 |
INTRODUCTION |
The Rb1 gene
product mediates numerous cellular functions including cell cycle
regulation (1, 2), maintenance of chromosomal integrity (3, 4),
cellular differentiation (5-7), and the survival of epithelial cells
(8-11). The Rb gene encodes a phosphoprotein (pRb) that
regulates the transition between G1 and S phases of the
cell cycle by transducing growth-inhibitory signals that arrest cells
in G1 (12-14). Functional regulation of pRb is cell
cycle-dependent, being strictly controlled by the activity
of cyclin-dependent kinases that regulate the state of pRb
phosphorylation. Dephosphorylated pRb inhibits the transcription of
genes via its interaction with members of the E2F family of
transcription factors. As the cell approaches the G1/S
border, cyclin D-cdk4/6 and cyclin E-cdk2 complexes sequentially
phosphorylate pRb. These events lead to the release of E2F and
subsequent activation of E2F-regulated genes, such as c-myc,
cyclin E, PCNA, and DNA polymerase, that are required for entry and
activation of the S phase.
The control of G1/S transit provides a dogmatic view of pRb
function, as cell cycle regulation is vastly more complex than the
simple scheme provided here. The overlapping function of two structurally related family members, p107 and p130, represents an
interactive mechanism in which pRb, p107, and p130 share the ability to
regulate different members of the E2F family and thus a variety of
target genes (13, 15, 16). Although structurally similar, there is
growing evidence supporting distinct cellular functions for each
Rb family member. All three proteins are differentially expressed during mouse development (17), and their ability to initiate
growth arrest is cell type-specific (18). In addition, these proteins
preferentially associate with different E2F family members. Although
pRb interacts with E2F1-4, there is apparent redundancy in the
regulation of both E2F4 and E2F5 by p107 and p130 (reviewed in Ref.
19). The similarities and differences between these proteins are also
apparent in mice carrying single or compound knockouts of
Rb, p107 and p130. When the Rb gene is deleted
through targeted disruption, the embryos die at 13 days of gestation
from defective development of erythroid and neuronal tissues (9-11).
In stark contrast, targeted disruption of either p107 or p130 does not
result in an obvious phenotype, and the mice remain viable (20).
Members of the Rb family are believed to play active roles
in tissue development by regulating a postmitotic state required for
cellular differentiation (5, 20, 21). The normal differentiation and
development of the prostate gland are critically dependent upon
androgenic steroids which, following binding to the androgen receptor,
transactivate or repress a number of transcriptional targets, including
cell cycle regulatory genes. It was recently demonstrated (8, 22-24)
that pRb is activated during androgen-stimulated epithelial
proliferation and during androgen ablation-induced apoptosis. The pRb
protein has also been shown to function as a transcriptional
co-activator of the androgen receptor (25). Taken together, these
results tentatively position Rb as a central mediator of
androgen action controlling the differentiation, growth, and death of
prostate epithelium; however, this hypothesis has not been tested.
Cell culture models currently available to study epithelial physiology
have a limited scope of relevance with regard to the pRb pathway. Many
transformed prostate cell lines, such as DU145, already exhibit
nonfunctional pRb due to mutation (26, 27). Viral oncogenes are common
tools used to immortalize cells in culture or study cell cycle
regulatory mechanisms. The viral oncogenes SV-40 large T antigen,
adenovirus E1A and E1B, and human papilloma virus E6 and E7
target and inactivate the pocket protein family members (pRb, p107, and
p130) as well as p53 (28-30). The use of viral oncogenes has provided
substantial insight into the function of the Rb family members and
their roles in regulating cell cycle, cell growth, and differentiation;
however, viral oncogenes are promiscuous in their interactions with
other cellular proteins and promote genomic instability making
interpretation of these experimental models difficult. The chromosomal
imbalances directly influenced by viral oncogenes have been identified
as either random or nonrandom genetic events and include gross
chromosomal translocations (31-33). Therefore, models that
specifically target and inactivate Rb, which minimize
complicating genetic alterations inherent with viral oncogenes, might
provide novel insight into the physiologic role of Rb in
epithelial cells.
The homozygous deletion of the Rb gene results in embryonic
lethality due to a variety of developmental abnormalities (9-11). The
embryonic lethality of the Rb knockout mutation has
prevented the development of Rb
/ epithelial cell lines,
and thus many of the functional aspects of pRb have not been
independently characterized in this cellular population. Recently, the
application of tissue recombination, using fetal tissue rescued from
Rb/ embryos and propagated in combination with wild type
(Rb+/+) prostate stromal tissue, has enabled the development
of Rb/ prostate (34-36) and mammary grafts (37). In the
current study, we describe the isolation and characterization of the
Rb/ prostate epithelial cell line, Rb/PrE, derived
from Rb/ fetal urogenital precursor tissue rescued from
embryonic mice (34). To our knowledge, this is the first in
vitro model to allow for the study of targeted Rb deletion on an epithelial population and provides a unique experimental platform with which to investigate the physiological consequences of
Rb deletion on the regulation of cell cycle,
differentiation, cell survival, and carcinogenesis.
| |
MATERIALS AND METHODS |
PCR Genotyping for Rb/ Tissues and Cell
Lines--
Heterozygous Rb+/ mice were purchased from The Jackson
Laboratories (Bar Harbor, ME). To increase litter size and thus the chances of Rb/ offspring, the heterozygotes were crossed to CD1
mice. The genetic identity of the offspring was confirmed by PCR
genotyping to identify the presence of the neomycin selection cassette
that was used to disrupt the Rb gene. The Rb heterozygotes were crossed ant the fetuses rescued, and the prostatic ductal tips
were then recombined as described previously (34). The Rb
status of tissue grafts was determined by PCR analysis. Control (The
Jackson Laboratories) and experimental DNA samples were amplified using
wild type- and mutant-specific primers and separated on 2% agarose
gels containing ethidium bromide. The protocol for PCR cycling
conditions was obtained from The Jackson Laboratory technical
support (micetech{at}jax.org). PCR primers (The Jackson Laboratories)
used for genotyping tissues and cell lines were as follows:
Rb knockout allele, forward 5'-AAT TGC GGC CGC ATC TGC ATC
TTT ATC GC-3' (oIMR025) and reverse 5'-GAA GAA CGA GAT CAG CAG-3'
(oIMR027); Rb wild type allele, forward 5'-AAT TGC GGC CGC
ATC TGC ATC TTT ATC GC-3' (oIMR025) and reverse 5'-CCC ATG TTC GGT CCC
TAG-3' (oIMR026) (10).
Serial Recombination of Rb/ Prostate Tissues--
Serial
tissue recombination was used to assess the ability of the
Rb/ epithelial cells to undergo multiple rounds of
growth and generate immortalized tissue. A ductal tip of ~300 µm
was micro-dissected from rescued Rb/ prostatic tissue
(34), recombined with normal rat urogenital mesenchyme (rUGM), and
grafted beneath the renal capsule of an intact male athymic mouse host.
After 1 month of growth, the host was sacrificed, and the graft was retrieved. The resultant 40 mg of prostatic tissue (38) was again
micro-dissected, and another 300-µm ductal tip was recombined with
fresh rUGM and grafted into a new mouse host to produce a "second
generation" graft. This recombination protocol, repeated 8 times,
resulted in ~13 epithelial population doublings for each round of
recombination and re-grafting as estimated by the number of epithelial
cells from the tissue weight using the Coffey equation 1g = 109 cells (39).
Generation of Wild type PrE and Rb/PrE
Cells--
Rb/ prostate grafts were established in nude
mice, and then the ductal fragments were recombined with rUGM as
described previously (34). A portion of each excised graft was fixed
for histological examination, and the rest were utilized to create Rb/PrE epithelial cultures. The tissue was minced with a scalpel and forceps and plated onto a tissue culture dish coated with collagen substrate in a minimal volume of medium to allow for attachment of the tissue to the matrix. These tissues were grown in
RPMI 1640 (BioWhittaker). Culture media were supplemented with ITS (5 µg/ml insulin, 5 µg/ml transferrin, 5 ng/ml selenium,
Collaborative Research), BPE (10 µg/ml bovine pituitary extract,
Sigma), epidermal growth factor (10 µg/ml, Collaborative Research),
cholera toxin (1.0 µg/ml, Sigma), amphotericin B (250 µg/ml,
fungizone, Invitrogen), dexamethasone (5 µM, Sigma), 200 mM L-glutamine (Invitrogen), and 100 units/ml
penicillin G and 100 units/ml streptomycin (BioWhittaker). This
formulation supports the growth of epithelial cells while retarding the
growth of the fibroblast cells. Approximately 10-14 days after
plating, the tissue pieces were removed from the cultures, and
selection was initiated with 200 µg/ml G418 (BioWhittaker). Once
large areas of epithelial cells became established, cells were passaged
1:3 by trypsinization. Between passage 5 and passage 10, cultures were
gradually switched to a medium containing only 5% FBS,
L-glutamine (Invitrogen), and 100 units/ml penicillin G and
100 units/ml streptomycin (Invitrogen) in RPMI 1640 termed "5% FBS
growth medium."
Wild type PrE cultures were generated from prostates excised from
6-week-old strain-matched male CD1 mice (Harlan Laboratories) following
euthanasia with CO2. Prostate tissues were minced and plated in the defined BPE-containing culture medium described above.
These cultures were maintained identical to the Rb/PrE described
here without the addition of G418.
The wild type control cell line, termed PrE, utilized for comparison
herein, spontaneously immortalized in culture and was therefore
utilized as a control for spontaneous immortalization of mouse
prostatic epithelial cultures.
Species Determination of Cells Utilizing Hoechst 33258 Staining--
Mouse Rb/PrE cells were grown on Falcon chamber
slides to 70% confluence and fixed with 100% ethanol on ice for 5 min
before washing with two changes of cold PBS. Fixed cells were then
stained with Hoechst 33258 dye (5 µg/ml, Sigma) for 1 min at room
temperature. Following staining, cultures were again washed three times
in cold PBS, wet-mounted (Biomeda Corp.), and photographed using a
Zeiss Axioskop fluorescent microscope to confirm that the cells were of
mouse origin (40).
Long Term Serum-free Growth Analysis--
Wild type and
Rb/PrE (5 × 105) cells were plated into 60-mm
culture dishes containing 5% FBS growth medium. Three days after plating (termed "Day 0"), culture media were changed to serum-free media consisting of RPMI 1640 (BioWhittaker), 100 units/ml penicillin G, and 100 units/ml streptomycin (without phenol red and without L-glutamine). Cultures were fed with the aforementioned
serum-free medium every 3 days and counted at the indicated times,
where each time point is the average and S.D. of triplicate dishes. Viable cell counts were analyzed by trypan blue exclusion. Photographs were taken on a Nikon Diaphot 200 with a Nikon digital camera.
Re-grafting of Rb/PrE Cells by Cellular
Recombination--
At passage 21, Rb/PrE cells were utilized to
generate prostate grafts via cellular recombination. To prepare grafts,
2.5 × 105 urogenital mesenchymal cells and 1 × 105 Rb/PrE cells were combined in a collagen matrix as
described previously (41). These grafts were then transplanted beneath the renal capsule of an adult, male nude mouse host (Charles River Laboratories) and grown for 1 month. Host animals were then sacrificed, and the grafts were harvested, subjected to fixation, and evaluated utilizing immunohistochemical techniques.
Immunohistochemistry--
Tissue sections were deparaffinized in
Histoclear (National Diagnostic) and hydrated in graded alcoholic
solutions and PBS. Endogenous peroxidase activity was blocked with
0.5% hydrogen peroxide in methanol for 30 min and washed in PBS prior
to staining. Immunocytochemical staining for expression of cytokeratins
CK8, CK14, and CK18 in Rb/PrE cells was performed as described
previously (34). Cells were grown on chamber slides (Falcon) coated
with 5 µg/ml fibronectin prior to fixation with 100% ethanol for 5 min on ice. Cells were then washed with PBS, and glucose oxidase substrate was employed in conjunction with the mouse and rabbit staining kits from Vector Laboratories. Staining for the AR and estrogen receptor 
(ER), mDLP, and p63 was also repeated
with the re-grafted Rb/PrE cells (passage 20). Following
growth in nude mouse hosts, the re-grafted tissue was harvested and
subjected to the staining procedure as described previously (34,
42).
Western Blot, Antibodies, and Protein
Analysis--
Primary antibodies were obtained as follows:
anti-E-cadherin (Transduction Laboratories, C20820), anti-Rb
(Pharmingen, 14001A), anti-p107 (Transduction Laboratories, R27020),
anti-p130 (Santa Cruz Biotechnology, SC-318), anti-p21 (Pharmingen,
556430), anti-p53 (Oncogene, OP29), anti-AR (Santa Cruz Biotechnology, SC-816), anti-estrogen receptors and 
(Santa Cruz Biotechnology, SC-542 and SC-8974, respectively), anti-cyclin D1 (Santa Cruz Biotechnology, SC-8396), anti-cyclin E1 (Santa Cruz Biotechnology, SC-481), anti-PCNA (Santa Cruz Biotechnology, SC-9857),
anti-actin (Santa Cruz Biotechnology, SC-1615), and high molecular
weight pan-cytokeratin (Z0622, Dako). Horseradish
peroxidase-conjugated secondary antibodies were obtained as follows:
donkey anti-mouse (Amresco, E974), goat anti-rabbit (Bio-Rad), and
donkey anti-goat (Bio-Rad). For protein analysis, cultured cells were
lysed on ice in 50 mM Tris, pH 7.5, 120 mM
NaCl, 0.5% Nonidet P-40, 1 mM EGTA, and protease
inhibitors (40 µM phenylmethylsulfonyl fluoride, 5 µg/ml leupeptin, 50 µg/ml aprotinin, 200 µM sodium
orthovanadate). Following centrifugation, the supernatants were
collected, quantitated using a Bradford microtiter assay, and denatured
with a reducing 2× sample loading buffer for 5 min at 100 °C. All
proteins were then separated on Tris/glycine pre-cast NOVEX gels and
analyzed utilizing the ECL detection system (Amersham Biosciences) as
described previously (43).
Growth Kinetics of PrE and Rb/PrE Cells--
Cultures of PrE
and Rb/ PrE were maintained in RPMI 1640 (BioWhittaker) containing
5% dextran-coated, charcoal stripped FBS, termed "5% CCS growth
medium," and compared with cultures grown in 5% FBS growth medium.
For these experiments, PrE and Rb/PrE cells were plated at a
density of 5 × 105 cells into 100-mm culture dishes,
and viability and cell number were assessed via trypan blue
(Invitrogen) exclusion on various days after plating.
Flow Cytometric Evaluations--
PrE and Rb/PrE
(1 × 106, passages 15-20) were plated into 100-mm
dishes and analyzed 4 days later to determine log phase cell cycle
profiles. To ascertain growth arrest in confluent cultures, cells were
plated at a higher density (2 × 106 cells per dish)
and retained in culture for a total of 15 days. Serum-containing medium
was replaced every 3 days on the long term cultures. Cells were
harvested by trypsinization, fixed, and stained with
4',6-diamidino-2-phenylindole, dihydrochloride (DAPI). Prior to
examination of cells by flow cytometry, 1 × 106 cells
were collected by trypsinization, centrifuged, and reconstituted with
800 µl of PBS (without calcium or magnesium). One drop of an internal
trout DNA control (Reiss Enterprise, 1007) was added to each sample.
Then, 3.5 ml of cold 100% ethanol was added dropwise while mixing for
fixation. Samples were then incubated for 1 h on ice prior to
centrifugation and re-hydrated with 1 ml of PBS for 15 min on ice.
Again, all samples were centrifuged and reconstituted with 1.5 ml of
DAPI staining buffer (100 mM Tris, pH 7.4, 150 mM NaCl, 1 mM CaCl2, 0.5 mM MgCl2, and 0.1% Nonidet P-40). A final concentration of 1.0 µg/ml DAPI (Molecular Probes, D-1306) was utilized for flow cytometric evaluations. Flow cytometry was carried out at the University of Michigan Flow Cytometry Core facility using BD
Biosciences FACSVantage SE model 127. Data were acquired to
105 events per sample. MultiCycle software (Phoenix Flow
Systems, San Diego, CA) was utilized to estimate the percentage of
cells in G1, S, and G2M phases of the cell
cycle populations.
Spectral Karyotype (SKY) Analysis and Comparative Genomic
Hybridization (CGH)--
Metaphase chromosomes from cultures of wild
type PrE and Rb/PrE were obtained by mitotic shake off after 1 h of colcemide (1 µg/ml) treatment. Slides were hybridized with SKY
kits, prepared from flow-sorted chromosomes, and detected 72 h
later as described previously (44). Images of 10-15 metaphase cells
were acquired using a DMRXA microscope (Leica, Wetzlar, Germany)
equipped with a custom-designed SKY-3 optical filter (Chroma
Technology, Brattleboro, VT), a spectral cube, and a charge-coupled
device camera (Hamamatsu, Bridgewater, NJ). Analysis was performed with
SkyView software (Applied Spectral Imaging, Ltd., Migdal Haemek,
Israel) as described elsewhere (45). For CGH analysis, DNA was prepared
under high salt conditions. Biotin-labeled DNA was derived from PrE and
Rb/PrE cultures and co-precipitated with digoxigenin-labeled
reference DNA obtained from sex- and strain-matched Rb+/
and/or Rb+/+ mice. DNA was hybridized to sex-matched normal
murine (C57) lymphocyte metaphase chromosomes and detected, and images
were acquired with Q-CGH software (Leica Imaging Systems,
Cambridge, UK) (46).
| |
RESULTS |
Genotype and Phenotype of PrE and Rb/PrE
Cells--
Because the Rb gene was disrupted by the
insertion of the neomycin resistance cassette (47), the use
of cellular recombination (34), employing neomycin-sensitive, wild type
rUGM, and neomycin-resistant Rb/PrE mouse epithelium allowed for
the specific selection of Rb/PrE from the wild type stroma.
Following selection on neomycin for several weeks, PCR genotyping
revealed that the Rb/PrE cells exhibited only the larger 420-bp
mutant PCR product compared with the 400-bp wild type PCR product or
the mixed PCR products of the heterozygous control cells (Fig.
1A). Western analysis of Rb/PrE cells revealed complete loss of pRb protein
expression while maintaining E-cadherin (epithelial-cadherin) protein
expression (Fig. 1B), confirming the Rb/
genotype and epithelial lineage of the Rb/PrE cells. Rb/PrE
cultures growing in 5% FBS growth medium were photographed at
subconfluent and confluent densities utilizing phase-contrast
microscopy (Fig. 1C). Staining Rb/PrE cells with Hoechst
33258 dye revealed homogeneous, punctate nuclear patterns
characteristic of mouse cells, whereas control rat epithelial cells
exhibited a non-punctate staining pattern, confirming that Rb/PrE
cells were not derived from rat tissues (Fig. 1C).
View larger version (49K):
[in this window]
[in a new window]
|
Fig. 1.
Verification of
Rb/PrE genotype and
murine epithelial histology. A, primer sets 025-026 (Rb+/+) and 025-027 (Rb/) amplified only wild
type (Rb+/+) alleles (400 bp) from wild type mouse tail DNA
(lane 1) and wild type PrE cells (lane 2). The
same primer sets amplified wild type and mutant alleles from
heterozygous control cells (lanes 3 and 4). Only
the mutant allele (420 bp) was amplified with these primer sets from
passage 4 and passage 10 Rb/PrE cells (lanes 5 and
6). B, Rb/PrE cells lack the 110-kDa pRb
protein but strongly express E-cadherin, which also serves as a loading
control. C, phase-contrast microscopy of Rb/PrE cells at
subconfluent (1) and confluent (2) densities. The
punctate nuclear staining pattern of the Hoechst 33258 stain is evident
in mouse Rb/PrE cells (3) and absent in the Dunning rat
control epithelial cells (4).
|
|
Analysis of Rb/PrE Survival and Growth in the Absence of
Serum--
Rb appears to play a paradoxical role in the regulation of
cell survival, as evidence has emerged supporting both pro- and anti-apoptotic functions of the protein (4). Wild type PrE and
Rb/PrE cells cultured beyond passage 8 are maintained in media
supplemented with 5% FBS. We wanted to determine whether Rb/PrE
cells could continue to proliferate in the absence of serum containing
growth factors and to determine whether the loss of pRb would
compromise survival. Viable cell counts of PrE and Rb/PrE cultures
were documented under serum-free conditions (Fig.
2A). Rb/PrE cells remained
viable and continued to proliferate for more than 50 days in serum-free
media, whereas the wild type PrE cells ceased to proliferate and
exhibited a marked loss in viability. These results suggest that the
Rb/PrE cultures can circumvent cell death programs in the absence
of growth factors.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 2.
Rb/PrE cells are
viable in the absence of serum and are immortal in vitro
and in vivo. A, wild type PrE
and Rb/PrE cells were grown under serum-free conditions, and viable
cells were counted at the indicated times. Results are reported as the
average percent relative viability from three independent experiments.
B, the number of passages for which the Rb/PrE cells
have been cultured by trypsinization in serum-containing media.
C, wild type PrE cells and Rb/PrE cells were utilized
for tissue recombination, and the number of viable serial passages of
the grafts obtained is indicated (C).
|
|
We rationalized that the loss of Rb would allow cells to override
growth-restraining or apoptotic mechanisms that may result in
immortalization. We could demonstrate that the Rb/PrE epithelial cells did not senesce and continued to proliferate for more than 120 passages (Fig. 2B). These results are in contrast to
previous studies of Rb/ mouse embryonic fibroblasts in
which these cells senesced at early passage (19, 48). Even though
Rb/PrE cells are immortal in culture, spontaneous immortalization
of murine cells has been reported (49), and we have observed the
spontaneous immortalization of approximately half of our wild type PrE
cultures (data not shown). To determine that the immortalization of
Rb/PrE cells resulted from artifactual effects of cell culturing or
was a specific outcome of Rb disruption, we examined the
immortalization of Rb/PrE cells in recombinant tissue grafts
in vivo. PrE and Rb/PrE grafts were subjected to
multiple rounds of serial tissue recombination with rUGM in male
athymic-mouse hosts. Serial regrafting was repeated for up to seven
additional passages in vivo, for a total of eight in
vivo passages. Each in vivo passage (to expand epithelial cell growth) represents ~13 population doublings. The wild
type grafts only proliferated and survived through three rounds of
serial regrafting. In contrast, the Rb/PrE grafts survived eight
rounds of recombination in vivo and were still viable at the
termination of the experiment (Fig. 2C).
The Loss of Rb Does Not Influence Prostatic Histodifferentiation
and Morphogenesis in Vitro and in Vivo--
Rb is believed
to play an active role in tissue development by regulating a
post-mitotic state required for cellular differentiation, including
androgen-mediated differentiation of prostate epithelium (reviewed in
Ref. 50). Morphologic examination and E-cadherin expression in Fig. 1
confirmed the epithelial lineage of the mouse Rb/PrE cells. To
examine more closely the state of Rb/PrE prostatic differentiation
in culture cells, we examined a number of prostatic and epithelial
markers. Cytokeratin expression was assessed in Rb/PrE cells
revealing low expression of the basal epithelial marker, cytokeratin
14, and high expression of the luminal epithelial marker, cytokeratin
18 (Fig. 3, A and
B, respectively), indicating a mixed epithelial population
dominated by a luminal phenotype. Western blot analysis of Rb/PrE
cultures at early and later passage also revealed strong androgen
receptor (AR) and estrogen receptor (ER) expression (Fig.
3C), whereas the estrogen receptor immunoreactivity was
negative (data not shown). These results demonstrated that the
Rb/PrE cells continued to express prostate-specific markers even
after many passages in culture.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 3.
Rb/PrE cells
retain prostate-specific markers. Phase-contrast microscopy of
subconfluent Rb/PrE cultures immunostained with antibodies against
cytokeratin 14 (A) and cytokeratin 18 (B).
C, PrE and Rb/PrE cellular lysates were resolved by
SDS-PAGE and probed with anti-AR and anti-ER antibodies at various
passages.
|
|
To determine whether Rb/PrE cells could recapitulate prostate
histodifferentiation and morphogenesis in vivo, the
Rb/PrE (passage 20) and wild type PrE (passage 20) were recombined
with rUGM and grafted into intact male athymic host mice.
Hematoxylin and eosin staining revealed that the wild type
rUGM+PrE grafts (Fig. 4A) and
the rUGM+Rb/PrE grafts (Fig. 4B) were indistinguishable. Both grafts exhibited normal prostate glandular morphology and stained
positively for AR and for the murine, prostate-specific dorsal lateral
protein marker (mDLP), confirming the murine prostate lineage of this
graft (Fig. 4, C and D). It has been reported previously (42) that p63, a homologue of p53, is expressed in basal
epithelium of glandular tissue, such as the prostate gland, and that
p63 is required for prostatic development. Staining with a p63 antibody
revealed strong staining in a population of basal cells just beneath
the luminal component in both wild type and Rb/ grafts (Fig. 4,
E and F). These findings confirm the ability of
Rb/PrE cells to recapitulate normal prostatic histodifferentiation and morphogenesis in vivo.
View larger version (110K):
[in this window]
[in a new window]
|
Fig. 4.
Rb/PrE cells
recapitulate prostatic histodifferentiation and morphogenesis in
vivo. Hematoxylin and eosin (H & E) staining
of wild type UGM+PrE grafts (A) and UGM+Rb/PrE grafts
(B) that were re-grafted into a male nude mouse host.
Immunohistochemistry revealed AR-positive luminal cells (C)
and positive staining for the mDLP (D). p63 positive basal
cell staining in the UGM+PrE grafts (E); UGM+Rb/PrE
grafts (F).
|
|
Rb/PrE Cells Retain Intact p107/p130 and p53/p21
Pathways--
Beyond the established role of pRb in cell cycle arrest,
the involvement of p107 and p130 in the inhibition of cell growth has
also been documented in several cell types (18). In normal cells the
levels of p130 and p107 change dramatically during the cell cycle. When
cells are engaged in cycle and moving through S phase, the expression
of p130 is low; however, when cells are forced to exit the cell cycle
by serum withdrawal or contact inhibition, p130 protein accumulates
rapidly (13, 15). The expression of p107 is modulated in an opposing
manner to p130. p107 is present at high levels in cycling cells, and
like pRb is hyperphosphorylated; however, when cells exit the cycle in
response to serum withdrawal or contact inhibition, p107 is rapidly
dephosphorylated and protein levels decrease. To determine whether the
immortalized, serum-independent phenotype of Rb/PrE cells resulted
exclusively from Rb loss and not from ancillary loss of p107
or p130, we examined the expression of these proteins in control PrE
and Rb/PrE cells in either subconfluent, serum-free, or high
density culture conditions. In subconfluent cultures, the hyper- and
hypophosphorylated forms of p107 were detected in the control PrE and
Rb/PrE cells. Western blot analysis of both PrE and Rb/PrE
cells revealed that p107 was dephosphorylated, and protein levels were
dramatically reduced in serum-free and high density cultures (Fig.
5A). From the same lysates,
levels of p130 were slightly elevated in untreated PrE cells compared
with Rb/PrE but accumulated to the same extent in both cell lines
under serum-free and high density culture conditions (Fig.
5B). These results demonstrated that p107 and p130 respond normally to G1 growth arrest signals in Rb/PrE cells
and wild type PrE control cultures.
View larger version (57K):
[in this window]
[in a new window]
|
Fig. 5.
Rb/PrE cells
exhibit functional p107/p130 and p53/p21 pathways. PrE and
Rb/PrE cells cultured for 5 days at subconfluent density
(SC) in serum-free media (SF) or at high density
(HD) for 5 days were analyzed for levels of p107 protein
(A), p130 protein (B), and an E-cadherin
loading control (C) by Western blot. PrE and
Rb/PrE cells were UV-irradiated and harvested at 48 h, and protein lysates were prepared. Shown are the levels of p53
protein (D), p21 protein (E), and an actin
loading control (F) as measured by Western blot.
|
|
Physiologic exit from the cell cycle and the induction of growth arrest
have been linked to p53 and p53-dependent cell cycle inhibitors p19Arf and p21Cip1 (51). Cell cycle
exit in response to non-physiologic signals such as treatment with
ionizing radiation and DNA-intercalating agents have also been shown to
be regulated by the p53-dependent activation of p21.
To confirm that the immortalizing effects observed in Rb/PrE cells
was a direct result of Rb loss and not a result of ancillary
loss of p53 and/or p21, the expression and functionality of these
proteins in Rb/PrE cells were examined. Western blot analysis of
Rb/PrE cells demonstrated that induction of p53 protein was
comparable with those observed in wild type cells following UV
irradiation (Fig. 5C). Levels of p21 were also induced following UV irradiation in both Rb/PrE and wild type cells (Fig.
5D). These results suggest that the p53 pathway is
functional in the Rb/PrE cultures.
Growth Kinetics of PrE and Rb/PrE Cultures--
To compare
growth characteristics of PrE to Rb/PrE cells, proliferation
studies were performed on passages 20-22 of these two cell lines. At
the indicated times, the cells were trypsinized and counted, and the
viability was assessed by trypan blue exclusion assay. Fig.
6 is a representative experiment
demonstrating PrE and Rb/PrE growth kinetics. The results
demonstrate that the 5% FBS growth medium supported the growth of both
cell lines. Additionally, there was a 2-fold enhanced proliferation
rate of the Rb/PrE cells over the control PrE cells. To compare the growth kinetics in the absence of hormones and steroids, we performed this same experiment with cells cultured in dextran-coated,
charcoal-stripped 5% serum growth medium (5% dextran-coated, charcoal
stripped (DCC)). We found that under this culture condition, both cell
lines grew equally well in the presence or absence of steroids.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 6.
Growth kinetics of PrE and
Rb/PrE cells. PrE
and Rb/PrE cells were cultured in 5% serum-containing media or
charcoal-stripped (DCC) serum. Viable cells were counted
utilizing trypan blue exclusion at the days indicated. w.t.,
wild type.
|
|
Analysis of Cell Cycle Re-entry in Serum-free Synchronized PrE and
Rb/PrE Cells--
We next wanted to investigate the mechanism(s)
that may be responsible for the increased growth kinetics of the
Rb/PrE cells. The expression levels of cell cycle regulatory
proteins were assessed in synchronized wild type and Rb/PrE cells
as these cells re-entered the cycle following serum starvation. At the
indicated times, the cells were harvested and analyzed for protein
expression of cyclin D1, cyclin E1, PNCA, and actin. PrE cells
exhibited slightly higher levels of cyclin D1 expression compared
with the Rb/PrE cells; however, cyclin D1 increased in
both lines as they reenter the cell cycle (Fig.
7). In contrast, expression of cyclin E1 and PCNA is significantly elevated in Rb/PrE compared with the wild
type control. Both PCNA and cyclin E1 have been found in complex with
E2F transcription factors at the G1/S phase border of the
cell cycle and are believed to play a critical role in the activation
of several S phase-specific proteins (52, 53). Therefore, the loss of
Rb and liberation of E2F1 likely promotes the expression of
E2F target genes, such as cyclin E1 and PCNA, that in turn drive the
cells into DNA synthesis, resulting in the enhanced growth kinetics
profile, as seen in the Rb/PrE cells.
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 7.
Cyclin analysis during re-entry into cell
cycle. PrE and Rb/PrE cells were cultured in the absence of
serum for 10 days. Serum-containing medium was added back at time 0, and the cultures were harvested at the indicated times. Protein lysates
were prepared and analyzed for cyclin E1, D1, and PCNA as compared and
reported by Western analysis where actin (1) is
the loading control for cyclin E1, D1 and actin
(2) is the loading control for PCNA. w.t., wild
type.
|
|
Rb/PrE Cells Have an Increased DNA Ploidy--
To understand
better the functional consequences of Rb loss on cell cycle
regulation in prostate epithelial cells, we evaluated the potential of
these cultures to undergo G1 arrest under high density
culture conditions in the presence of serum growth factors. PrE and
Rb/PrE cultures were grown in 5% FBS and analyzed at subconfluent
and high density culture conditions by flow cytometry. At subconfluent
culture conditions, there was no significant difference in the
distribution of cells in the various phases of the cell cycle between
PrE and Rb/PrE cells. When PrE and Rb/PrE cultures were
maintained at high density (15 days) in 5% FBS growth medium, some
differences were noted in the DNA ploidy (Fig.
8, C and D). The
was an increase in DNA content of Rb/PrE cells compared with wild
type PrE cells as indicated by the increases of the mean G1
value of 124 versus 194 units in the Rb/PrE cells.
However, in high density cultures, the distribution of cells in
G1 is similar, with PrE and Rb/ exhibiting 61 and 70%
G1, respectively. In multiple experiments, we noted
differences in the distribution of cells in S phase, where the
Rb/PrE cultures had ~2-fold higher S phase content than wild type
PrE cells. Therefore the data shown in Fig. 8 revealed that whereas
both PrE and Rb/PrE cells have similar G1 distribution,
the Rb/PrE cells have increased DNA content (ploidy) as compared
with wild type.
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 8.
Flow cytometric analysis of log phase and
high density cultures of wild type and
Rb/PrE cells. Cell
cycle analysis of subconfluent, log phase PrE (A) and
Rb/PrE (B) cultures under serum-containing growth
conditions. C, high density culture (15 days) PrE;
D, Rb/PrE in serum-containing media. Flow cytometric
analysis was analyzed by multicycle software.
|
|
Karyotyping of Rb/PrE Cell Line--
To determine the extent
of chromosomal abnormalities that may have resulted following
Rb loss, chromosomal integrity of the Rb/PrE cells was
analyzed by spectral karyotyping (SKY). As shown (Fig.
9A), metaphase
chromosomes from two cell lines representing an early passage (passage
14) and a later passage (passage 25) were obtained by mitotic shake off
and subjected to SKY analysis. Both passages were hypertriploid with
gains corresponding to 4-6 copies of chromosome 19 in all cells and a
recurring loss of chromosome 4 (2 copies of chromosome 4 in 60% of
cells). The Rb/PrE cells also revealed recurring but lower level
gains of chromosome 15, as well as a rare translocation involving
chromosomes Y and 17. The ISCN karyotype of the passage 14 Rb/PrE
cells showed the following: 59-70 XY, +X[10], +Y[9],
der(Y)t(Y;17)(B1;D1) [2], 4[6], +5[5], +8[3], +9[5],
+10[5], +11[4], +15[10], +16[4], +17[3], +18[3],
+19(×2)[10] (cp10). The ISCN karyotype of 10 cells of the
passage 25 Rb/PrE was similarly hypertriploid and had a ISCN karyotype as follows: 61-74 XY, +X[8],+Y[6], 4[6],
+5[5], 6[3], +7[2], +8[5], +9[6], +10[8], 11[6],
+12[5], +13[4], +14[3], 15[3], +15[3], +16[4],
+19(×2)[10] (cp10). No cells were karyotypically identical, and each
cell exhibited an average of 8 chromosomal gains or losses from a modal
number of 3.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 9.
SKY analysis of
Rb/PrE cells.
A, metaphase analysis of passage 40 Rb/PrE cell line by
spectral karyotyping. Arrows show 6 copies of chromosome 19 in an otherwise near triploid cell line (arrows).
Asterisk shows recurrent translocation between Y and 17 seen
in 20% of cells. B, CGH analysis showing gains of
chromosome 19 and loss of chromosome 4 in passage 40 Rb/PrE cells.
Computed profiles of chromosome 4 and 19 show the degree of loss or
gain, respectively, compared with sex- and age-matched +/+ DNA. Mode
value is black line, and gain and loss of one copy are
depicted as fine green and red lines,
respectively. Blue line represents the DNA profile of
sample. The green line indicates gains, and the red
line indicates loss, shown next to idiogram.
|
|
To assess the above changes on the genomic DNA level, CGH was performed
with DNA extracted from Rb/PrE cell lines at passages 10, 20, and
40, and control DNA was extracted from strain-matched wild type mouse
DNA. Results showed that all three Rb/PrE DNA samples from passages
10, 20, and 40 shared loss of chromosome 4 and gains of chromosome 19, although by passage 40 there was significant amplification of
chromosome 19 as well as one copy gains of chromosomes 6, 11, and 15. These data strongly suggest that the homozygous loss of Rb
resulted in chromosomal changes including the loss of mouse chromosome
4 and significant gains of chromosome 19 over the wild type control
(Fig. 9B and data not shown).
| |
DISCUSSION |
Until now, the lethal nature of the Rb knockout
precluded the establishment of Rb/ epithelial cell
lines, where definitive experiments to investigate physiological roles
for Rb in specific epithelial populations could be
performed. The current study describes the successful establishment of
an Rb/ prostate epithelial cell line that was
rescued from fetal urogenital precursor tissue. The
resultant cell line, termed Rb/PrE, was utilized for the physiologic examination of Rb deletion in a specific
epithelial population.
Historically, the use of transforming oncogenes, such as E1a, E6/E7,
and large T antigen, have been useful to address Rb
function; however, these reagents do not exclusively target
Rb and promote such genomic instability that experimental
interpretation is difficult. The chromosomal imbalances directly
influenced by viral oncogenes have been identified as predominantly
random genetic events (31-33). One of the central objectives of this
study was to delineate and characterize the physiological function of
pRb in a prostate epithelial population, if possible, with minimal
complications of genetic instability inherent in tumor cells and cells
transformed with viral oncogenes. Therefore, the state of chromosomal
integrity was essential in the characterization of the Rb/PrE line.
SKY analysis revealed that the deletion of Rb in prostate
epithelium gave rise to aberrations that are consistent with
immortalization. Aneuploidy in the Rb/PrE cells characterized by
the loss of chromosome 4 and gains of chromosome 19 were the prominent,
recurring events. When these cells were analyzed at passages 14 and 25, Rb/PrE cells were hypertriploid with 59-70 and 61-74 chromosomes per cell, respectively, compared with a hyperdiploid state of 42-43
chromosomes per cell in the PrE wild type controls at passage 14 (data
not shown) where normal mouse cells have 40 chromosomes. At passage 72 the Rb/PrE cells were hypertriploid to hexaploid exhibiting 61-136
chromosomes. Our control, wild type PrE cell line, which had
spontaneously immortalized in culture, also demonstrated loss of
chromosome 4 and one copy gains of chromosomes 2, 11, 18, and 19. The
copy number of chromosome 19 in the Rb/PrE cells increased with
passage number up to 12 copies of chromosome 19 detected at passage 72. Rb/PrE cells also demonstrated recurring but lower level gains of
chromosomes 5, 6, 11, 15, and 18 at passages 20-72. In addition to
these aberrations, the early passage of Rb/PrE cells also had a
rare (i.e. 2 of 10 cells) structural aberration involving
chromosomes Y and 17.
The gain of chromosome 19 was more prominent in the Rb/PrE cells
than in the wild type PrE cells (12 copies versus 1) and may
reflect a more specific outcome of Rb disruption. Trisomy 19 has been implicated in early tumorigenesis in murine hepatocellular carcinoma and disorders of the murine central nervous system (54, 55).
Furthermore, gains of hamster chromosome 3q, which shares homology with
mouse chromosome 19, have also been associated with immortalization
in vitro (56, 57). These data suggest that the high level
gain of chromosome 19 may convey an early selective advantage on
the Rb/PrE cells due to an increased copy number of
growth-related genes and/or true oncogenes that reside on chromosome 19.
The loss of chromosome 4 is a recurring event in both the wild type PrE
cells and Rb/PrE cells and, at a threshold level, may result in
part from unrelated events due to culturing. However, evidence does
exist that suggests a selective pressure may, in part, be caused by
this Rb null genotype. Although the Rb/PrE cells are not
completely immune to genetic alterations, the chromosomal changes
described here are minimal and reflect a specific increase in
chromosome 19. The Rb/PrE cells are not prone to the more frequent
genetic translocations attained in cells associated with p53 mutation
tumorigenesis, viral-oncogenic transformation. The finding is that the
Rb/PrE line, although susceptible to chromosomal gain, does not
exhibit chromosomal rearrangement or translocation due to an intact p53
repair mechanism.
Numerous studies (9-11) have suggested that pRb plays an essential
role in embryonic development, and the deletion of Rb in a
variety of models resulted in marked abnormalities in the
differentiation of specific cell types (5-7). In this study, we have
demonstrated that despite the lack of pRb protein expression,
Rb/PrE cells continue to express markers of terminally
differentiated prostatic epithelium and that these cells are fully
capable of recapitulating normal prostatic morphogenesis in
vivo, complete with expression of prostate-specific secretory
proteins. It has been suggested that cellular differentiation can be
divided into the following three general steps: cell cycle exit,
protection from apoptosis, and tissue-specific gene expression (58).
Our findings suggest that the loss of Rb does override the ability of
Rb/PrE cells to growth arrest. A role for Rb disruption in
immortalization is more strongly suggested by the in vivo
experiments demonstrating that only the Rb/PrE grafts and not wild
type grafts were immortal. The Rb/PrE cells were also able to
survive in serum-free media, a condition that eventually induces cell
death in the wild type controls. Finally, we found that Rb
loss does not affect the ability of prostate epithelium to undergo
normal prostatic histodifferentiation and recapitulate prostate
morphogenesis. This is in agreement with another study demonstrating
that the loss of Rb did not adversely affect the normal development of
murine mammary gland (37).
The Rb gene product regulates the transition between
G1 and S phases in the cell cycle and functions in
transducing growth-inhibitory signals that arrest cells in
G1. Deletion of Rb in a variety of cancer cell
types has been associated with a deregulated cell cycle and
endo-reduplication (59). When PrE and Rb/PrE cells were analyzed
for differences in expression of cell cycle regulatory proteins,
enhanced expression of cyclin E1 in conjunction with increased PCNA
levels was noted in the Rb/PrE cells, whereas cyclin D1 levels were
reduced. Increased expression of cyclin E and PCNA in the Rb/PrE
cells was likely due to the liberation of E2F and subsequent activation
of transcription. The finding that cyclin D1 was repressed in the
Rb/PrE cultures might be explained by an independent
transcriptional mechanism by which E2F-1 and SP1 cooperate to repress
cyclin D1 transcription at specific sites in the cyclin D1 promoter
(60). The observation that E2F may regulate such opposing outcomes of
different cell cycle regulatory targets suggests a clear dissociation
of these two pathways in prostate epithelium. The Rb/PrE cells are
more active in DNA synthesis, and this may be attributed to the
increased cell growth kinetics and loss of growth arrest potential of
these cells. These results support the hypothesis that the loss of
Rb does result in a more proliferative and aneuploid
phenotype and contribute to alterations in cell cycle regulatory
proteins possibly through a constitutively active E2F1. A downstream
effect of this enhanced proliferative activity may result in the
compression of the G2/M phase as noted in the Rb/PrE
cultures at high density. This experiment provides an excellent example
of the regulatory control provided by pRb at the G1/S
border where, in the absence of pRb, E2F is free to activate genes such
as cyclin E1 and PCNA that drive DNA synthesis. These findings suggest
that the loss of Rb on a specific epithelial population may
circumvent growth inhibitory constraints that support the immortalized
phenotype and promote epithelial survival in the absence of growth factors.
Cellular senescence has been associated with a reduction in telomere
length. This hypothesis was supported by studies demonstrating a direct
link between limited cell division and progressive telomeric shortening. It was also postulated that maintenance of telomere length
by telomerase might override senescence. Such a causal role for
telomerase in cellular senescence has been demonstrated by transfection
of the human telomerase reverse transcriptase cDNA into normal
cells resulting in their ability to bypass senescence and prolong the
life span in vitro (61). We found that telomerase activity
was elevated in both Rb/PrE, but because telomerase activity was
also elevated in the spontaneously immortalized control wild type PrE
cells, it is impossible to conclude that there is a specific regulatory
role for Rb in telomerase expression. These results do suggest,
however, that telomerase activity may be important in the
immortalization of mouse epithelial cells.
The similarities and differences between pRb, p130, and p107 are
apparent in mice carrying single or compound knockouts of the
corresponding genes. When the Rb gene is deleted through
homologous recombination, embryos die at 13 days of gestation from
defective development of erythroid and neuronal tissues (9-11). In
stark contrast, targeted disruption of either p107 or p130 is not
lethal and does not result in an obvious phenotype (20, 62). Recent studies (19, 48) have demonstrated that mouse embryonic fibroblasts (MEFs), harboring a triple knockout of pRb, p107, and p130, are resistant to G1 arrest signals and do not undergo
senescence in culture. These cells also exhibit some features of
transformed cells, such as focal proliferation on monolayers and
anchorage-independent growth; however, individual mutants do not
undergo transformation. As with the Rb/ fibroblasts,
Rb/ prostate epithelial cells do not exhibit
characteristics of transformation. Although mutation of individual Rb
family members in MEFs resulted in minor alterations in both cell cycle
regulation and DNA damage response, these cells still remained
sensitive to G1 arrest signals and subsequently entered
into senescence. In general, the cell cycle studies of Rb/PrE
epithelial cells supported those findings in MEF cells, in that Rb
deletion resulted in only minor changes in cell cycle regulation. Taken
together, these results suggest that cell cycle regulation and
cellular transformation of Rb/ MEFs and Rb/PrEs are not attained by a single mutation of the Rb gene and is
likely influenced by the simultaneous disruption of all three pocket protein family members. In contrast to the findings in
Rb/ MEFs, our results suggested that the loss of Rb does
support the immortalization of prostate epithelium, particularly
in vivo.
The Rb gene encodes a key regulatory component of the
cell cycle that is frequently disrupted in many human cancers,
including adenocarcinoma of the prostate gland (reviewed in Ref. 50). In a recently published model of mouse prostate cancer (34), the use of
Rb/PrE prostatic tissue and its response to hormonal carcinogenesis
were described. In that study, Rb/ prostatic tissue was
highly susceptible to hormone-induced malignant transformation. This
model has drawn much interest due to the recapitulation of several key
features of human prostate cancer, namely in its progression from
dysplasia to carcinoma accompanied by the loss of the basal epithelium.
The role of pRb in human neoplasia, including prostate carcinoma, has
been the subject of rigorous investigation for a number of years;
however, the specific function of pRb in the
,
,
TOP
ABSTRACT
INTRODUCTION
