establishment of two human prostate cancer cell lines derived … · vol. 3. 2493-2500. dee,nber...
TRANSCRIPT
Vol. 3. 2493-2500. Dee,nber 1997 Clinical Cancer Research 2493
Establishment of Two Human Prostate Cancer Cell Lines Derived
from a Single Bone Metastasis1
Nora M. Navone,2 Matilde Olive, Mustafa Ozen,
Rodney Davis, Patricia Troncoso, Shi-Ming Tu,
Dennis Johnston, Allan Pollack, Sen Pathak,
Andrew C. von Eschenbach, and
Christopher J. LogothetisThe University of Texas M. D. Anderson Cancer Center. Houston.
Texas 77030
ABSTRACTHuman prostate cancer cell lines are particularly diffi-
cult to establish, and most existing cell lines do not exhibit
features commonly seen in human prostate cancer. Most
available models either grow only in vivo as xenografts or
are androgen insensitive and fail to express prostate-specific
antigen (PSA). The lack of functionally relevant model sys-
tems of advanced prostate cancer has limited prostate can-
cer research and therapy development. Of 30 processed
samples derived from patients with prostate cancer, we
established two cell lines (MDA PCa 2a and MDA PCa 2b)
that express PSA and androgen receptor, grow in vitro, and
are androgen sensitive. Cells from these lines produced tu-
mors in nude mice when injected either s.c. or orthotopically
(intraprostatic). Both cell lines were established from a bone
metastasis of a patient whose cancer was exhibiting andro-
gen-independent growth. Although both were derived from
two samples of the same specimen, they have different ge-
netic features (as assessed by karyotype analysis) and dif-
ferent phenotypes (e.g., morphology and growth rate). It is
likely that they are distinct clones isolated by the use of
different culture procedures and reflect the genetic hetero-
geneity of the tumor. These new cell lines are the first
available derived from a bone metastasis of an androgen-
independent prostatic adenocarcinoma that grow both in
vivo and in vitro and have retained PSA expression and
androgen sensitivity. They therefore constitute important
model systems to address critical questions related to the
Received 5/1/97: revised 8/19/97: accepted 8/25/97.
The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked
adi’ertise,nent in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
I This work was supported in part by grants from CaP CURE (Associ-
ation for the Cure of Cancer of the Prostate) and the John S. Dunn
Research Foundation of Houston, Texas. and by NIH Grant RRO 4999-
01.2 To whom requests for reprints should be addressed. at Department of
GU Medical Oncology. M. D. Anderson Cancer Center, I 5 15 HolcombeBoulevard, Box 13, Houston, TX 77030. Phone: (713) 792-2830: Fax:
(713) 745-1625.
androgen-independent growth of human prostate cancer
and to the complex process of bone metastasis.
INTRODUCTIONProstate cancer is the most frequently diagnosed and the
second most deadly cancer in men ( I ). Although localized
prostate cancer may be cured by radical prostatectomy or radi-
ation therapy. the prospect ofcure diminishes significantly when
prostate cancer cells escape beyond the confines of the gland (2.
3). Patients with regional or distant metastases are usually
treated with androgen deprivation therapy. Although most pa-
tients initially respond to this therapy. given sufficient time it is
believed that all eventually relapse with androgen-independent
disease. The androgen-independent phenotype of human pros-
tate cancer is defined in the clinic as the ability of prostate
cancer cells to grow despite castrate plasma concentrations of
testosterone (2-4). No current standard therapy consistently
confers a significant survival benefit to patients who develop
androgen-independent prostate cancer (2-4).
The lack of functionally relevant in vitro models has been
one of the major obstacles to advancing our understanding of the
molecular mechanisms of this disease as well as to developing
effective therapies. Most available human prostate cancer mod-
els either grow only in vito as xenografts (5-9) or (with the
exception of the LNCaP cell line: Ref. 10) lack some of the
common features of clinical prostate cancer (androgen sensitiv-
ity and expression of PSA: Refs. 5-9).
The purpose of this work was to develop model systems of
advanced prostate cancer. Our specific objectives were: (a to
screen human prostate cancer tissue samples derived from mul-
tiple sites for their growth in vitro and in vito: (h) to optimize
their growth conditions: and (c) to characterize the cells grown
ii? vitro. We processed 30 tissue samples derived from primary
prostate cancer of various histological grades and from prostate
cancer tissue derived from different metastatic sites. Initial
attachment and cell growth were obtained in cells from nine
specimens: of these only two were established as cell lines
(MDA PCa 2a and MDA PCa 2b). The cell lines were derived
from a bone metastasis of a patient whose cancer was clinically
progressing despite a castrate concentration of testosterone and
resulting in spinal cord compression requiring a laminectomy.
Both lines express PSA3 and AR. grow in i’ii’o and in vitro, and
are androgen sensitive. These are the first cell lines available
derived from a bone metastasis of prostate cancer that express
PSA and are androgen responsive. The lines are also the second
and third human prostate cancer cell lines to produce PSA and
are androgen sensitive. Our cell lines have now grown in vitro
3 The abbreviations used are: PSA. prostate-specific antigen: AR. an-
drogen receptor: DHT. dihydrotestosterone: EGF. epidernial growth
factor: FBS. fetal bovine serum.
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
2494 New Prostate Cancer Cell Lines
for 24 months (MDA PCa 2a is in passage 56, and MDA PCa 2b
is in passage 50) and have been used to establish tumors in nude
mice.
MATERIALS AND METHODS
Quality Control
Samples from different areas of the primary tumors ormetastatic lesions were obtained for tissue culture. Tumor sam-
pies ranged from 1.0 to 0.5 cm in the largest dimensions. A
tissue section immediately adjacent to that submitted for culture
was obtained for detailed histopathological analysis and deter-
mination of percentage of tumor in the sample.
Collection and Initial Processing of Tissue Specimens
Aseptically removed specimens were kept in cold (4#{176}C)
sterile tubes with HEPES-buffered saline ( 1 1) until processed.
The specimens were mechanically disaggregated with sterile
crossed scalpels to obtain pieces of about 1 mm3. Many of the
samples disaggregated into small clumps of tumor cells in the
washing medium as they were cut. In these cases, subsequent
enzymatic disaggregation of the tissue was not necessary. In
other cases, after cutting, the tissue was resuspended in five
volumes of digestion solution [Collagenase Sigma type 1 (40
units/mi of growth medium); Sigma Chemical Co., St. Louis,
MO] and incubated at 37#{176}C.The length of the incubation varied
from 30 mm to overnight, depending on the degree of fibrosis of
the specimen. Effective disaggregation of the tissue was moni-
tored over time by gently moving the flask.
Culture of Tissue Specimens
Every tissue specimen to be cultured was incubated with at
least three different media [HITES (12), BRFF-HPC1, and
PFMR4-1A (1 l)}. HITES medium is modified RPMI 1640
(GIBCO BRL, Gaithersburg, MD) with hydrocortisone, insulin,
transferrin, and selenium. (Estradiol was not included in our
formulation.) BRFF-HPC1 medium (Biological Research Fac-
ulty and Facility, Inc., Jamsville, MD) has a very rich formula-
tion and includes cholera toxin, DHT, insulin, EGF, and hydro-
cortisone. The formulation for PFMR4-1A (kindly provided by
Dr. Donna Peehl) includes cholera toxin, insulin, EGF, and
hydrocortisone (1 1). When enough tissue was available, we
cultured the isolated cells with and without serum. In most
cases, the presence of serum was necessary to obtain initial
attachment and cell growth. Collagen-coated dishes were also
used to encourage cell attachment and growth (1 1). Daily mon-
itoring of the cultures allowed us to alter culture conditions as
necessary.
In Vitro Cell GrowthTo determine population doubling times of the MDA PCa
2a (passage 30) and MDA PCa 2b (passage 30) cell lines, cells
were seeded onto 6-well plates at a density of 4 X 10” cells/well
in BRFF-HPC1 medium with 20% FBS and in F12K(+) me-
dium supplemented with 20% FBS. F12K(+) is F12K (Life
Technologies, Inc., Gaithersburg, MD) supplemented with 25
ng/ml cholera toxin, 10 ng/ml EGF, 40 mn phosphoetha-
nolamine, 10 pg/mI hydrocortisone, 45 m�t selenous acid, and 5
mg/mi insulin. The sources and preparation of these supple-
ments have been described previously (1 1). The cultures were
harvested at days 7 and 14, and the cells were counted using a
hemocytometer. Population doublings were calculated as the
average of three independent cell counts derived from different
wells. The anchorage-independent colony-forming assay in
semisolid medium was performed as described previously (13).
Cell Growth Induction by Androgens
The kinetics of growth induction of the MDA PCa 2a and
MDA PCa 2b cell lines by DHT was determined by the crystal
violet mitogenic assay (14, 15). Cells were seeded on 24-well
plates at a density of iO� cells/well and allowed to attach for
48 h in F12K(+) with 20% FBS. On the third day, designated as
day 0, the cultures were washed twice in PBS and cultured for
7 or 14 days in F12K(+) with 15% charcoal-stripped FBS and
10� t2 M tO 106 M DHT (Sigma). The medium was changed on
days 3, 7, and 10. Each concentration of DHT was assayed in
triplicate. The mitogenic activity of DHT was determined by the
uptake and elution of crystal violet dye by the cells in each well
(14, 15). A small aliquot of eluted dye was placed in a 96-well
plate, and the absorbance of each well at 560 nm was measured
with a Titerek Muitiscan TCC/340 (Flow Laboratories, McLean,
VA). Control experiments demonstrated that the absorbance was
directly proportional to the number of cells in each well.
Tumorigenicity of MDA PCa 2 Cells in a Nude Mouse
Assay
Six- to 8-week-old male athymic (nude) mice (BALB/c
strain; Charles River Laboratory, Wilmington, MA) were used
for the tumorigenicity assay of MDA PCa 2 cells. The animals
were housed aseptically in the Athymic Animal Facility of the
M. D. Anderson Cancer Center and manipulated under surgical
aseptic conditions in a laminar flow hood. Cells to be injected
into the mice were trypsinized, washed, counted, and resus-
pended at 4-S x 106 cells/l00 �iI of growth medium for the s.c.
injection. To perform the intraprostatic injection, 1-2 X 106
MDA PCa 2a cells were suspended in 20-30 p.1 of growth
medium.
s_c. Injection. Four mice were each injected s.c. with 5 X
106 MDA PCa 2a cells/site at each of four sites. Two sites were
injected with Matrigel ( 1 : 1 v/v; Collaborative Biomedical Prod-
uctlBecton Dickinson, Bedford, MA) and two with growth
medium. Six mice were each injected s.c. with 4 X 106 MDA
PCa 2b cells at two sites. Three mice received cell injections
with Matrigel (1 : 1 v/v) and three with growth medium.
Intraprostatic Injection. Fourteen mice were injected
with 2 X 106 MDA PCa 2a cells into the prostate (20-30 �i.l).
Mice were anesthetized with methoxyflurane, and a lower mid-
line incision was made. Tumor cell suspensions were injected
into the dorsal prostatic lobes using a 30-gauge needle with a
1-mi disposable syringe and a calibrated push button-controlled
dispensing device (Hamilton Syringe Company, Reno, NV).
The abdominal wound was closed in one layer with wound clips
(Autoclip; Clay Adams, Parsippany, NY).
Surgical Castration of Male Mice
The animals were anesthetized with methoxyflurane under
sterile conditions. A midline or bilateral scrotal incision was
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 2495
made to expose the testicles. The testis were pulled away from
the surrounding connective tissue until the spermatic cord was
visible. The cord was ligated 1-2 cm above the junction of the
cord and the testis. The cord was subsequently cut, the testis was
removed, and the scrotum was closed with wound clips (Auto-
clip; Clay Adams) or nonabsorbable sutures. The inhalation
agent was removed before closure of the scrotum to reduce the
recovery period. Each animal was observed until it was able to
flee or defend itself. No postanesthesia analgesics were given.
The incision site was examined 24 h after the procedure.
PSA Blood Levels
Blood from the mice was obtained at regular intervals from
a small incision in the main tail vein. Serum was separated from
the blood, and PSA was measured using a dual-reactive enzy-
matic immunoassay kit with a lower limit of sensitivity of 0.4
ng/ml (Hybritech, Inc., San Diego, CA).
Chromosome Preparation and Giemsa Banding
Freshly fed (24 h earlier) and approximately 80% confluent
cell cultures were used for chromosome preparations. Prior to
harvest, cells were treated with Colcemid (final concentration,
0.04 p.g/mi) for 30 mm at 37#{176}C.After centrifugation. the cell
pellet was treated with a 0.06 M KC1 hypotonic solution for 20
mm and fixed in acetic acid-methanol (1 :3 by volume). Slides
were prepared following the routine air-drying procedure. 0p-
timaily aged slides were used for the induction of Giemsa
banding (16). At least 15 metaphase plates from each cell line
harvests were photographed using the Genetiscan (PSI, Inc.,
Houston, TX) and four metaphases from each sample were cut
and karyotyped in detail.
Protein Extraction and Western Blot
Protein extracts from cells growing in monolayer were
obtained using TRIPA buffer and following standard procedures
(17). Cell lysates were centrifuged at 100,000 X g for 45 mm at
4#{176}C.The supernatants were normalized for protein content, and
50-100 �i�g of protein per lane (depending on the molecular
weight of the protein to be detected) was fractionated on 7.5-
12% SDS polyacrylamide. Gels were blotted to a Hybond-ECL
nitrocellulose membrane (Amersham Corp., Arlington Heights,
IL), which was probed and washed following the instructions for
the enhanced chemiluminiscence Western blotting detection
system (Amersham Corp.).
Statistical Methods
The relationship between tumor size and PSA production
before and after castration was examined using the analysis of
covariance (18).
RESULTS
In 19 months, we processed 30 tissue samples, including
specimens derived from primary prostate cancer of various
histological grades. as well as prostate cancer tissue derived
from different metastatic sites. Initial attachment and cell
growth were obtained in nine specimens, but only two human
prostate cancer cell lines, MDA PCa 2a and MDA PCa 2b, have
been successfully established. These new cell lines, both ob-
tamed from one patient, express PSA and AR, grow in vito and
in vitro, and are androgen sensitive.
Clinical History
The patient was a 63-year-old black male who presented
with impending T9 cord compression at an outside hospital in
September 1993. He underwent spinal decompression and tumor
resection followed by bilateral orchiectomy. Pathological exam-
ination showed a poorly differentiated adenocarcinoma. He
came to M. D. Anderson Cancer Center in October 1993 with a
near-complete neurological recovery and good pain control.
He did well until June 1995, when he again showed evi-
dence of progression with a rising PSA in the presence of
castrated plasma levels of testosterone (14 ng/dl). He was mi-tiated on treatment consisting of suramin and doxorubicin. A
lesion at Tl2 continued to grow during treatment and was
encroaching on the spinal cord, causing significant right flank
pain. In August 1995, he underwent elective resection of a T12
paraspinal mass. At that time, his PSA was 1 14 ng/ml. Patho-
logical examination revealed metastatic adenocarcinoma. The
tumor showed strong staining for PSA. No staining for chro-
mogranin was present. The resected tumor fragments measured
in aggregate 12 X 12 X 5 cm. Two different areas of grossly
viable tumor were sampled for tissue culture studies. The con-
trol sections for both samples showed histologically similar
carcinoma.
The patient received postoperative radiation therapy but
continued to deteriorate clinically. Palliative estramustine and
etoposide were given without benefit. He developed pulmonary
lymphangitic spread of tumor in his lungs and suffered from
significant bone marrow suppression. He died of complications
from his cancer in December 1995. His PSA at the time of death
was 1158 ng/mi.
Initial Isolation of Epithelial Cells and in Vitro Growth
Two samples from different areas of the tumor resected on
August 1995 were used for culture. They were processed ac-
cording to the procedure described in “Materials and Methods”
and digested with coilagenase (Sigma type I , 40 units/mI in
growth media) at 37#{176}Cfor 3 h.
We used different culture conditions in the two tumor
samples to isolate the epithelial cells from contaminant stromal
cells and to obtain in vitro growth. The two conditions that were
successful are detailed below.
Isolation of Cells Growing in Suspension. MDA PCa 2a
cells were initially maintained in HITES with 10% FBS for 1
month. Epithelial cells did not attach under those conditions
and, therefore, were effectively separated from fibroblasts. After
1 month, the nonadherent cells were transferred to a collagen-
coated dish with BRFF-HPC1 medium and 10% FBS. After
transfer to that medium, the cells grew as monolayers.
Use of Low Percentage Serum in the Growth Medium
to Favor Epithelial Growth over Fibroblasts. MDA PCa 2b
cells were grown as a monolayer in the presence of fibroblasts
in BRFF-HPC1 supplemented with 2% FBS for 2 months until
the fibroblasts died, at which time the FBS was increased to
10%.
The MDA PCa 2 cell lines have sustained continuous in
vitro growth as monolayer cultures since their initial isolation
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
: �� �-,,
.�‘- �
����:--- #{149}.
.. ��:
. .-�.... e._�. �
,: � � ;:�;.�:
2496 New Prostate Cancer Cell Lines
Fig. 1 Phase-contrast micrographs of human prostate cancer cell lines.A, MDA PCa 2a passage 21 ; B, MDA PCa 2b passage 15. Note thatMDA PCa 2a cells are larger and, unlike MDA PCa2b, form tightintercellular junctions.
(24 mo). MDA PCa 2a is in passage 56, and MDA PCa 2b is in
passage 50. MDA PCa 2a and MDA PCa 2b are propagated in
either BRFF-HPC1 with 20% FBS or F12K(+) with 20% FBS.
Several vials of each cell line (5 X 106 cells/vial) have been
frozen at every third passage. Recovery of frozen cells after 6
months was good; 60-70% of cells were viable and grew. The
cells are free of Mycoplasma as assessed by PCR amplification
of tissue culture supernatant using genus-specific PCR primers
(19).
MDA PCa 2 cells were originally derived from different
areas of the same tumor specimen and have some elements of
morphology in common; they grow in clumps, form layers, and
never reach confluence. However, as shown in Fig. 1 , there are
some morphological differences between the two cell lines.
MDA PCa 2a cells are bigger, and unlike MDA PCa 2b, form
tight intercellular junctions as seen under light microscopy.
The population doubling times also differed in the two cell
lines. For MDA PCA 2a growing in BRFF-HPC1 with 20%
FBS, the population doubling time was 82 h at 7 and at 14 days.
When grown in F12K(+) with 20% FBS, the doubling time was
87-93 h at 7 and at 14 days. For MDA PCa 2b cells growing in
BRFF-HPC1 w/20% FBS, the doubling time was 42-53 h at 7
and at 14 days; growing in F12K(+) w/20% FBS, it was 6 1-73
h at 7 and at 14 days. The difference in growth rates and the
morphological differences described above suggest that the two
cells lines originated from different clones within the tumor
isolated by the different culture methods.
Clonal Growth Assay in Soft Agarose
Both MDA PCa 2a and MDA PCa 2b formed colonies in
soft agarose. Their colony-forming efficiency was 2-4%, which
is within the expected range for human cancer cell lines.
Karyotype Analysis of MDA PCa 2a and MDA PCa 2b
Giemsa-banded karyotype analysis was performed on
these two cell lines before and after growth in nude mice.
Before injection in nude mice, MDA PCa 2a at passages 6
and 18 contained between 49 and 94 chromosomes, with a
mode at 63. There were five consistent clonal marker chro-
mosomes present in this cell line (Ml-M5), tentatively iden-
rifled as: Ml, del(3q); M2, t(5q;?;l5q); M3, del(8p); M4,
t(5q;7q;14q); and MS. del(17p12). A typical Y chromosome
was present in most metaphases, and some cells had more
than one copy of the Y chromosome. A karyotype showing
marker chromosomes and other anomalies is shown in Fig.
2a. MDA PCa 2b at passage 10 contained between 44 and 92
chromosomes, with a mode at 47. This cell line had seven
clonal marker chromosomes, including the five that were
found in MDA PCa 2a (Fig. 2b). The two new marker
chromosomes were tentatively identified as ml, del(2p) and
m2, t(4;?). A typical Y chromosome was absent in some cells
but not in others.
To characterize the cells that produced tumors in nude
mice, tumor samples from mice were minced, processed with
coliagenase, and grown in vitro for 3-6 weeks prior to harvest-
ing for karyotype analysis. This analysis showed human cells
with the same characteristic markers found in the original cell
lines.
Regulation of MDA PCa 2 Cell Growth by Androgens in
Vitro
To determine if the MDA PCa 2 cells were androgen
sensitive, we evaluated the in vitro mitogenic effects of DHT
in charcoal-stripped FBS. As shown in Fig. 3, peak responses
were seen with 1 X lO_8 M DHT after 14 days in MDA PCA
2a (P < 0.001) and after 7 days in MDA PCa 2b cells (P <
0.001), when compared with controls grown in hormone-free
medium.
Expression of AR and PSA in MDA PCa 2 Cells
Expression of AR and PSA was examined by Western blot
analysis of MDA PCa 2 cell lysates. The human prostate cancer
cell line LNCaP was used as a positive control for AR and PSA
expression (5, 10, 20, 21); the androgen-insensitive cell lines
PC3 and DU14S served as negative controls (5, 20, 21). As
shown in Fig. 4, both AR and PSA are expressed in each of the
MDA PCa 2 cell lines. These results are in agreement with the
androgen-sensitive phenotype of MDA PCa 2 cells described in
the previous section.
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
B
�#{246}� �,M3
� �:* � �4
1. � � I
�i( IiMl M5 ml m2
Fig. 2 Giemsa-banded karyotypes of human prostate cancer cell lines.
A, MDA PCa 2a at passage 6: B, MDA PCa 2b at passage 10. M.
markers (Ml to M5) common to both cell lines. ni. unique markers ml
and m2, present only in MDA PCa 2b. Both types of marker chromo-
somes from an additional metaphase spread are shown at the bottom of
each karyotype. Note the presence of two copies of chromosome Y in
MDA PCa 2a and an apparent absence of a typical chromosome Y inthis cell ofthe MDA PCa 2b. Unidentified markers (UM) present in both
cell lines are also shown.
Clinical Cancer Research 2497
A � � � “� I�
I I I;�3 I. III ;;�$� zA�
a.,, !�*� � ��‘ !p1:! #{216}�
UM
*1k!‘� .� �1 I� �4:;il �
ml Ml m2 M2
M5
.�e* �
Tumorigenicity of MDA PCa 2 Cells in an Athymic Mouse
Assay
MDA PCA 2a
The tumorigenicity of MDA PCa 2a was assessed by two
methods: s.c. injection and intraprostatic injection in athymic
mice.
s.c. injection. Seven weeks after injection, six of the
eight sites injected with tumor cells and Matrigel had a 47-170-
mm3 tumor. Nine weeks after injection, three of eight sites
injected with medium (no Matrigel) had tumors larger than 40
mm3. No metastases were detected by macroscopic examination
of mice organs at necropsy. Tumor volume was calculated as
length X width X height X 0.5236 (the formula for an ellip-
soid). PSA blood levels of the tumor-bearing mice were deter-
mined at regular intervals and were found to increase over time
after s.c. injection of MDA PCa 2a cells. At week 8, PSA blood
levels ranged from 40 to 150 ng/ml.
Intraprostatic injection. Ten weeks after injection, 10
of 14 mice injected showed increased values of PSA in their
blood (24-170 ng/mi). Eleven weeks after injection, 7 of 14
mice that received injections had palpable tumors. Seventeen
weeks after injection, the tumors were 200-800 mm3.
MDA PCa 2b
The tumorigenicity of MDA PCa 2b cells was assessed
through s.c. injection. Six weeks after injection, all six sites
injected with Matrigel had tumors ranging in size from 40 to
80 mm3; three of six sites injected with growth medium had
small tumors. The mice injected with Matrigel had serum
PSA values ranging from 370 to 667 ng/ml, whereas the mice
injected without Matrigel had values ranging from 53.4 to
129.5 ng/ml. Eight weeks after injection, all 12 sites injected
had tumors. Intraprostatic injections were not done with
MDA PCa 2b cells, due to the high rate of tumor induction
achieved through s.c. injections. In vivo, MDA PCa 2b grew
faster than MDA PCa 2a.
The ability of MDA PCa 2 cells to induce tumors that
secrete PSA as a function of tumor size in athymic mice is
compelling evidence that these cell lines originated from human
prostatic cancer. In addition, these results suggest that this
system is a useful model to study human prostate cancer. Ex-
pression of PSA is an invaluable marker of prostate cancer
progression (2-4), because tumor growth can be monitored by
PSA blood levels, and loss of PSA expression occurs only in
very late stages during the progression of the disease to a more
aggressive phenotype.
Effect of Castration of Mice Bearing Tumors Produced by
MDA PCa 2 Cells
The effect of castration on the in vivo growth of MDA PCa
2 cells was assessed by monitoring tumor size and PSA blood
levels after castration of mice bearing tumors produced by s.c.
injection of the cells. Male athymic mice were injected with 4 X
106 cells with medium (no Matrigel). Mice were castrated when
their tumors were 40 mm3 or larger. This occurred at 8-1/2
weeks in 3 of 10 mice injected with MDA PCa 2a cells and at
6 weeks in S of 9 mice injected with MDA PCa 2b cells. As
shown in Figs. 5 and 6, the tumors stopped growing or de-
creased in size, and PSA blood levels dropped dramatically after
castration of the mice injected with either cell line. These results
demonstrate that the MDA PCa 2 cell lines are androgen re-
sponsive.
PSA Blood Levels as a Function of Tumor Size
Four regression lines (MDA PCa 2a before and after cas-
tration, and MDA PCa 2b before and after castration) using
tumor weight to predict PSA were calculated. Because tumor
weight is a measure of cellular mass and PSA is normally
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
3.5
3
2.5
2
0 1.5
.5
0
3.5
3
2.5
�2
C
.5
0
0 10-12 10-’#{176} 10-8 10-60 10.12 10-10 10.8 10-6
DHT(M) DHT(M)
AR � llOkD
b- Actin � � .� 43 kD
PSA .� 33kD
Fig. 4 AR and PSA expression in human prostate cancer cell lines.Western blot analysis was performed with polyclonal antibody to human
AR (Novocastra) and with antibody to human PSA (Dako, Inc.,
Carpenteria, CA). �3-Actin (Mr 43,000) was used as a loading control.
DU145, negative control for AR and PSA; LNCaP, positive control for
AR and PSA; PC3, negative control for AR and PSA.
2498 New Prostate Cancer Cell Lines
MDA PCa 2a I 14 days MDA PCa 2b I 7 days
Fig. 3 Growth of MDA PCa 2cells in vitro in the presence
of different concentrations of
DHT. Each data point is the av-erage of three independent cx-
periments: bars, SD. O.D., ab-
sorbance at 560 nm.
c$ ..Oc�1 C�1
c8 c�
v�
zLI�
<
U�a
< rC9c1�
-�
analyzed as a logarithm, both weight and PSA were log (base
10) transformed for the analysis. The correlation of all re-
gression lines was above 0.89, and for every unit of log
increment in tumor weight, there was 2/5 of log increments in
PSA produced by MDA PCa 2a and 3/4 of log increments in
PSA produced by MDA PCa 2b. The amount of PSA pro-
duced by tumor weight remained the same in MDA PCa 2b
after castration but the relationship was lost in MDA PCa 2a
after castration. Based on these calculations, we estimated
that MDA PCa 2a produces 0.43 (± 0.15) ng/mi of PSA per
g of tumor, and MDA PCa 2b produces 0.67 (± 0.08) ng/ml
of PSA per g of tumor.
DISCUSSION
We have established two new human prostate cancer cell
lines, MDA PCa 2a and MDA PCa 2b. Evidence that these cell
lines originated from human prostatic cancer includes the pres-
ervation of PSA expression and androgen sensitivity.
MDA PCa 2 cells constitute an important model system for
the study of the clinically defined androgen-independent growth
of human prostate cancer. These cell lines were derived from a
patient with prostate cancer who relapsed two years after sur-
gical castration. At the time of relapse, the tumor from which the
cells were derived was actively growing, producing symptoms
that necessitated surgery. This objective evidence of disease
progression in the presence of castrate plasma levels of testos-
terone (14 ng/dl) is the clinical definition of androgen-indepen-
dent growth used at the University of Texas M. D. Anderson
Cancer Center and indicates the presence of an androgen-inde-
pendent tumor (4, 22).
Recurrent prostate cancer is a spectrum of diseases that
differ in clinical presentation and prognosis (2-4). A recent
report proposes a classification of prostate cancers as hormone-
naive, androgen-independent/hormone-sensitive. or hormone-
independent based on their hormone sensitivity (4). Androgen-
independent/hormone-sensitive cells are categorized as
androgen-independent because they grow in the presence of
castrate plasma levels of testosterone, but their growth rate
responds to other hormonal manipulations (e.g. , adrenal andro-
gen blockade and antiandrogens). The importance of this dis-
tinction is that it is generally recognized that a subpopulation of
patients who have relapses after one hormone treatment may
still respond to a second hormone treatment (4). The MDA PCa
2 cells, which are derived from an androgen-independent tumor
but are sensitive to androgens in vitro and in vito, are presumed
to represent the androgen-independent/hormone-sensitive tumor
type.
To our knowledge, the LNCaP cell line is the only other
existing human prostate cancer cell line that is androgen sensi-
tive. LNCaP was derived from a metastatic site (supraclavicular
lymph node) of a patient with an androgen-independent pros-
tatic adenocarcinoma (10). In contrast to MDA PCa 2a and 2b,
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A VCastration B150�
Castration
Fig. 5 Tumor size (A) and PSA blood levels
(B) produced by s.c. injection of MDA PCa 2a
cells into athymic mice. At 8-1/2 weeks after
cell injection, 3 of 10 mice that received injec-
tions had tumors larger than 40 mm3. Each
symbol represents a different mouse. The ar-
row indicates the time when the mice were
castrated. Note that after castration of the mice,
the tumors decrease in size or stop growing,
and the PSA blood levels drop dramatically.
A Castration
Sc N � 0’� © � r� � ‘4 (0 �
- ,- ,- ,-
Weeks after cell injection Weeks after cell injection
300�
250�
200�
150�
100�
50�
CastrationB
E
a
.�
00
(�1 ‘l� �C 00 ©
Weeks after cell injection
Fig. 6 Tumor size (A) and PSA blood levels (B)produced by s.c. injection of MDA PCa 2b cellsinto athymic mice. At 6 weeks after injection. five
of nine mice that received injections had tumors
larger than 40 mm3. Each symbol represents a
different mouse. The arrow indicates the time
when the mice were castrated. Note that aftercastration of the mice, the tumors decrease in sizeor stop growing, and the PSA blood levels dropdramatically.
Weeks after cell injection
Clinical Cancer Research 2499
EE
0
Ea
EE
1�
0
Ea
the LNCaP cell line is only weakly tumorigenic when injected
s.c. into male nude mice, but the tumorigenicity can be increased
by coinoculation with Matrigel or fibrobiasts (23). The complex
role of androgens in prostate cancer progression has been high-
lighted in a recent publication (24). In that report, an androgen-
repressed human prostate cancer cell line derived from a patient
with androgen-independent prostate cancer was characterized.
Those findings imply that a subpopuiation of cells is inhibited
by androgens. The authors suggest that the androgen-repressed
phenotype may represent late events in prostate cancer progres-
sion (24).
A clinical hallmark of prostate cancer progression to the
androgen-independent state is the development of osseous me-
tastases: up to 84% of patients have bone metastases at autopsy.
This is the only malignancy to consistently produce osteobiastic
lesions (2, 23, 25). The marked clinical affinity that prostate
cancer has for bone is responsible for much of the clinical
morbidity that is a result of this cancer. Our cell lines may serve
as a useful model of the complex process of prostate cancer
metastasis and progression within bone. The availability of this
new model system is essential for the study of these interactions.
The marker chromosomes characteristic of these prostate
cancer cell lines contain regions that have been identified by
allelotyping studies and comparative genomic hybridization
analysis as sites of common allelic gains or losses in clinical
prostate cancer specimens, i.e., 8p (26-28), 7q (27, 29-32), Sq
(27, 28), 3q (27), 2q (27), and l7p (22, 27). Ofparticular interest
is the presence of an interstitial or intercalary deletion of Yip 12,
with preservation of Y/pl 1 and Ylpl3. This may suggest the
presence of an additional tumor suppressor gene in 17p along
with p53, which is located at Y7pl3.l (22). It is also of interest
that Cher et a!. (27) found an increased frequency of ailelic
gains in the 4q25-28 region among African-Americans, because
the MDA PCa 2 cell lines which were derived from an African-
American patient have three to four copies of chromosome 4,
reflecting amplification of the 4q25-28 region. A marker chro-
mosome involving chromosome 4 was also present in the MDA
PCa 2b cell line.
The MDA PCa 2 cell lines, although obtained from differ-
ent samples of the same specimen, have unique genetic features
(as assessed by karyotype analysis) and phenotypes (different
morphology and different growth rate). It is likely that these two
cell lines are distinct clones that reflect the genetic heterogeneity
of the tumor and will provide important model systems to
cellular interaction that may participate in prostate cancer pro-
gression within bone.
ACKNOWLEDGMENTSWe gratefully acknowledge the technical assistance of Dr. P. A.
Steck in the colony-forming assay and C. R. Sikes for technical assist-
ance in the surgical procedures in mice.
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
2500 New Prostate Cancer Cell Lines
REFERENCES1. Parker, S. L., Tong. T., Boiden, S., and Wingo, P. A. Cancer
Statistics. CA Cancer J. Clin., 47: 5-29, 1997.
2. Catalona, W. J. Management of cancer of the prostate. N. Engl.J. Med., 331: 996-1004, 1994.
3. Gittes. R. F. Carcinoma of the prostate. N. Engl. J. Med.. 324:
236-245. 1991.
4. Scher. H. I., Steineck, G., and Kelly, W. K. Hormone-refractory (D3)
prostate cancer. Refining the concept. Urology. 46: 142-148, 1995.
5. Ellis, W. J.. Vassella, R. L., Buhler, K. R., Bladou, F., True, L. D.,
Bigier. S. A., Curtis, D., and Lange, P. H. Characterization of a novelandrogen-sensitive. prostate-specific antigen-producing prostate carci-
noma xenograft: LuCaP 23. Clin. Cancer Res., 2: 1039-1048, 1996.
6. Kaighn, M. E., Shankar, N., Ohnuki, Y., Lechner, F., and Jones,
L. W. Establishment and characterization of a human prostatic carci-
noma cell line (PC-3). Invest. Urol.. 17: 16-23. 1979.
7. Stone, K. R., Mickey. D. D., Wunderli, H., Mickey, G. H., andPaulson, D. F. Isolation of a human prostate carcinoma cell line
(DU145). Int. J. Cancer, 21: 274-281, 1978.
8. Iizumi, T.. Yazaki, T., Kanoh, S., Kondo, I., and Koiso, K. Estab-
lishment of a new prostatic carcinoma cell line (TSU-PR1). J. Urol..
137: 1304-1306, 1987.
9. van Weerden, W. M., de Ridder. X. M. A., Verdaasdonk,
C. L., Romijn, J. C., van der Kwast, T. H., Schroder, F. H., and vanSteenbrugge, G. J. Development of seven new human prostate tumorxenografts and their histopathological characterization. Am. J. Pathol.,149: 1055-1062, 1996.
10. Horoszewicz, J. S.. Leong, S. S., Kawinski, E., Karr, J. P., Rosenthal,H., Chu, T. M., Mirand, E. A., and Murphy, G. P. LNCaP model of humanprostatic carcinoma. Cancer Res., 43: 1809-1818, 1983.
1 1 . Peehl, D. M. Culture of human prostatic epitheliai cells. in:
R. I.R. A. Freshney (ed), Culture ofEpithelial Cells. pp. 159-180. NewYork: Wiley-Liss, 1992.
12. Freshney. R. I. Culture of animals cells. A manual of basic tech-
niques, Ed. 3. pp. 353-354. New York: Wiley-Liss. 1994.
13. Pershouse, M. A., Stubblefield, E., Hadi, A., and Killary, A. M.Analysis of the functional role of chromosome 10 loss in human glio-
biastomas. Cancer Res., 53: 5043-5050, 1993.
14. Gillies, R. J., Didier, N., and Denton. M. Determination of cell
number in monolayer cultures. Anal. Biochem.. 159: 109-1 13, 1987.
15. Kanamarus, H., and Yoshida, 0. Assessment of in vitro lymphokine
activated killer (LAK) cell activity against renal cancer cell lines and its
suppression by serum factors using crystal violet assay. Urol. Res., 17:
259-264. 1989.
16. Pathak, S. Chromosome banding techniques. J. Reprod. Med., 17:
25-28, 1976.
17. Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: A
Laboratory Manual, Ed. 2, pp. 18.60-18.75. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory. 1989.
18. Zar, J. H. Biostatistical Analysis, Ed. 3, pp. 179-306. EnglewoodCliffs: Prentice Hail, 1996.
19. Dussurget. 0., and Roulland-Dussoix, D. Rapid, sensitive PCR-
based detection of mycoplasmas in simulated samples of animal sera.
Appi. Environ. Microbiol.. 60: 953-959, 1994.
20. Tilley, W. D.. Wilson. C. M.. Marcelli. M.. and McPhaul. M. J.
Androgen receptor gene expression in human prostate carcinoma cell
lines. Cancer Res., 50: 5382-5386, 1990.
21. Trapman, J.. Ris-Staipers, C., van der Korput, J. A.. Kuiper, G. G.,
Faber, P. W., Romijn, J. C.. Mulder, E., and Brinkmann, A. 0. Theandrogen receptor: functional structure and expression in transplanted
human prostate tumors and prostate tumor cell lines. J. Steroid Biochem.
Mol. Biol., 378: 837-842, 1990.
22. Navone, N. M.. Troncoso, P.. Pisters, L. L., and Goodrow, T. L. p53protein accumulation and gene mutation in the progression of human
prostate carcinoma. J. Natl. Cancer Inst.. 85: 1657-1669, 1993.
23. Gleave, M., Hsieh, J. T., Gao. C., von Eschenbach, A. C., and
Chung, L. W. K. Acceleration of human prostate cancer growth in vivo
by factors produced by prostate and bone fibroblasts. Cancer Res.. 51:
3753-3761, 1991.
24. Zhau, H. Y. E., Chang. S-M., Chen, B-Q., Wang. Y., Zhang, H..Kao, C., Sang, Q. A., Pathak, S. J., and Chang. L. W. K. Androgen-
repressed phenotype in human prostate cancer. Proc. Natl. Acad. Sci.
USA, 93: 15152-15157. 1996.
25. Jacobs, S. C. Spread of prostatic cancer to bone. Urology, 21: 337,
1983.
26. Kunimi, K., Bergerheim. U. S. R.. Larsson. I-L., Ekman, P.. and
Collins, V. P. Allelotyping of human prostatic adenocarcinoma. Genom-
ics, 11: 530-539, 1991.
27. Cher, M. L.. Bova. G. S., Moore, D. H.. Small, E. J., Carroll, P. R.,Pin, S. S., Epstein. J. I.. Isaacs, W. B.. and Jensen. R. H. Genetic
alterations in untreated metastases and androgen-independent prostate
cancer detected by comparative genomic hybridization and allelotyping.
Cancer Res., 56: 3091-3102, 1996.
28. Visakorpi, T., Kailioniemi, A. H., and Syvanen. A. C. Genetic
changes in primary and recurrent prostate cancer by comparative
genomic hybridization. Cancer Res., 55: 342-347, 1995.
29. Oakahashi. S. Frequent loss of heterozygosity at 7q3 1 .1 in primary
prostate cancer is associated with tumor aggressiveness and progression.
CancerRes.,55:4li4-4i19, 1995.
30. Jones, E., Zhu, X. L., and Rohr, L. R. Aneusomy of chromosomes7 and 17 detected by FISH in prostate cancer and the effects of selection
i,l vitro. Genes Chromosomes Cancer, 3: 163- 170. 1994.
31. Alcaraz, A., Takahashi, S., Brown, J. A., Herath, J. F., Bergstralh,
E. J.. Larson-Keller, J. J.. Lieber, M. M.. and Jenkins. R. B. Aneuploidyand aneusomy of chromosome 7 detected by fluorescence in situ hy-
bridization are markers of poor prognosis in prostate cancer. Cancer
Res., 54: 3998-4002. 1994.
32. Zenklusen, J. C., Thompson. J. C.. Troncoso, P.. Kagan, J.. and
Conti, C. J. Loss of heterozygosity in human prostate carcinomas: a
possible tumor suppressor gene at 7q31.l. Cancer Res., 54: 6370-6373,
1994.
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1997;3:2493-2500. Clin Cancer Res N M Navone, M Olive, M Ozen, et al. from a single bone metastasis.Establishment of two human prostate cancer cell lines derived
Updated version
http://clincancerres.aacrjournals.org/content/3/12/2493
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/3/12/2493To request permission to re-use all or part of this article, use this link
Research. on August 2, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from