases
using microarray data (Unigene, www.ncbi.nlm.nih.gov兾UniGene兾
Hs.Home.html), EST analyses (GeneCards, Weizmann Institute of
Science, http:兾兾bioinformatics.weizmann.ac.il兾cards兾), or serial
analysis of gene expression (National Center for Biotechnology
Information, www.ncbi.nlm.nih.gov兾SAGE兾) show moderate to
high levels of expression of CSF-1R sequences in human prostate
cancer-derived samples. It is difficult to evaluate the contribution of
expression from the prostate epithelial component of the tumor
versus the stromal component, which can include tissue macro-
phage-derived cells by such techniques. We evaluated prostate
cancer-derived clonal cell lines and commercially available pros-
tate-derived basaloid epithelial cell populations (PrEC). We could
detect low levels of CSF-1R mRNA expression in LNCaP, DU145,
and PrEC cells by RT-PCR analysis but not in the PC3 or LAPC4
cell lines (Fig. 3A). BT-20 cells are derived from a human breast
Fig. 1. Immunohistochemical analysis of CSF-1R in mouse developing pros-
tate. Tissue sections from 17 days past conception as well as 3 or 10 days and
8 weeks of age after birth of wild-type mice (C57BL兾6) stained with polyclonal
rabbit anti-CSF-1R antibody (Santa Cruz Biotechnology, lot K130, 1:200 dilu-
tion) based on the ABC method [DAKO or EnVison System (DAKO)]. The
specificity of this antibody was confirmed by serial sections stained with the
antibody after preincubation with immunizing peptide. Ten-day prostate and
liver sections from Csf1r⫺兾Csf1r⫺mice (23) were also stained with rabbit
anti-CSF-1R antibody at the same dilution.
Fig. 2. CSF-1Rexpression analyses in the prostate tumors of TRAMP mice. (A)
CSF-1R expression in age-matched control mouse prostate and in the prostate
tumor of a TRAMP mouse was examined by IHC with polyclonal rabbit anti-
CSF-1R antibody (Santa Cruz Biotechnology, lot K130, 1:200 dilution) or after
preincubation with immunizing peptide developed with the EnVison System
(DAKO). (B) Immunoblot and (C) RT-PCR analyses of CSF-1R expression in
single-cell suspensions prepared from a TRAMP tumor (23 weeks) before and
after depletion of CD11b-positive macrophages by Dynabead magnetic bead
technology as described in Materials and Methods. The M-NFS-60 cell line was
derived from a myelogenous leukemia (36).
14406
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.222537099 Ide et al.
cancer and are used as a positive control of CSF-1R expression (30).
All three of the prostate cell preparations positive for CSF-1R
expression were also positive for the CSF-1 ligand RNA, raising the
possibility of an autocrine or paracrine type of stimulation in some
situations. Western blot analysis of extracts from this panel of cell
lines (Fig. 3B) demonstrates the CSF-1R mature-sized protein (165
kDa) in BT-20, DU145, and PrEC cells, with lower levels in LNCaP
and borderline to undetectable levels in the PC3 and LAPC4.
Human prostate cancer tissue specimens were examined by
IHC for CSF-1R expression with rabbit anti-CSF-1R peptide
antiserum (Santa Cruz Biotechnology). Reactivity, shown as red
in cancerous glands, was much higher than in noncancerous
glands in the same field (Fig. 4A). Control rabbit IgG (Fig. 4B)
did not stain serial sections, and preincubation with the immu-
nizing peptide blocked reactivity of the antiserum (Fig. 10, which
is published as supporting information on the PNAS web site).
CSF-1R staining was heterogeneous and predominately cyto-
plasmic, with some membranous staining in prostate cancer (Fig.
4A) and prostatic intraepithelial neoplasia (PIN) (Fig. 4C),
which is consistent with prior reports of breast cancer staining
(44). The presence of CSF-1R was also detected in macrophages
near prostate cancers and within the lumens of some glands. The
level of expression in macrophages was much higher than
prostate cancer cells on a per-cell basis (Fig. 10). We confirmed
the identity of macrophages stained with anti-CSF-1R by double
staining with CD68 (Fig. 4D) (45). The CSF-1R antibody stained
placental syncytiotrophoblasts and villous macrophages at the
same dilution (Figs. 4Eand 10) and was inhibited with the
immunizing peptide (Fig. 4F). This antibody stained tumors
derived from xenographs of PC-3 cells engineered to express
CSF-1R (Fig. 11, which is published as supporting information
on the PNAS web site). A series of rat anti-human CSF-1R
monoclonal antibodies reactive with the ectodomain of the
native receptor (46) were evaluated for IHC but did not recog-
nize the antigen after tissue fixation.
IHC Analysis of CSF-1R Expression in Primary Prostate Cancers Using
Tissue Microarrays. We examined samples from 215 informative
cases, which had not received neoadjuvant therapy, by IHC
analysis using tissue microarrays. Specimen cores with normal
histology (230), benign prostatic hypertrophy (118), PIN (45),
and tumor (716) were evaluated. Sections were scored as unde-
Fig. 3. Expression analyses of CSF-1R in human prostate cancer cell lines. (A)
RT-PCR analysis. The PCR products at representative cycles were shown: CSF-1R,
CSF-1, and prostate-specific antigen, cycle 40 and

-actin, cycle 25. The primers
sequences are listed in Materials and Methods.(B) Immunoblotting analysis.
Protein lysates of each 2 ⫻105cells were subjected to SDS gel electrophoresis and
polyclonal rabbit anti-human CSF-1R antibodies (1:250 dilution) were used for
this analysis, as described in Materials and Methods. Human breast cancer cell line
BT-20; human prostate cancer cell lines PC3, DU-145, LNCap, and LAPC4, and
normal basaloid PrEC prostate cells were used in these analyses. PC-3 infected
with human CSF-1R expressing retrovirus was used as a control. Fig. 4. CSF-1R immunostaining in human primary prostate cancer with poly-
clonal rabbit anti-human CSF-1R antibodies (Santa Cruz Biotechnology) based on
the ABC method (DAKO) or the EnVision⫹System (DAKO). (A) Stronger staining
in tumor cells was seen compared with adjacent normal glands. (B) Control IgG
staining on a serial section. (C) Specific immunostaining was observed in the
cytoplasm of PIN as well as malignant epithelial tumor cells. (D) Macrophage
origin defined by CD68 staining (arrow: luminal macrophage in prostatic gland)
using the double-staining method as described in Materials and Methods. CD68
staining used DAB (brown) and CSF-1R staining used Fast red (red purple) as
substrates. (E) CSF-1R staining was detected in brush border of placental syncy-
tiotrophoblasts (arrows) and villous macrophages (arrowheads). (F) After prein-
cubation with the immunizing peptide, the staining for CSF-1R was blocked.
Ide et al. PNAS
兩
October 29, 2002
兩
vol. 99
兩
no. 22
兩
14407
MEDICAL SCIENCES
tectable, weak, or strong-staining. The majority of all specimens
(whether normal or malignant) showed some staining, but a clear
and important trend is the higher percentage of cases of PIN and
prostate cancer with strong staining (Fig. 5). Further analysis will
be needed to accurately correlate such staining results with
clinical and pathological variables. We have noted that the most
intense and uniform staining for lesions within the prostate
occurred in areas of PIN or carcinoma of histological Gleason
grade 3 or 4. Examples of representative tissue microarray IHC
results are shown in Fig. 12, which is published as supporting
information on the PNAS web site.
IHC Staining of Activated CSF-1R in Metastatic Sites by Tyrosine
723-Phosphospecific Peptide Antibody. Prostate cancer can spread to
distant sites, including lymph nodes, soft tissues, and, most com-
monly, bone and bone marrow (47). We used antibodies recogniz-
ing CSF-1R and specialized activation-specific reagents raised
against the phosphorylated tyrosine 723 peptide (PY723) of
CSF-1R (39). Phosphorylation of this tyrosine is important for the
CSF-1R-driven phenotypic traits of anchorage-independent growth
and metastasis (42, 48). Specimens of metastases were obtained
that stained positive for prostate-specific antigen, confirming that
they were of prostate cancer origin (data not shown). In each of the
specimens, carcinoma cell areas showed staining for the CSF-1R
and the PY723 modification of CSF-1R, which was blocked by the
cognate peptide (Fig. 6; also see Fig. 13, which is published as
supporting information on the PNAS web site).
Discussion
Our results correlate the expression of the CSF-1 receptor with
murine prostate development, proliferation, and cancer progres-
sion. Murine prostate carcinoma, initiated by distinct oncogenic
signals, demonstrated enhanced expression of this receptor.
Several cultured human prostate cancer-derived cell lines, pri-
mary cultures of prostate epithelium, and a large proportion of
human prostate cancer specimens, most prominent in PIN
lesions and carcinomas of histological grades 3 and 4, were
positive for CSF-1R expression. The mechanisms regulating this
pattern remain unknown.
Role of the CSF-1 Receptor in Steroid-Regulated Epithelial Develop-
ment. Mammary gland development in Csf1
op
兾Csf1
op
mice revealed
a lactational defect secondary to incomplete ductal growth, with
precocious development of the lobuloalveolar system (49). Crossing
Csf1
op
兾Csf1
op
mice with a mammary cancer-susceptible strain of
mice did not affect the incidence or growth of primar y tumors but
delayed their development to invasive metastatic carcinomas asso-
ciated with lessened infiltration and function of tumor-associated
macrophages (50). CSF-1 signaling provides a critical function for
mammary gland development during pregnancy, lactation, and
cancer progression (24).
Prostate development depends on steroid sex hormones, includ-
ing androgen. Androgen may not directly stimulate the prolifera-
tion of normal prostate epithelial cells (10, 11). Paracrine factors,
which are produced by the mesenchyme and regulated by steroid
hormones, play a critical role. Several growth factors that act as
paracrine mediators have been identified, including IGF-I, fibro-
blast growth factor (FGF)-7, FGF-10, HGF, and transforming
growth factor-

(11, 51). Our observations suggest that CSF-1 is a
candidate for such a paracrine factor.
CSF-1兾1R and Bone Metastasis. The most favored site of metastasis
of prostate cancer is bone (47). The ability of prostate cancer to
incite an osteoblastic reaction suggests there are bidirectional
pathways between prostate cancer cells and osteoblasts and
between osteoclasts and bone stromal cells. Prostate cancer
metastases in bone have an extensive bone resorption ability, and
Fig. 5. Immunohistochemical staining distribution of prostate by anti-CSF1R
antibody on tissue microarrays. Two hundred fifteen prostatectomy cases
encompassing 1,109 informative tissue spots on three TMAs were used for
analysis grading on a 0–2 scale (0 ⫽negative, 1 ⫽weak, and 2 ⫽strong
staining). Immunostaining conditions were as described in Materials and
Methods. NL, normal glands; BPH, benign prostatic hyperplasia; PIN, prostatic
intraepithelial neoplasia; PCA, prostate cancer. Tissue spot histology and
grading were confirmed on hematoxylin兾eosin slides.
Fig. 6. Immunohistochemical analysis of CSF-1R and activated CSF-1R expres-
sion in metastatic prostate cancer. Tissue sections from primary and metastatic
prostate cancer (aorta) were stained with polyclonal rabbit anti-CSF-1R (Santa
Cruz Biotechnology) antibody using the EnVision⫹System (DAKO) and polyclonal
rabbit anti-PY723 CSF-1R antibody using both the EnVision⫹system and the
tyramide signal amplification system (NEN). Coincidence of CSF-1R with activated
CSF-1R in some sections was confirmed by serial sections. The specificity of
anti-PY723 CSF-1R antibody was confirmed after preincubation with immunizing
PY723 peptide (10
M) blocked reactivity (39).
14408
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.222537099 Ide et al.
osteoclastogenesis may play a role in the establishment of bone
metastasis (52).
CSF-1 is an important regulator of hematopoiesis and bone
resorption (19, 20). Studies performed with Csf1
op
兾Csf1
op
mice
established that CSF-1 function is essential for proliferation and
differentiation of osteoclasts and is locally synthesized by bone
marrow stromal fibroblasts and osteoblasts. In addition to effects
on cell proliferation, differentiation, and invasion, the mem-
brane-bound CSF-1 may mediate adhesion (53).
Other Receptor Tyrosine Kinases Expressed in Prostate Cancer. Sev-
eral protein tyrosine kinases, such as PDGF receptor (PDGF-R),
c-met, HER-2兾neu, and IGF-1R, are expressed in prostate cancer,
including metastatic sites (2). CSF-1R is more closely related
structurally to the receptors for PDGF than to other PTK receptor
subfamilies (54). Immunohistochemical analyses show moderate to
strong PDGF-R
␣
expression in PIN and primary prostate cancers
but weak or absent expression in nonneoplastic prostate epithelium
and stroma (13). The expression of PDGF-R
␣
protein by metastatic
prostate cancer was also confirmed by IHC in a series of bone
marrow metastases (13). HER-2兾neu overexpression in primary
prostate cancer and in metastatic sites of prostate tumors was noted
before and after androgen-deprivation therapy (55). HGF is over-
expressed in prostate cancer, and its receptor c-met expression in
prostate cancer has been linked to higher histologic-grade and
-stage disease (56). The finding of activated CSF-1R in metastatic
tumors supports a role for CSF-1兾CSF-1R signaling in prostate
cancer progression.
We thank Norman Greenberg (Baylor College of Medicine) for providing
TRA MP mice; Dr. Hong Wu (UCLA) for providing PTEN⫹兾⫺mice; Drs.
E. Richard Stanley and Xu-Ming Dai (Albert Einstein College of Medicine)
for Csf1r⫺兾Csf1r⫺tissues; Martine Roussel (Department of Tumor Cell
Biology, St. Jude Children’s Research Hospital) and Charles Sherr (Howard
Hughes Medical Institute at St. Jude Children’s Research Hospital) for
reagents and valuable advice; J. C. White for preparation of the manuscript
and figures; Yoon Kim, Shirley Quan, Benjamin Rafii, Adam Gottesfeld,
Gregory Ferl, Sheila Tze, and Gregg Kanter for excellent technical assis-
tance; and Drs. Tetsuro Watabe, Robert E. Reiter, and members of the
Witte laboratory for helpful discussions. We thank the members of the
Human Tissue Research Center at UCLA for processing and sectioning of
tissues. We thank Dr. Arie Belldegrun (Department of Urology, UCLA) for
help in providing specimens used in the constr uction of the tissue microarray
used in these studies. Portions of this study were supported by a CaP CURE
grant (to O.N.W.), National Institutes of Health Grant GM08243-13 (to
D.B.S.), and funds from the Tina and Richard Carolan Prostate Research
Fund (to A.P.). O.N.W. is an Investigator and H.I. an Associate of the
Howard Hughes Medical Institute.
1. Blume-Jensen, P. & Hunter, T. (2001) Nature 411, 355–365.
2. Djakiew, D. (2000) Prostate 42, 150–160.
3. Ware, J. L. (1999) Cancer Metastasis Rev. 17, 443–447.
4. Gibbs, J. B. (2000) J. Clin. Invest. 105, 9–13.
5. Schwaab, T., Lewis, L. D., Cole, B. F., Deo, Y., Fanger, M. W., Wallace, P.,
Guyre, P. M., Kaufman, P. A., Heaney, J. A., Schned, A. R., et al. (2001)
J. Immunother. 24, 79–87.
6. Ko, Y. J., Small, E. J., Kabbinavar, F., Chachoua, A., Taneja, S., Reese, D.,
DePaoli, A., Hannah, A., Balk, S. P. & Bubley, G. J. (2001) Clin. Cancer Res.
7, 800–805.
7. Bergan, R. C., Waggle, D. H., Carter, S. K., Horak, I., Slichenmyer, W. &
Meyers, M. (2001) Urology 57, 77–80.
8. Yu, Q. & Toole, B. P. (1997) Dev. Dyn. 208, 1–10.
9. Gakunga, P., Frost, G., Shuster, S., Cunha, G., Formby, B. & Stern, R. (1997)
Development (Cambridge, U.K.) 124, 3987–3997.
10. Abate-Shen, C. & Shen, M. M. (2000) Genes Dev. 14, 2410 –2434.
11. Hayward, S. W. & Cunha, G. R. (2000) Radiol. Clin. North Am. 38, 1–14.
12. Signoretti, S., Waltregny, D., Dilks, J., Isaac, B., Lin, D., Garraway, L., Yang, A.,
Montironi, R., McKeon, F. & Loda, M. (2000) Am. J. Pathol. 157, 1769 –1775.
13. Chott, A., Sun, Z., Morganstern, D., Pan, J., Li, T., Susani, M., Mosberger, I.,
Upton, M. P., Bubley, G. J. & Balk, S. P. (1999) Am. J. Pathol. 155, 1271–1279.
14. Robinson, D. R., Wu, Y. M. & Lin, S. F. (2000) Oncogene 19, 5548 –5557.
15. Sherr, C. J., Rettenmier, C. W., Sacca, R., Roussel, M. F., Look, A. T. &
Stanley, E. R. (1985) Cell 41, 665–676.
16. Roth, P. & Stanley, E. R. (1992) Curr. Top. Microbiol. Immunol. 181, 141–167.
17. Cecchini, M. G., Dominguez, M. G., Mocci, S., Wetterwald, A., Felix, R.,
Fleisch, H., Chisholm, O., Hofstetter, W., Pollard, J. W. & Stanley, E. R. (1994)
Development (Cambridge, U.K.) 120, 1357–1372.
18. Pollard, J. W. & Stanley, E. R. (1996) Adv. Dev. Biochem. 4, 153–193.
19. Cecchini, M. G., Hofstetter, W., Halasy, J., Wetterwald, A. & Felix, R. (1997)
Mol. Reprod. Dev. 46, 75–83.
20. Ryan, G. R., Dai, X. M., Dominguez, M. G., Tong, W., Chuan, F., Chisholm,
O., Russell, R. G., Pollard, J. W. & Stanley, E. R. (2001) Blood 98, 74–84.
21. Pollard, J. W. (1997) Mol. Reprod. Dev. 46, 54–61.
22. Yoshida, H., Hayashi, S. I., Kunisada, T., Ogawa, M., Nishikawa, S., Okamura,
H., Sudo, T., Shultz, L. D. & Nishikawa, S. I. (1990) Nature 345, 442–444.
23. Dai, X. M., Ryan, G. R., Hapel, A. J., Dominguez, M. G., Russell, R. G., Kapp,
S., Sylvestre, V. & Stanley, E. R. (2002) Blood 99, 111–120.
24. Sapi, E. & Kacinski, B. M. (1999) Proc. Soc. Exp. Biol. Med. 220, 1–8.
25. Kacinski, B. M. (1997) Mol. Reprod. Dev. 46, 71–74.
26. McDermott, R. S., Deneux, L., Mosseri, V., Vedrenne, J., Clough, K.,
Fourquet, A., Rodriguez, J., Cosset, J. M., Sastre, X., Beuzeboc, P., et al. (2002)
Eur. Cytokine Network 13, 121–137.
27. Horoszewicz, J. S., Leong, S. S., Chu, T. M., Wajsman, Z. L., Friedman, M.,
Papsidero, L., Kim, U., Chai, L. S., Kakati, S., Arya, S. K. & Sandberg, A. A.
(1980) Prog. Clin. Biol. Res. 37, 115–132.
28. Kaighn, M. E., Narayan, K. S., Ohnuki, Y., Lechner, J. F. & Jones, L. W. (1979)
Invest. Urol. 17, 16 –23.
29. Stone, K. R., Mickey, D. D., Wunderli, H., Mickey, G. H. & Paulson, D. F.
(1978) Int. J. Cancer 21, 274–281.
30. Filder man, A. E., Bruckner, A., Kacinski, B. M., Deng, N. & Remold, H. G.
(1992) Cancer Res. 52, 3661–3666.
31. Sapi, E., Flick, M. B., Gilmore-Hebert, M., Rodov, S. & Kacinski, B. M. (1995)
Oncogene 10, 529–542.
32. Klein, K. A., Reiter, R. E., Redula, J., Moradi, H., Zhu, X. L., Brothman, A. R.,
Lamb, D. J., Marcelli, M., Belldegrun, A., Witte, O. N.,etal.(1997) Nat . Med.
3, 402–408.
33. Greenberg, N. M., DeMayo, F., Finegold, M. J., Medina, D., Tilley, W. D.,
Aspinall, J. O., Cunha, G. R., Donjacour, A. A., Matusik, R. J. & Rosen, J. M.
(1995) Proc. Natl. Acad. Sci. USA 92, 3439–3443.
34. Dubey, P., Wu, H., Reiter, R. E. & Witte, O. N. (2001) Cancer Res. 61,
3256–3261.
35. Stiles, B., Gilman, V., Khanzenzon, N., Lesche, R., Li, A., Qiao, R., Liu, X. &
Wu, H. (2002) Mol. Cell. Biol. 22, 3842–3851.
36. Nakoinz, I., Lee, M. T., Weaver, J. F. & Ralph, P. (1990) J. Immunol. 145, 860 –864.
37. Era, T. & Witte, O. N. (2000) Proc. Natl. Acad. Sci. USA 97, 1737–1742.
38. Kononen, J., Bubendorf, L., Kallioniemi, A., Barlund, M., Schraml, P.,
Leighton, S., Torhorst, J., Mihatsch, M. J., Sauter, G. & Kallioniemi, O. P.
(1998) Nat. Med. 4, 767–768.
39. Flick, M. B., Sapi, E., Perrott a, P. L., Maher, M. G., Halaban, R., Carter, D.
& Kacinski, B. M. (1997) Oncogene 14, 2553–2561.
40. Whitacre, D. C., Chauhan, S., Davis, T., Gordon, D., Cress, A. E. & Miesfeld,
R. L. (2002) Cell Growth Differ. 13, 1–11.
41. Guilbert, L. J. & Stanley, E. R. (1986) J. Biol. Chem. 261, 4024–4032.
42. Roussel, M. F., Downing, J. R. & Sherr, C. J. (1990) Oncogene 5, 25–30.
43. Savarese, D. M., Valinski, H., Quesenberry, P. & Savarese, T. (1998) Prostate
34, 80–91.
44. Maher, M. G., Sapi, E., Turner, B., Gumbs, A., Perrotta, P. L., Carter, D.,
Kacinski, B. M. & Haffty, B. G. (1998) Clin. Cancer Res. 4, 1851–1856.
45. Pulford, K. A., Sipos, A., Cordell, J. L., Stross, W. P. & Mason, D. Y. (1990)
Int. Immunol. 2, 973–980.
46. Ashmun, R. A., Look, A. T., Roberts, W. M., Roussel, M. F., Seremetis, S.,
Ohtsuka, M. & Sherr, C. J. (1989) Blood 73, 827–837.
47. Scher, H. I. & Yagoda, A. (1987) Am. J. Med. 82, 6–28.
48. Roussel, M. F., Dull, T. J., Rettenmier, C. W., Ralph, P., Ullrich, A. & Sherr,
C. J. (1987) Nature 325, 549–552.
49. Pollard, J. W. & Hennighausen, L. (1994) Proc. Natl. Acad. Sci. USA 91,
9312–9316.
50. Lin, E. Y., Nguyen, A. V., Russell, R. G. & Pollard, J. W. (2001) J. Exp. Med.
193, 727–740.
51. Stein, J., Borzillo, G. V. & Rettenmier, C. W. (1990) Blood 76, 1308 –1314.
52. Zhang, J., Dai, J., Qi, Y., Lin, D. L., Smith, P., Strayhorn, C., Mizokami, A.,
Fu, Z., Westman, J. & Keller, E. T. (2001) J. Clin. Invest. 107, 1219–1220.
53. Uemura, N., Ozawa, K., Takahashi, K., Tojo, A., Tani, K., Harigaya, K., Suzu,
S., Motoyoshi, K., Matsuda, H. & Yagita, H. (1993) Blood 82, 2634–2640.
54. Hamilton, J. A. (1997) J. Leukocyte Biol. 62, 145–155.
55. Osman, I., Scher, H. I., Drobnjak, M., Verbel, D., Morris, M., Agus, D., Ross,
J. S. & Cordon-Cardo, C. (2001) Clin. Cancer Res. 7, 2643–2647.
56. Humphrey, P. A., Zhu, X., Zarnegar, R., Swanson, P. E., Ratliff, T. L., Vollmer,
R. T. & Day, M. L. (1995) Am. J. Pathol. 147, 386–396.
Ide et al. PNAS
兩
October 29, 2002
兩
vol. 99
兩
no. 22
兩
14409
MEDICAL SCIENCES
