Germline Genetic Variation Modulates Tumor
Progression and Metastasis in a Mouse Model of
Neuroendocrine Prostate Carcinoma
Shashank J. Patel1, Alfredo A. Molinolo2, Silvio Gutkind2, Nigel P. S. Crawford1*
1Metastasis Genetics Section, Cancer Genetics Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland, United States of
America, 2Oral and Pharyngeal Cancer Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland, United States of
Neuroendocrine (NE) differentiation has gained increased attention as a prostate cancer (PC) prognostic marker. The aim of
this study is to determine whether host germline genetic variation influences tumor progression and metastasis in C57BL/6-
Tg(TRAMP)8247Ng/J (TRAMP) mouse model of aggressive NEPC. TRAMP mice were crossed to the eight progenitor strains
of the Collaborative Cross recombinant inbred panel to address this. Tumor growth and metastasis burden were quantified
in heterozygous transgene positive F1 male mice at 30 weeks of age. Compared to wild-type C57BL/6J-Tg(TRAMP)824Ng/J
males, TRAMP x CAST/EiJ, TRAMP x NOD/ShiLtJ and TRAMP x NZO/HlLtJ F1 males displayed significant increases in tumor
growth. Conversely, TRAMP x WSB/EiJ and TRAMP x PWK/PhJ F1 males displayed significant reductions in tumor growth.
Interestingly, despite reduced tumor burden, TRAMP x WSB/EiJ males had an increased nodal metastasis burden. Patterns of
distant pulmonary metastasis tended to follow the same patterns as that of local dissemination in each of the strains. All
tumors and metastases displayed positive staining for NE markers, synaptophysin, and FOXA2. These experiments
conclusively demonstrate that the introduction of germline variation by breeding modulates tumor growth, local metastasis
burden, and distant metastasis frequency in this model of NEPC. These strains will be useful as model systems to facilitate
the identification of germline modifier genes that promote the development of aggressive forms of PC.
Citation: Patel SJ, Molinolo AA, Gutkind S, Crawford NPS (2013) Germline Genetic Variation Modulates Tumor Progression and Metastasis in a Mouse Model of
Neuroendocrine Prostate Carcinoma. PLoS ONE 8(4): e61848. doi:10.1371/journal.pone.0061848
Editor: Irina U. Agoulnik, Florida International University, United States of America
Received November 30, 2012; Accepted March 14, 2013; Published April 19, 2013
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for
any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: This work was supported by the Intramural Research Programs of the National Human Genome Research Institute and National Institute of Dental and
Craniofacial Research, US National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
Prostate cancer (PC) is the second leading cause of cancer
mortality in men . Despite the reduced mortality that has arisen
with an increased availability of tools for early diagnosis and the
success of hormone ablation therapies, advanced and recurrent PC
remains incurable. Conventional assessment of PC prognosis relies
heavily upon the Gleason grading scale to assess the histological
severity of the primary tumor. However, this pathological grading
system suffers from significant interobserver variability, and as
a consequence, there is a need for additional means of assessing
Neuroendocrine differentiation (NED), which is defined as the
emergence of single or small clusters of neuroendocrine cells in
conventional prostatic adenocarcinomas through trans-differenti-
ation, has been gaining increasing attention as a potential
prognostic marker . Adenocarcinomas with significant NED
are typically associated with a poorer prognosis, with NED being
increased in high-grade and high stage tumors, and particularly in
hormone treated and castration resistant disease . Neuroendo-
crine prostate carcinomas (NEPC), which includes carcinoid
tumor and small-cell carcinoma, are heterogeneous tumors
containing mainly NE cells, and are associated with a poor
outcome [2,4]. Although NEPC are rare and comprise less than
2% of all tumors [5,6], a sizable subset of conventional prostate
adenocarcinomas displays NED. One plausible factor influencing
the induction of NED is germline genetic variation. Despite the
fact that recent studies have identified novel genes orchestrating
this aggressive disease [2,7,8], it remains elusive that how host
germline genetic variation influences susceptibility to NED or
The influence of germline variation on tumor progression has
been proven in other cancer types by introducing germline
polymorphism into mouse models of cancer by breeding. This type
of strategy was utilized in the FVB/N-Tg(MMTV-PyVT)634Mul/
J mouse model of mammary tumorigenesis to identify progression
and metastasis quantitative trait loci (QTLs). Subsequent fine
mapping studies allowed for the characterization of Brd4, Sipa1
and Rrp1b as novel breast cancer metastasis susceptibility genes
[9,10,11,12]. The clinical significance of these was established by
demonstrating that germline polymorphisms within these genes
are associated with poor prognosis in multiple human breast
cancer cohorts [13,14,15].
The TRAMP transgenic mouse model was utilized in this study
to investigate the role of host germline polymorphism in the
development of aggressive prostate carcinoma. This model utilizes
PLOS ONE | www.plosone.org1April 2013 | Volume 8 | Issue 4 | e61848
the androgen-responsive minimal probasin promoter (PB) to
initiate SV40 virus T antigen (TAg)-mediated tumorigenesis in
the prostatic epithelium. It is a particularly useful model for
studying progression and metastasis, since by 30 weeks of age mice
typically display a high incidence of multi-organ metastases, which
is not a characteristic of any other prostate cancer mouse model
[16,17]. Also, studies have shown that NE phenotype tumors in
TRAMP mice share certain key molecular features with NED and
NEPC in humans, such as tumor growth regulation by high
expression of AURKA, MYCN, FOXA2 and SIAH2 [7,8,18].
In this study, germline polymorphism was introduced by
breeding to determine the extent to which host genetic variation
influences tumor progression and metastasis in this model.
Specifically, TRAMP mice were crossed to the eight progenitor
strains of the Collaborative Cross (CC) recombinant inbred panel,
which are a genetically diverse mapping panel that captures nearly
90% of the known variation present in laboratory mice [19,20]
This study conclusively demonstrates that the introduction of
germline variation by breeding significantly modulates tumor
growth, local metastasis burden, and distant metastasis frequency
in TRAMP mice.
Materials and Methods
Animal Husbandry and Genotyping
C57BL/6J-Tg(TRAMP)824Ng/J (TRAMP-B6) mice and eight
CC progenitor strains, C57BL/6J (B6), A/J (A), 129S1/SvImJ
(129S), NOD/LtJ (NOD), NZO/HlLtJ (NZO), CAST/EiJ
(CAST), PWK/PhJ (PWK) and WSB/EiJ (WSB), were obtained
from Jackson Laboratories (Bar Harbor, ME). TRAMP-B6
females were crossed with eight CC progenitor strains to generate
eight different F1 strains of mice (TRAMP-B6, TRAMP-A,
TRAMP-129S, TRAMP-NOD, TRAMP-NZO, TRAMP-CAST,
TRAMP-PWK and TRAMP-WSB, respectively) hemizygous for
PB-TAg transgene. TRAMP-B6 females were bred with non-
transgenic B6 males for colony maintenance. This study was
carried out in strict accordance with the recommendations in the
Guide for the Care and Use of Laboratory Animals of the National
Institutes of Health. The protocol was approved by the National
Human Genome Research Institute Animal Care and Use
Committee (protocol number: G-09-2). All necropsies were
performed under sodium pentobarbital anesthesia followed by
cervical dislocation, and all efforts were made to minimize
suffering. The mouse tail genomic DNA was extracted from F1
progeny with HotSHOT method  for genotyping analysis. To
identify the hemizygous PB-TAg transgene positive F1 (TRAMP
F1) mice, PCR screening was performed as described .
TRAMP F1 male mice were sacrificed by pentobarbital
overdose at 30 weeks of age or humane endpoint, whichever
were achieved first. Humane experimental endpoints for this study
were rapid weight loss, hunched posture, labored breathing,
trauma, impaired mobility, dysuria, or difficulty in obtaining food
or water. Prostatic tumor, seminal vesicles, lungs, liver, and lymph
nodes were harvested from TRAMP F1 males. Prostatic tumor
and seminal vesicles were weighed to quantify tumor burden.
Tumor and seminal vesicle weights were normalized to age of the
mouse at the time of euthanasia by multiplying the actual weight
by the inverse ratio of the age in days at the time of death and the
pre-designated experimental endpoint (i.e. 210 days). Visible
enlarged lymph nodes in para-aortic region were weighed to
quantify metastatic lymph node burden. These were again
normalized to age of the mouse at the time of euthanasia. Lungs
were collected to determine isolated tumor cell infiltrates in lung
parenchyma and microscopic metastatic lesions. Other organs
displaying macroscopic metastatic lesions through gross observa-
tion were also collected for histology. These collected tissues were
fixed in buffered formalin (10% w/v phosphate buffered
formaldehyde, Fisher Scientific) or Z-fix (zinc buffered formalde-
hyde, Anatech Ltd.) overnight and then transferred to 70%
ethanol. Fixed tissues were embedded in paraffin, sectioned to
a thickness of 4 mm and stained with hematoxylin and eosin
(H&E). Histology slides were scanned with Scanscope Digital
microscope (Aperio, Vista, CA).
Tissues were collected and sectioned as described above for
immunohistochemical (IHC) analyses. Histopathological analyses
of these tissues were undertaken by Dr. A. Molinolo. Tissues were
stained with the following primary antibodies: anti-synaptophysin,
1:200 (Life Technologies, Carlsbad, CA); anti-cytokeratin 8, 1:200
(CK8; ab59400, Abcam, Cambridge, MA); and anti-SV40 Large
T-antigen (TAg), 1:400, (BD Pharmingen, San Diego, CA).
Paraffin blocks were sent to Center of Comparative Medicine,
University of California, Davis, CA to perform IHC staining with
anti-Foxa1 (C-20), 1:1000 and anti-Foxa2 (P-19), 1:1000, BD
Transduction Laboratories, San Diego, CA.
Time Course Experiments
TRAMP F1 males of B6, WSB, PWK and NOD backgrounds
were allocated to the groups and separated throughout different
time points. For each of these strains of TRAMP F1 mice, three
mice were assigned for each time point. These mice were sacrificed
at 4, 8, 12, 16, 20 and 25 weeks; prostate and seminal vesicles were
harvested for histopathological analysis. Prostatic intraepithelial
neoplasia (PIN) lesion in TRAMP mice has been termed as
atypical hyperplasia of TAg in previous literature . However,
lesions of prostate epithelium were categorized into three
categories, hyperplasia, PIN and carcinoma, following the
conventional classification . Prostates from these F1 males were
collected at 8 weeks of age and subjected to IHC staining with
anti-TAg to determine the strain specific differences in TAg
expression in prostatic epithelium.
Quantitative and Statistical Analysis
MedCalc (Mariakerke, Belgium) was used to analyze phenotype
data. Statistical significance levels for observed differences were
calculated with the Mann-Whitney U test. Kaplan-Meyer survival
curves were plotted for survival analyses and hazard ratios
calculated using the Cox proportional hazards model. Imagescope
software (Aperio, Vista, CA) was used to quantify the nuclear
expression of TAg in prostatic epithelium. The significance levels
for the strain specific differences in TAg expression were
determined with Fisher’s exact test. Correlation coefficient and
level of its significance were equated with Spearman’s rank
analysis. P,0.05 was considered significant.
Germline Genetic Variation Modulates Tumor Associated
Mortality and Tumor Burden in TRAMP Mice
The aim was to assess tumor growth and metastasis in TRAMP
F1 males at 30 weeks of age. However this goal was frequently not
achievable owing to factors relating to the overall health of the
mouse necessitating euthanasia for humane reasons listed above.
Tumor-associated mortality was therefore followed for TRAMP
F1 mice, and significant differences were observed in a number of
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org2April 2013 | Volume 8 | Issue 4 | e61848
the F1 strains (Fig. 1a). Compared to wildtype C57BL/6J-
Tg(TRAMP)824Ng/J males, tumor associated mortality was
significantly higher in TRAMP-NOD (hazard ratio [HR]=2.26
[1.76–2.92]; P,0.001), TRAMP-CAST (HR=3.03 [1.77–5.16];
P,0.001) and TRAMP-NZO (HR=1.18 [1.02–1.35]; P=0.022).
TRAMP-A, TRAMP-129S, TRAMP-WSB and TRAMP-PWK
males did not show any significant difference in their survival times
in comparison to wildtype C57BL/6J-Tg(TRAMP)824Ng/J
To account for differences in tumor associated mortality, tumor
weights were adjusted in all strains to account for age of the mouse
at the time of euthanasia/death. Compared to wildtype C57BL/
6J-Tg(TRAMP)824Ng/J males (age adjusted mean 6 SEM of
tumor burden=1.5460.65 g,
(n=42) and TRAMP-NZO (n=57) males displayed a significant
increase in tumor growth (tumor burden=9.8661.16 g and
5.6660.97 g, respectively; P,0.001; Fig. 1b). TRAMP-CAST
animals (n=24) also displayed a significant increase in tumor
growth, albeit at a somewhat lesser level of significance (tumor
burden=4.5361.09 g; P=0.050; Fig. 1b). Conversely, both
TRAMP-WSB (n=68) and TRAMP-PWK (n=37) males dis-
played a significant reduction in tumor growth (tumor bur-
den=0.7960.31 g [P=0.001] and 0.0260.01 g [P,0.001], re-
spectively; Fig. 1b). Indeed, only 29/68 TRAMP-WSB mice and
1/37 TRAMP-PWK mice developed macroscopic tumors, weigh-
ing $0.2 g. TRAMP-129S (n=35) and TRAMP-A (n=68) did
not show significant difference in tumor growth compared to wild-
type TRAMP mice. These data conclusively demonstrate that the
introduction of germline variation by breeding significantly
modulated tumor burden in the TRAMP mouse.
As was the case in the previous studies [24,25], seminal vesicle
tumors were observed in TRAMP F1 males. Seminal vesicle
weight was therefore recorded as an indicator of this type of tumor
growth. Seminal vesicles from TRAMP-CAST (n=15), TRAMP-
PWK (n=37), TRAMP-WSB (n=68) and TRAMP-A (n=65)
showed significantly less tumor growth (age adjusted seminal
vesicle burden=0.2060.03 g, 0.3160.02 g, 0.5760.04 g and
1.0860.08 g respectively;
C57BL/6J-Tg(TRAMP)824Ng/J males (seminal vesicle bur-
den=1.3660.10 g, n=39; Fig. 1c). However, comparatively,
TRAMP-NZO displayed significantly higher tumor growth
(seminal vesicle burden=2.1660.04 g, n=52; P=0.009; Fig. 1c).
n=35), both TRAMP-NOD
Germline Variation does not Affect TAg Expression in
The dramatic reduction in prostate tumor burden seen in the
TRAMP-PWK and TRAMP-WSB strains compared to other F1
mice could be due to suppression of TAg transgene in the prostatic
epithelium. To eliminate this possibility, IHC analysis for TAg
transgene expression was performed using prostates from 8 week
old wildtype C57BL/6J-Tg(TRAMP)824Ng/J, TRAMP-PWK,
TRAMP-WSB and TRAMP-NOD mice. The percentage of
prostatic acinar cells expressing TAg from anterior, lateral, ventral
and dorsal lobes was quantified. The total amount of cells
expressing TAg in the prostate did not vary significantly in
TRAMP-PWK, TRAMP-WSB and TRAMP-NOD mice com-
pared to wildtype C57BL/6J-Tg(TRAMP)824Ng/J (P.0.1; Fig.
S1). These data indicate that TAg expression does not differ
among strains and that the observed phenotypic differences are
due to germline genetic variation.
Figure 1. Genetic background influences prostate and seminal
vesicle tumor burdens, and tumor associated mortality in
TRAMP mice. (a) Kaplan-Meier survival curve for TRAMP F1 strains
with a significantly increased tumor-associated mortality compared to
wildtype C57BL/6J-Tg(TRAMP)824Ng/J (B6) mice. (b) Average age
adjusted prostate tumor burden in TRAMP F1 strains. (c) Average age
adjusted seminal vesicle tumor burden in TRAMP F1 strains. Dotted line
at 0.6 g on y-axis represents the maximum weight of seminal vesicles
recorded in transgene negative TRAMP F1 mice at 30 weeks of age. Bar
graphs represent average weight of prostate tumor or seminal vesicles
6 SEM. *P,0.050 and **P,0.001. Values in parentheses below x-axis
represent the number of animals evaluated in each group.
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org3April 2013 | Volume 8 | Issue 4 | e61848
Strain Specific Differences in Prostate Carcinoma
Prostate carcinoma in the TRAMP model exhibits multi-organ
metastasis, primarily to lymph nodes and also to lungs, liver,
kidney and bones . In order to examine the effect of strain-
specific modifier loci on tumor cell capacity to disseminate to
distant organs via lymphatic and blood vessels, the incidence of
metastasis were tracked in pelvic and renal lymph nodes, lungs,
liver and kidney in TRAMP F1 mice. Strain-specific variation was
observed in the incidence of metastases in multiple organs (Fig. 2a
and 2b). IHC for SV40-TAg and androgen receptor were
performed to better characterize these lesions. Regardless of strain
background, all metastases stained positive for SV40-TAg and
negative for AR (Fig. S2).
A high incidence of metastasis was observed in some F1 strains.
For example, 40/42 (95%) of TRAMP-NOD males developed
metastases in local lymph nodes, 12/42 (29%) developed liver
metastases, and 22/29 (75%) developed pulmonary metastases.
Conversely, a low incidence of metastasis was observed in other
strains. For example, only 14/68 (21%) TRAMP-WSB males
developed nodal metastases, 8/41 (20%) pulmonary metastases
and 3/41 (4%) hepatic metastases.
Several of the F1 strains exhibited striking differences in age
adjusted local nodal metastasis burden (Fig. 2c) compared to the
actual frequency of nodal metastasis (Fig. 2a). The wildtype
C57BL/6J-Tg(TRAMP)824Ng/J strain exhibited the lowest age
adjusted local metastasis burden with an average local nodal
weight of 0.1260.02 g. Five F1 strains had a significantly
increased age adjusted local nodal metastasis compared to this:
TRAMP-NZO (met burden=0.3160.12; P=0.012), TRAMP-A
(met burden=0.5360.20; P=0.049), TRAMP-CAST (met bur-
den=1.5360.45; P,0.0001) and TRAMP-WSB (met bur-
Characterization of the NE Phenotype in TRAMP Tumors
NE phenotype prostate tumor cells in TRAMP mice display
similar morphological and molecular characteristics to human
NED and NEPC cells, including the expression of IHC markers,
synaptophysin, chromogranin A and FOXA2, and loss of
expression of cytokeratin 8 (CK8) [2,4,7,8,26]. IHC for the NE
marker synaptophysin was performed in tumors and metastatic
lesions. Regardless of strain background, all macroscopic tumors
($0.2 g) stained positive for synaptophysin (Fig. 3). Tumor cells
infiltrating lung, liver and lymph node tissues also stained positive
for synaptophysin (Fig. 4). These observations establish that tumor
burden and nodal metastatic burden, used as quantitative traits in
this study, represent phenotypes of NEPC.
To further characterize the NE phenotype, IHC was performed
for the transcription factors FOXA1 and FOXA2, and the luminal
epithelial marker CK8. As expected, CK8 staining was absent in
all TRAMP F1 tumors (data not shown). Staining for FOXA2
mirrored that of synaptophysin, and was observed in all tumors.
However, such consistency was not observed when comparing
FOXA1 and FOXA2 staining patterns with some foci staining
strongly for both FOXA1 and FOXA2, while others stained
strongly for only either one of these transcription markers (Fig. 3 g–
l). Such differences in FOXA1 and FOXA2 staining were not
strain-dependent and most likely represent heterogeneity of the
overall process of tumorigenesis.
Germline Factors Mediate NE Carcinoma Initiation
To investigate differences in tumor initiation and progression,
serial histological analysis of prostate and seminal vesicles was
performed in some of the strains with the most divergent
Tg(TRAMP)824Ng/J strain (TRAMP-NOD, TRAMP-WSB and
TRAMP-PWK). Hyperplastic lesions were evident in dorsolateral
and anterior prostates in all strains as early as 4 weeks. By 8 weeks
Figure 2. Effect of paternal genotype on metastatic efficiency of carcinoma cells in TRAMP mice. (a) Incidence of macroscopic
metastases in lymph nodes, kidney and liver modulated by paternal genotype. Data not shown for TRAMP-PWK since macroscopic metastases were
not observed in this strain. (b) Incidence of lung metastases modulated by paternal genotype. These incidence counts include the presence of
isolated NE tumor cells, microscopic or macroscopic metastases in lung parenchyma. (c) Average age adjusted lymph node metastasis burden
affected by paternal genotype. The total nodal metastasis burden for less enlarged lymph nodes (,0.1 g) was rounded to 0.1 g. Bar graphs represent
average nodal burden 6 SEM. *P,0.050 and **P,0.001 for comparisons vs. wildtype C57BL/6J-Tg(TRAMP)824Ng/J (B6). Data not shown for TRAMP-
PWK since nodal metastases were not observed in this strain. Values in parentheses below x-axis represent the number of animals evaluated in each
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org4 April 2013 | Volume 8 | Issue 4 | e61848
of age, these hyperplastic lesions invariably progressed to PIN. The
severity and number of hyperplastic foci and PIN lesions increased
with age in these F1 mice, and no strain differences were noted.
Seminal vesicle hyperplasia was benign in nature and
progressed to become papillary adenomas or phyllodes-like tumor
consisting of stromal and epithelial proliferation protruding into
the glandular lumen. NE carcinomas consist of solid masses of
tumor cells with atypical and pleomorphic cells without nuclear
polarity and scant cytoplasm. The histological features of NE
carcinoma and seminal vesicle hyperplasia did not differ between
TRAMP-WSB and TRAMP-PWK strains. Similarly, histological
Tg(TRAMP)824Ng/J mice in all other F1 strains studied (data
not shown). NE carcinomas were found to be arising from ventral
and/or dorsolateral lobes of prostate in TRAMP mice. These
strains exhibited differences in the incidence of seminal vesicle
hyperplasia and NE carcinomas (Table S1).
The earliest time point when the NE lesion was evident in
TRAMP-NOD mice was 8 weeks. The incidence of NE
carcinoma increased with age, and by 16 weeks, 100% of
TRAMP-NOD mice (6/6) displayed NE carcinoma lesions or
Figure 3. Representative photomicrographs of strain-specific differences in the expression of NE markers in prostate tumors from
TRAMP F1 mice. Panels (a), (b) and (c) show H&E staining of the tumors; Panels (d), (e) and (f) show IHC staining of tumors with NE marker
synaptophysin; Panels (g), (h) and (i) show IHC staining with FOXA1; Panels (j), (k) and (l) show IHC staining with FOXA2. Arrows point to normal
prostate luminal cells, showing positive staining for FOXA1 and no staining for the NE markers, synaptophysin and FOXA2. Bar denotes 100 mm.
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org5 April 2013 | Volume 8 | Issue 4 | e61848
tumor. Conversely, only 1 out of 15 wildtype C57BL/6J-
Tg(TRAMP)824Ng/J males developed NE tumor at 16 weeks of
age. In our time-course study, all TRAMP-WSB mice developed
seminal vesicle hyperplasia by 25 weeks of age (6/6), but none had
NE carcinoma. Given that no tumors were evident in TRAMP-
PWK mice by 30 weeks of age, a set of these mice were aged to 50
weeks (n=4). Interestingly, only 1 of these TRAMP-PWK mice
developed NE tumors and died at 38 weeks of age. Another
TRAMP-PWK mouse developed only hyperplasia, and 2 of these
4 TRAMP-PWK developed severe hyperplasia and phyllodes-like
proliferations at 50 weeks. These data emphasize previous
observations that NE lesions arise independently of the PIN
lesions in the TRAMP model . More importantly, the
observed strain-specific differences demonstrate that paternal
genotypes substantially affect the initiation and kinetics of NE
tumorigenesis in TRAMP F1 mice.
Although NE carcinoma represents only 1–2% of prostate
malignancies, this form of PC is highly aggressive and has a poor
survival rate [3,27]. Additionally, NED is present in almost .90%
of high grade adenocarcinomas of prostate . More important-
ly, androgen depletion therapy strongly stimulates NED [29,30].
The characteristics of these types of tumor, including aggressive
tumor growth, multi-organ metastases and relapse of disease after
castration or androgen depletion are well recapitulated by
TRAMP mouse model [24,31,32]. The overall aim of this work
is to identify germline susceptibility genes that are associated with
Figure 4. Representative photomicrographs of metastatic lesions displaying NE phenotype in organs harvested from TRAMP F1
mice. Panels (a), (c) and (e) show H&E staining of the metastatic lesions in the lymph node, lungs and liver, respectively. Panels (b), (d) and (f) show
strong positive IHC staining of metastatic PC cells with synaptophysin. Bar denotes 50 mm.
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org6April 2013 | Volume 8 | Issue 4 | e61848
an increased risk of NED development. The potential clinical
significance of this is that it may prove possible to type
polymorphic germline modifier genes at the time of PC diagnosis
to more accurately identify men at risk of progression, and thus
initiate more aggressive therapies in these metastasis-prone
Germline predisposition to diseases displaying a complex, non-
Mendelian inheritance pattern is controlled by polymorphisms
within multiple genes, each of which likely have a small effect in
isolation . The identification of modifier genes of tumor
progression and metastasis has proven particularly difficult using
genome wide and familial association studies. An alternate
approach has been to map disease modifier loci using mouse
models of human disease [34,35]. Selection of the appropriate
mouse strain(s) for this type of study is of critical importance to
maximize genetic power and increase the likelihood of identifying
genes relevant to human disease [36,37]. This study was carried
out to determine whether germline modifiers modulate the
development of aggressive NEPC in TRAMP model. Our data
have proven that this is indeed the case, and the information
gained from these analyses will be used to direct future genetic
In this study, polymorphisms carried by NOD males were
shown to increase tumor growth and dissemination. Conversely,
progression was significantly suppressed by modifiers inherited
from PWK males, with NE tumorigenesis being almost completely
suppressed in this strain. A particularly interesting observation was
that TRAMP-CAST and TRAMP-WSB strains displayed a par-
ticularly high propensity for metastasis, where the nodal metastasis
burden frequently exceeded the primary tumor weight (Fig. 2C).
Indeed, the disease course in the TRAMP-WSB strain was
particularly unusual when one considers that only 29/68 de-
veloped macroscopic tumorigenesis and that the overall age
adjusted prostate and seminal vesicle tumor burden in this strain
Tg(TRAMP)824Ng/J mice (see Fig. 1a and 1b, respectively).
The dissociation of tumor growth and metastasis in this strain
underscores the differing mechanisms involved in these two
processes and makes TRAMP-WSB a particularly important
strain for future genetic mapping studies.
Additionally, our time-course experiment data showed that
prostate and seminal vesicle tumor initiation are controlled by
inherited polymorphic genes in these strains. This time-course
study demonstrated that there is dissociation among the experi-
mental groups between the development of neuroendocrine
prostate tumors, which are high in the TRAMP-NOD strain but
low in the other groups and seminal vesicle hyperplasia, which is
high in the wildtype C57BL/6J-Tg(TRAMP)824Ng/J and
TRAMP-WSB strains but low in TRAMP-NOD and TRAMP-
PWK. Overall, these data imply that these differing components of
tumorigenesis in TRAMP mice are influenced by different
germline polymorphisms. Interestingly, another study reported
that NE carcinoma occurs in 20% of TRAMP mice on B6
background and 100% of mice on FVB background, and in ,50%
of TRAMP-B6 x FVB F1 mice. This suggests that the interaction
of modifiers present in B6 mice decreased NEPC incidence
compared to TRAMP-FVB mice . However, the present study
demonstrates that modifiers present in NOD mice can override
the suppressive effect of B6 strain modifiers to promote NEPC
incidence. Collectively, the presented data demonstrate that
polymorphic genes derived from paternal strains likely influence
tumor progression at multiple stages of disease development.
We do however acknowledge that the TRAMP model suffers
a number of limitations. Specifically, the histological features of
TRAMP tumors are much more reminiscent of human NEPC,
which is relatively rare in humans, rather than the far more
common adenocarcinoma with NED. The precise origins of the
neoplastic process in the TRAMP mouse were clearly described in
the seminal work of Chiaverotti et. al. . This study determined
that two distinct neoplastic cellular lineages evolve within TRAMP
tumors. The first lineage, which was termed ‘abnormal hyperpla-
sia of TAg’, was determined to follow a benign course and does
not progress beyond pre-malignant PIN lesions. The second
lineage, which accounts for the aggressive malignancies observed
in TRAMP mice, arises spontaneously and has a strong neuro-
endocrine phenotype. Our work confirms the findings of this
important study. Specifically, although abnormal hyperplasia of
TAg and PIN was observed in all F1 animals, the incidence or
kinetics of these did not differ between strains. As observed by
Chiaverotti et. al. , strain differences appear to exert the
majority of their influence upon the NE component of TRAMP
tumorigenesis. Another limitation of this model is the relative
infrequency of skeletal metastases . Initial analyses demon-
strated a slight increase in the frequency of metastases to long
bones in more F1 mice displaying a more aggressive phenotype
(e.g. TRAMP-NOD). However, the overall incidence of bone
metastases remained low and it is unlikely that formal quantifi-
cation would yield any data of value in the cohort sizes described
Regardless of the limitations of the TRAMP model, we argue
that the modifier genes identified in future studies performed on
the basis of the experiments performed here are likely to be
applicable not only to NEPC but also more generally to the
pathogenesis of aggressive human PC. Prostate adenocarcinomas
and NE carcinomas derived from both human and mice all possess
the capability to metastasize to lymph nodes, lungs and liver. This
indicates that these two distinct tumor cell types may well have
acquired similar molecular characteristics under the influence of
common allelic determinants to promote a more metastatic
phenotype. This is supported by the fact that metastatic disease
is enhanced in recurrent PC after androgen depletion therapy,
which is similar to the highly metastatic behavior displayed by
NEPC cells lacking androgen receptors . In addition, both
androgen-independent PC and NEPC cells over-express tran-
scription factors such as FOXA1, indicating the involvement of
common factors in the transcriptional regulation in these cells
[39,40,41]. These observations support the notion that common
pathways may well be involved in the modulation of adenocar-
cinoma-associated NED and NE prostate cancer progression.
In conclusion, this study establishes that germline polymorph-
isms are significant modulators of NE tumor initiation and growth
kinetics in the TRAMP model. TAg-mediated tumorigenesis and
metastasis have been characterized in the progenitor strains of the
CC recombinant inbred panel, which will facilitate future modifier
mapping studies. Identification of the candidate genes and
characterization of genetic polymorphisms associated with this
aggressive form of prostate disease will facilitate an improved
understanding of this aspect of PC pathogenesis and potentially
facilitate the identification of individuals at increased risk of NED.
Additionally, selective use of these TRAMP F1 strains in
preclinical chemotherapeutic trials for specifically targeting
metastasis or tumor growth of NEPC may prove to be a valuable
of TAg transgene in prostate epithelia of TRAMP F1
Strain specific differences in the expression
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org7 April 2013 | Volume 8 | Issue 4 | e61848
mice. Transgene was found to be expressing predominantly in
ventral and dorsolateral prostates (representative photomicro-
graphs (a), (b) and (c)) of TRAMP F1 mice. (c). Percentage of
cells showing positive IHC staining with anti-TAg across different
strains of F1 mice from ventral, dorsal, lateral and anterior lobes of
prostate collected from 8 weeks old TRAMP F1 mice. Bar graph
represents average percent of cells stained positive 6 SD for
TRAMP F1 strains. Bar denotes 50 mm.
metastatic lesions. IHC analysis for androgen receptor (AR)
and SV40-TAg was performed to better characterize metastatic
lesions in TRAMP F1 strains. With regards to AR, immunostain-
ing was positive in a high proportion of the nuclei in a paraffin
embedded section of normal prostate (a). No staining was seen in
either lymph nodes (c) or lung metastases (e), regardless of strain
background. For the SV40-TAg, the normal prostate control is not
immunoreactive (b), but strong positivity was seen in the nuclei of
both lymph nodes (d) and lung metastases (f). The arrows point to
the metastatic area embedded in the lung parenchyma. Bar
denotes 50 mm.
Representative IHC analysis of TRAMP F1
NE carcinoma. Histopathological analysis of prostates from those
F1 strains displaying greatest phenotypic variability compared to
TRAMP-B6 was performed at various different ages. The number
of mice in this time course experiment with evidence of various
neoplastic lesions for each strain is displayed in the table.
Strain related variation in pathological progression of
We thank Dr. Robert Cardiff for his valuable discussions and counseling.
Additionally, we would like to thank Drs. Kathy Kelly, Kent Hunter and
Minnkyong Lee for their critical appraisal of this manuscript.
Conceived and designed the experiments: SJP AAM SG NPSC. Performed
the experiments: SJP AAM. Analyzed the data: SJP AAM. Contributed
reagents/materials/analysis tools: SJP AAM. Wrote the paper: SJP AAM
1. Gloeckler Ries LA, Reichman ME, Lewis DR, Hankey BF, Edwards BK (2003)
Cancer survival and incidence from the Surveillance, Epidemiology, and End
Results (SEER) program. Oncologist 8: 541–552.
2. Sun Y, Niu J, Huang J (2009) Neuroendocrine differentiation in prostate cancer.
Am J Transl Res 1: 148–162.
3. Cindolo L, Cantile M, Vacherot F, Terry S, de la Taille A (2007)
Neuroendocrine differentiation in prostate cancer: from lab to bedside. Urol
Int 79: 287–296.
4. Shappell SB, Thomas GV, Roberts RL, Herbert R, Ittmann MM, et al. (2004)
Prostate pathology of genetically engineered mice: definitions and classification.
The consensus report from the Bar Harbor meeting of the Mouse Models of
Human Cancer Consortium Prostate Pathology Committee. Cancer Res 64:
5. Fixemer T, Remberger K, Bonkhoff H (2002) Apoptosis resistance of
neuroendocrine phenotypes in prostatic adenocarcinoma. Prostate 53: 118–123.
6. Grobholz R, Griebe M, Sauer CG, Michel MS, Trojan L, et al. (2005) Influence
of neuroendocrine tumor cells on proliferation in prostatic carcinoma. Hum
Pathol 36: 562–570.
7. Qi J, Nakayama K, Cardiff RD, Borowsky AD, Kaul K, et al. (2010) Siah2-
dependent concerted activity of HIF and FoxA2 regulates formation of
neuroendocrine phenotype and neuroendocrine prostate tumors. Cancer Cell
8. Beltran H, Rickman DS, Park K, Chae SS, Sboner A, et al. (2011) Molecular
Characterization of Neuroendocrine Prostate Cancer and Identification of New
Drug Targets. Cancer Discov 1: 487–495.
9. Hunter KW, Broman KW, Voyer TL, Lukes L, Cozma D, et al. (2001)
Predisposition to efficient mammary tumor metastatic progression is linked to
the breast cancer metastasis suppressor gene Brms1. Cancer Res 61: 8866–8872.
10. Park YG, Zhao X, Lesueur F, Lowy DR, Lancaster M, et al. (2005) Sipa1 is
a candidate for underlying the metastasis efficiency modifier locus Mtes1. Nat
Genet 37: 1055–1062.
11. Crawford NP, Qian X, Ziogas A, Papageorge AG, Boersma BJ, et al. (2007)
Rrp1b, a new candidate susceptibility gene for breast cancer progression and
metastasis. PLoS Genet 3: e214.
12. Lifsted T, Le Voyer T, Williams M, Muller W, Klein-Szanto A, et al. (1998)
Identification of inbred mouse strains harboring genetic modifiers of mammary
tumor age of onset and metastatic progression. Int J Cancer 77: 640–644.
13. Crawford NP, Ziogas A, Peel DJ, Hess J, Anton-Culver H, et al. (2006) Germline
polymorphisms in SIPA1 are associated with metastasis and other indicators of
poor prognosis in breast cancer. Breast Cancer Res 8: R16.
14. Crawford NP, Alsarraj J, Lukes L, Walker RC, Officewala JS, et al. (2008)
Bromodomain 4 activation predicts breast cancer survival. Proc Natl Acad
Sci U S A 105: 6380–6385.
15. Crawford NP, Walker RC, Lukes L, Officewala JS, Williams RW, et al. (2008)
The Diasporin Pathway: a tumor progression-related transcriptional network
that predicts breast cancer survival. Clin Exp Metastasis 25: 357–369.
16. Gingrich JR, Barrios RJ, Morton RA, Boyce BF, DeMayo FJ, et al. (1996)
Metastatic prostate cancer in a transgenic mouse. Cancer Res 56: 4096–4102.
17. Wang F (2011) Modeling human prostate cancer in genetically engineered mice.
Prog Mol Biol Transl Sci 100: 1–49.
18. Haram KM, Peltier HJ, Lu B, Bhasin M, Otu HH, et al. (2008) Gene expression
profile of mouse prostate tumors reveals dysregulations in major biological
processes and identifies potential murine targets for preclinical development of
human prostate cancer therapy. Prostate 68: 1517–1530.
19. Roberts A, Pardo-Manuel de Villena F, Wang W, McMillan L, Threadgill DW
(2007) The polymorphism architecture of mouse genetic resources elucidated
using genome-wide resequencing data: implications for QTL discovery and
systems genetics. Mamm Genome 18: 473–481.
20. Churchill GA, Airey DC, Allayee H, Angel JM, Attie AD, et al. (2004) The
Collaborative Cross, a community resource for the genetic analysis of complex
traits. Nat Genet 36: 1133–1137.
21. Truett GE, Heeger P, Mynatt RL, Truett AA, Walker JA, et al. (2000)
Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide
and tris (HotSHOT). Biotechniques 29: 52, 54.
22. Hurwitz AA, Foster BA, Allison JP, Greenberg NM, Kwon ED (2001) The
TRAMP mouse as a model for prostate cancer. Curr Protoc Immunol Chapter
20: Unit 20 25.
23. Chiaverotti T, Couto SS, Donjacour A, Mao J-H, Nagase H, et al. (2008)
Dissociation of Epithelial and Neuroendocrine Carcinoma Lineages in the
Transgenic Adenocarcinoma of Mouse Prostate Model of Prostate Cancer. The
American Journal of Pathology 172: 236–246.
24. Gingrich JR, Barrios RJ, Foster BA, Greenberg NM (1999) Pathologic
progression of autochthonous prostate cancer in the TRAMP model. Prostate
Cancer Prostatic Dis 2: 70–75.
25. Tani Y, Suttie A, Flake GP, Nyska A, Maronpot RR (2005) Epithelial-stromal
tumor of the seminal vesicles in the transgenic adenocarcinoma mouse prostate
model. Vet Pathol 42: 306–314.
26. Kaplan-Lefko PJ, Chen TM, Ittmann MM, Barrios RJ, Ayala GE, et al. (2003)
Pathobiology of autochthonous prostate cancer in a pre-clinical transgenic
mouse model. Prostate 55: 219–237.
27. Sella A, Konichezky M, Flex D, Sulkes A, Baniel J (2000) Low PSA metastatic
androgen- independent prostate cancer. Eur Urol 38: 250–254.
28. Vashchenko N, Abrahamsson PA (2005) Neuroendocrine differentiation in
prostate cancer: implications for new treatment modalities. Eur Urol 47: 147–
29. Sciarra A, Monti S, Gentile V, Mariotti G, Cardi A, et al. (2003) Variation in
chromogranin A serum levels during intermittent versus continuous androgen
deprivation therapy for prostate adenocarcinoma. Prostate 55: 168–179.
30. Berruti A, Mosca A, Tucci M, Terrone C, Torta M, et al. (2005) Independent
prognostic role of circulating chromogranin A in prostate cancer patients with
hormone-refractory disease. Endocr Relat Cancer 12: 109–117.
31. Hurwitz AA, Foster BA, Allison JP, Greenberg NM, Kwon ED (2001) The
TRAMP Mouse as a Model for Prostate Cancer. Current Protocols in
Immunology. Hoboken, NJ: John Wiley & Sons, Inc.
32. Gingrich JR, Barrios RJ, Kattan MW, Nahm HS, Finegold MJ, et al. (1997)
Androgen-independent Prostate Cancer Progression in the TRAMP Model.
Cancer Research 57: 4687–4691.
33. Balmain A, Gray J, Ponder B (2003) The genetics and genomics of cancer. Nat
Genet 33 Suppl: 238–244.
34. Talbot CJ, Nicod A, Cherny SS, Fulker DW, Collins AC, et al. (1999) High-
resolution mapping of quantitative trait loci in outbred mice. Nat Genet 21:
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org8 April 2013 | Volume 8 | Issue 4 | e61848
35. Hunter KW, Crawford NP (2008) The future of mouse QTL mapping to Download full-text
diagnose disease in mice in the age of whole-genome association studies. Annu
Rev Genet 42: 131–141.
36. Navarro SJ, Trinh T, Lucas CA, Ross AJ, Waymire KG, et al. (2012) The
C57BL/6J Mouse Strain Background Modifies the Effect of a Mutation in
Bcl2l2. G3 (Bethesda) 2: 99–102.
37. Hunter KW (2012) Mouse models of cancer: does the strain matter? Nat Rev
Cancer 12: 144–149.
38. Bostwick DG, Qian J, Pacelli A, Zincke H, Blute M, et al. (2002)
Neuroendocrine expression in node positive prostate cancer: correlation with
systemic progression and patient survival. J Urol 168: 1204–1211.
39. Mirosevich J, Gao N, Gupta A, Shappell SB, Jove R, et al. (2006) Expression
and role of Foxa proteins in prostate cancer. Prostate 66: 1013–1028.
40. Wang Q, Li W, Zhang Y, Yuan X, Xu K, et al. (2009) Androgen receptor
regulates a distinct transcription program in androgen-independent prostate
cancer. Cell 138: 245–256.
41. Gerhardt J, Montani M, Wild P, Beer M, Huber F, et al. (2012) FOXA1
promotes tumor progression in prostate cancer and represents a novel hallmark
of castration-resistant prostate cancer. Am J Pathol 180: 848–861.
Genetic Variation Modulates Metastasis
PLOS ONE | www.plosone.org9 April 2013 | Volume 8 | Issue 4 | e61848