Cancer Research TCM Europe 2010 Workshops
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 67, 7525, August 1, 2007. doi: 10.1158/0008-5472.CAN-07-0668
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liao, C.-P.
Right arrow Articles by Roy-Burman, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liao, C.-P.
Right arrow Articles by Roy-Burman, P.

Clinical Research

Mouse Models of Prostate Adenocarcinoma with the Capacity to Monitor Spontaneous Carcinogenesis by Bioluminescence or Fluorescence

Chun-Peng Liao1, Chen Zhong2, Gohar Saribekyan2, James Bading3, Ryan Park3, Peter S. Conti3, Rex Moats3, Anton Berns5, Wei Shi4, Zongxiang Zhou6, Alexander Yu Nikitin6 and Pradip Roy-Burman1,2

Departments of 1 Biochemistry and Molecular Biology, 2 Pathology, and 3 Radiology, Keck School of Medicine, University of Southern California; 4 Developmental Biology Program, Children's Hospital Los Angeles, Los Angeles, California; 5 Division of Molecular Genetics, Netherlands Cancer Institute, Amsterdam, The Netherlands; and 6 Department of Biomedical Sciences, Cornell University, Ithaca, New York

Requests for reprints: Pradip Roy-Burman, Department of Pathology, Keck School of Medicine, University of Southern California, 2011 Zonal Avenue, Los Angeles, CA 90033. Phone: 323-442-1184; Fax: 323-442-3049; E-mail: royburma{at}usc.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
The application of Cre/loxP technology has resulted in a new generation of conditional mouse models of prostate cancer. Here, we describe the improvement of the conditional Pten deletion model of prostate adenocarcinoma by combining it with either a conditional luciferase or enhanced green fluorescent protein reporter line. In these models, the recombination mechanism that inactivates the Pten alleles also activates the reporter gene. In the luciferase reporter model, the growth of the primary cancer can be followed noninvasively by bioluminescence imaging (BLI). Surgical castration of tumor-bearing animals leads to a reduced bioluminescence signal corresponding to tumor regression that is verified at necropsy. When castrated animals are maintained, the emergence of androgen depletion–independent cancer is detected using BLI at times varying from 7 to 28 weeks postcastration. The ability to monitor growth, regression, or relapse of the tumor with the use of BLI lead to the collection of tumors at different stages of development. By comparing the distribution of phenotypically distinct populations of epithelial cells in cancer tissues, we noted that the degree of hyperplasia of cells with neuroendocrine differentiation significantly increases in the recurrent cancer relative to the primary cancer, a characteristic which may parallel the appearance of a neuroendocrine phenotype in human androgen depletion–independent cancer. The enhanced green fluorescent protein model, at necropsy, can provide an opportunity to locate or assess tumor volume or to isolate enriched populations of cancer cells from tumor tissues via fluorescence-based technologies. These refined models should be useful in the elucidation of mechanisms of prostate cancer progression, and for the development of approaches to preclinical intervention. [Cancer Res 2007;67(15):7525–33]


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
There are two primary objectives in the modeling human prostate cancer in mice. The first is to recapitulate the pathophysiologic characteristics of the human disease in a "natural" manner in immunocompetent mice to facilitate the understanding of the complex molecular mechanisms underlying prostate cancer. The second is to use the models to develop or test new targeted therapies. Because the growth and progression of prostate cancer varies widely from animal to animal, postmortem tissue analysis of large cohorts of animals is often required to derive statistically meaningful data. The recent development of miniaturized noninvasive imaging techniques has made it possible to follow tumor progression in individual mice and other small animals (1). Among the different noninvasive imaging techniques with potential to provide tumor-specific information in individual living mice, bioluminescence imaging (BLI) has drawn much attention (2, 3). Recently, transgenic mouse models were developed to express firefly luciferase specifically in the prostate in an androgen-dependent fashion (4–6). In one case, such a system was combined with a SV40 T antigen–induced prostate tumor model for the purpose of following tumor development using BLI (5).

Several prostate epithelium–specific Cre/loxP systems have been developed (7–10), and were shown to be useful in generating prostate preneoplastic or neoplastic mouse models (7, 9–15). However, there are no reports of the incorporation of Cre/loxP-mediated luciferase activation in any of these systems. The Cre line, PB-Cre4, that we developed (8) has been proven to be robust in not only causing recombination of both alleles of a target gene (11–13, 16) but also in the efficient switching of up to four alleles in a single cell (14, 15). Encouraged by this powerful expression of Cre in the prostate, we proceeded to increase the efficiency and utility of the conditional Pten deletion (cPten–/–) model (12, 16, 17) by combining it with a conditional reporter allele based on the idea that the cells in which Pten alleles are deleted would also be subject to activation of the reporter allele by the same Cre-mediated recombination. We used the luciferase reporter line (18) that was shown to be useful in the visualization of spontaneous tumorigenesis in other tissue-specific conditional mouse models. Similarly, the ROSA26-EGFPloxP line (19) offered an opportunity to implement fluorescence imaging of the cPten–/– model.

Here, we present evidence that each of the conditional reporter alleles, luciferase and enhanced green fluorescent protein (EGFP) could be used to assess the primary tumor burden in the prostate. Furthermore, we show that longitudinal BLI in the combinatorial luciferase model can register the growth of the primary tumor, its regression through androgen depletion, and most importantly, in the recurrence of androgen depletion–independent (ADI) cancer (20), all in individual living mice.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Generation of compound transgenic mice. The luciferase reporter construct that was used to generate the luciferase transgenic line contained the ß-actin promoter/loxP/GFP/polyadenylation sequence/loxP/luciferase/polyadenylation sequence in 5' to 3' orientation (18). Although the transgene expression driven by the ß-actin promoter is ubiquitous in this mouse line, the transcription is terminated at the floxed polyadenylation site before the luciferase gene. However, there is luciferase gene expression when the floxed polyadenylation sequence is deleted by Cre-mediated recombination. In EGFP reporter mice, a floxed DNA fragment that includes PGK-cytosine deaminase/PGK-Puro/"Stop" sequences is inserted between the ROSA26 promoter and the EGFP coding region, so that EGFP is only expressed on Cre-mediated excision of this floxed DNA segment (19). To generate mice with conditional inactivation of Pten alleles and activation of luciferase or EGFP reporter, male mice of ARR2PB promoter-driven (21), prostate epithelium–specific Cre line PB-Cre4 on C57B/6xDBA2 background (8) were first crossed with homozygous floxed Pten mice on the 129/BALB/c background (22). The male offspring carrying floxed Pten alleles and PB-Cre4 transgene (cPten–/–) were then crossed to floxed luciferase (L) or floxed EGFP (G) reporter female mice carrying the homozygous floxed Pten gene. The background of the L reporter line was FVB/N (18), whereas that of the G line was C57BL/6 (19). The resulting cPten–/–L or cPten–/–G mice were consequently of mixed genetic background, a fact that was readily apparent from variable skin color ranging from white to brown to black. Only male compound mutant mice were used for the study. Nonrecombinant littermates, such as Ptenfloxed/floxedL or Ptenfloxed/floxedG without the Cre gene, served as controls. All mice were maintained under identical conditions and animal experimentation was conducted using the standards for humane care in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

BLI. Mice were given a single i.p. injection of ketamine (50 mg/kg) and xylazine (10 mg/kg) followed by i.v. injection of luciferin (50 mg/kg). After waiting for 4.5 min to allow proper distribution of luciferin, the mice were placed in the chamber of an IVIS 200 optical imaging system (Xenogen Corp.). Photons were collected for a period of 1 min, and images were analyzed using LIVING IMAGE software v. 2.50 (Xenogen). BLI of excised tissues was done after euthanasia of the animals previously injected i.v. with luciferin. Signal intensity was quantified for defined regions of interest as photon count rate per unit body area per unit solid angle subtended by the detector (units of photons/s/cm2/steradian). Images simulating the three-dimensional location of bioluminescent sources within mice were generated using the single-view diffuse tomography capability of the IVIS 200 and LIVING IMAGE 3D v. 2.50.

Immunohistochemistry and immunofluorescence. Immunohistochemical analysis of parallel paraffin sections of paraformaldehyde-fixed tissue was done by a modified avidin-biotin complex (ABC) technique as described previously (15). Briefly, antigen retrieval was done by boiling the slides in 10 mmol/L of citric buffer (pH 6.0) for 15 min. Antibodies to Ki67 (1:1,000 dilution; Novocastra Laboratories), cytokeratin 8 (CK8; 1:50, TROMA-1 antibody; Developmental Studies Hybridoma Bank, University of Iowa), cytokeratin 5 (CK5; 1:1,000; Covance), androgen receptor (AR; 1:60; Upstate), and synaptophysin (1:100 or 1:20; Dako) or Chromogranin A (1:100; Zymed Laboratories) were incubated overnight at 4°C with deparaffinized sections. Sections were subsequently incubated with biotinylated secondary antibody for 30 min at room temperature and then detected with the ABC Elite kit (Vector Laboratories) and 3,3'-diaminobenzidine (Sigma) as substrate. All transverse sections of the whole prostate were scanned by ScanScope (Aperio Technologies) with a 40x objective followed by lossless compression and assessment of all immunostainings in identical anatomic regions. Double immunofluorescence staining was done with rat monoclonal antibody CK8 (1:5) and rabbit polyclonal antibody synaptophysin (1:10), or mouse monoclonal antibody synaptophysin (1:2) and rabbit polyclonal antibody AR (1:10), followed by FITC-conjugated anti-rat or FITC-conjugated anti-mouse and rhodamine (TRITC)-conjugated anti-rabbit secondary antibodies (Jackson ImmunoResearch Laboratories). To stain cell nuclei, sections were incubated with a 10 µg/mL solution of 4',6-diamidino-2-phenylindole (DAPI) for 4 min.

Detection of EGFP in tissues and tissue sections. EGFP expression in tissues was initially examined under a Leica Z16 APO (Leica Microsystems Inc.) dissecting microscope with a GFP filter. The freshly collected tissues from euthanized mice were placed under the microscope and images recorded with Leica DFC300FX Digital Color Camera (Leica Microsystems Inc.) and ImagePro MC5.1 (Media Cybernetics Inc.). We used confocal microscopy following a modification of a previously described procedure in order to detect EGFP fluorescence in tissue sections (23). Briefly, tissues were removed from euthanized mice and then fixed for 2 h in 4% paraformaldehyde at room temperature. Fixed tissues were washed once in PBS and embedded in Tissue-Tek at room temperature. The embedded tissues were kept in the dark at 4°C overnight and then slowly frozen at –80°C in a container wrapped with an insulating material. Tissues were placed at –20°C for 30 min before sectioning. The frozen sections (10 µm) were covered with mounting medium containing DAPI. The slides were kept overnight at –20°C, then examined by Zeiss LSM-510 laser scanning confocal microscope (Carl Zeiss). Images were recorded using the LSM 510 software version 3.2 SP2.

Reverse transcription-PCR. Total RNA was extracted by TRIzol Reagent (Invitrogen) following the protocol recommended by the manufacturer. RNA (2 µg) was reverse-transcribed by random hexamers and SuperScript III reverse transcriptase (Invitrogen) in a volume of 21 µL. The reverse transcription reaction (2 µL) was used as a template in a 20 µL PCR reaction mixture using a primer set specific for the firefly luciferase reporter gene (24). An equivalent amount of RNA without reverse transcription served as negative control. As a positive control for amplification from the cDNA, we used the Gapdh primers, 5'-CAGCCTCGTCCCGTAGACAAAATGG-3'and 3'-TTCTGGGTGGCAGTGATGGCATGGA leading to a product of 520 bp.

Preparation of single-cell suspensions for FACS analyses. Freshly collected prostate tissues were minced with crossed scalpels (size 11 blades), transferred to a 5 mL tube and incubated in DMEM/F12 medium containing 10% fetal bovine serum, collagenase (1 mg/mL), hyaluronidase (1 mg/mL), and DNase I (1 µg/mL) at 37°C overnight on a rotator. Collagenase, hyluronidase, and DNase I were purchased from Sigma. After low-speed centrifugation, the single cells and cell clumps were collected and subjected to treatment with 0.05% Trypsin-EDTA for 10 min and then diluted with an equal amount of medium containing 10% fetal bovine serum. The preparation was passed through 100 and 40 µm cell strainers sequentially for the isolation of single cell suspensions for FACS analyses. GFP-expressing cells were detected using the FL1 channel (absorption spectra 530/30 nm) using a MoFlo Cell Sorter with Summit v.3.1 (Dako Cytomation).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Characterization of cPten–/–L and cPten–/–G mice. Considering that there could be significant differences in the recombination efficiency of different alleles in vivo (24), we first examined whether the reporter genes were expressed in a tissue-specific manner. For cPten–/–L mice, we administered luciferin, the substrate for luciferase by i.v. injection to male mice beginning at 7 weeks of age, and sequential images were recorded from the ventral side. Each experimental mouse was paired with a corresponding littermate control lacking the Cre transgene. A representative BLI picture of such a pair of mice taken ~5 min after luciferin injection is shown in Fig. 1A . The image shows a strong bioluminescent signal, specifically from an area corresponding to the prostate, whereas the control mouse shows only a very weak background signal. We also used three-dimensional imaging to further identify where the bioluminescence originated in living animals. For this purpose, we imaged a sexually mature (9-week-old) Cre-positive Ptenfloxed/wild-typeL male (no. 1492) that should have a noncancerous prostate but with Cre-mediated reporter activation. The ventral image (Fig. 1B) confirmed that the sources of relatively strong signals (dark dots) were located in the lower abdomen, and the lateral image showed the source to be concentrated anteriorly in the expected location of the prostate gland. When a cPten–/–L mouse (no. 2425) was imaged similarly, the results also corresponded with the location of the prostate and, as expected, displayed a much higher (~100-fold) signal intensity (Fig. 1B).


Figure 1
View larger version (42K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. In vivo and ex vivo bioluminescence images of cPten–/–L mice. A, comparison of in vivo BLI signals from the abdominal region in a 12-week-old cPten–/–L male (no. 1424), and its littermate control (no. 1423) that lacked the Cre transgene. B, three-dimensional imaging of a 9-week-old cPtenfloxed/wild-typeL male (no. 1492) and a 24-week-old cPten–/–L male (no. 2425) indicating the major source of BLI signal to be inside the abdomen at a location corresponding to the prostate. The BLI intensity from the noncancerous recombined mouse (no. 1492) appeared 2 logs lower than that observed in recombined cancerous mouse no. 2425. Reverse rainbow color table, photo density detected from the surface; the BlackRed color table, the source intensity from low to high using a color scale from black to red. C, ex vivo bioluminescence images of the UGS of a 15-week-old cPten–/–L (no. 2160) shown next to its control (no. 2159). The regions of the prostate (P), seminal vesicle (SV), spermatic cord (SC), and testes (T). D, BLI indicative of lymph node metastasis in an 18-week-old cPten–/–L (no. 1309). Besides the strong signal from the prostate area, some of the other "hotspots" corresponded to regional lumbar (L) and caudal (C) lymph nodes.

 
To confirm the prostate tissue–specificity of luciferase expression, we examined several other cPten–/–L and Ptenfloxed/floxedL pairs, ranging in age from 10 to 15 weeks by ex vivo BLI imaging. Mice were euthanized immediately following luciferin administration and the urogenital system (UGS) was isolated and imaged. An example of the images obtained 18 min after luciferin administration is displayed in Fig. 1C. It is clear that the UGS from the cPten–/–L mouse presented a high level of BLI signal from prostate lobes, although weak but still detectable signals were also found in regions of the seminal vesicles, spermatic cords, and testes. The signal intensity was either very low or not detected in the reproductive organs of the littermate controls when examined under the same conditions of isolation and imaging. Taken together, these results supported the validity of BLI to study prostate tumorigenesis in vivo in the cPten–/–L mice.

However, it should be noted that we encountered several factors, some related to the mixed genetic background of the animals that influenced the intensity of the BLI signal. We found that i.v. injection of the substrate was superior to i.p. injection in reducing the bioluminescence background. The strength of the light was also significantly affected by the fluid content in the bladder. Emitted light is absorbed by the skin or fur, and a full bladder significantly increases the depth and hence the attenuation of the signal (25). For these reasons, we routinely massaged all animals in an attempt to release urine, and also shaved mice with dark fur to reduce attenuation of the light. These limitations hindered comparisons of signal intensity between animals and between different time points within the same animal. Besides indicating the primary cancer, BLI was also effective in revealing lymph node metastases in a few instances. This was illustrated in the case of an 18-week-old cPten–/–L mouse shown in Fig. 1D. The dorsal image of this mouse revealed three "hotspots" superior to the prostate region. When the image of the spots was enlarged and flipped left to right, it matched very well with the ventral location of the caudal and lumbar lymph nodes.7 These lymph nodes, located in proximity to the bifurcation of the aorta in the abdominal cavity and close to the backbone, were previously identified as preferential sites for metastasis in orthotopic mouse models of prostate cancer (26, 27). The frequency of detection of lymph node metastases in living animals by the BLI approach was, however, quite low because we could identify only three such cases from ~120 cPten–/–L live mice that were imaged. Because the lymphovascular metastatic lesions in the cPten–/– model are mostly microscopic in size, and for that matter many could even be missed during histology sectioning (12), it is not surprising to find the sensitivity of BLI to detect metastases in this model to be much lower than that of histology sectioning. In addition, it remains to be determined whether changes in the mixed genetic background in cPten–/–L relative to cPten–/– mice might have altered the incidence of micrometastases.

The prostate tissues of different cPten–/–G male mice ranging in age from 16 to 52 weeks were evaluated with UV epifluorescence microscopy. A view of the fluorescence images derived from UGS of a cPten–/–G mouse relative to that of controls is presented in Fig. 2 . The autofluorescence in the prostate tissue is minimal in the Ptenfloxed/floxedG mouse lacking the Cre gene (Fig. 2A). The only tissue exhibiting significant autofluorescence seemed to be the bladder. In comparison with this background, a Cre-containing noncancerous prostate of a Ptenfloxed/wild-typeG mouse showed fluorescence, although with variable distribution of intensities within the lobes (Fig. 2B). Although the dorsolateral prostate and ventral prostate had uneven but still notable fluorescence throughout, the anterior prostate exhibited the least and only focal fluorescence. This apparent inefficiency in recombination or marker gene expression in the anterior prostate appeared similar to the pattern that we described earlier in PB-Cre4/R26R double transgenic mice (8). However, the overall efficiency in EGFP expression in the mixed genetic background of the current EGFP mice was relatively weak. To what extent this property might be a consequence of the efficiency of recombination of the floxed EGFP allele or an epigenetic modification of marker gene expression is unclear.


Figure 2
View larger version (52K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Fluorescence analysis of tissues from cPten–/–G males. A, dorsal and ventral views of UGS of a 24-week-old Ptenfloxed/wild-typeG male mouse lacking the Cre gene (no. 465) shown under bright-field or UV. B, similar views of a 48-week-old cPtenfloxed/wild-typeG mouse (no. 343) carrying the Cre gene. C, views of the prostate of cPten–/–G mouse (no. 464) for which the control was no. 465 (A). On the photomicrographs obtained under bright-field, the regions of dorsolateral prostate (DLP), anterior prostate (AP), ventral prostate (VP), seminal vesicle (SV), and bladder (B) are indicated. Bottom right, the glandular structure within the tumor in VP (arrow). D, confocal photomicrographs after nuclear DAPI staining, EGFP fluorescence, and their overlay are illustrated using the prostate of a 20-week-old cPten–/–G male (no. 111) relative to its control (no. 112).

 
It is noteworthy, however, that tumor areas in the prostate, irrespective of the lobe of origin, had significantly higher levels of fluorescence. This is shown in Fig. 2C in a 24-week-old cPten–/–G mouse for which the littermate control is presented in Fig. 2A. Of the 15 cPten–/–G mice sacrificed to date, fluorescence from EGFP in the prostate tumors was detected as early as 4 months of age. Confocal microscopic analysis of another set of animals, aged 20 weeks, identified prostatic epithelium as the primary source of GFP fluorescence in the cPten–/–G mouse (Fig. 2D), with the stromal cells showing only scattered spots of autofluorescence detectable in both the recombinant and control tissues. Similar to the cPten–/–L system, lymph nodes with prominent fluorescence were rarely identified in the cPten–/–G mouse except for a single caudal node that exhibited focally intense fluorescence (data not shown). Although autofluorescence of tissue materials could generally complicate the analysis of EGFP fluorescence, we did not encounter, as described above, a high level of autofluorescence in the tissue of primary interest, the prostate. The bladder had detectable autofluorescence, as did tissues of the gastrointestinal system which contains bacterial flora. There was some variable and patchy fluorescence in testes and spermatic cords irrespective of the presence or absence of the Cre gene, consistent with autofluorescence.

Longitudinal monitoring of prostate tumor growth, regression, and recurrence in cPten–/–L mice. Beginning at 6 to 8 weeks of age, a cohort of mice was monitored using BLI at intervals of ≥2 weeks up to ~52 weeks. Images obtained from one such mouse are presented in Fig. 3A , and the measured BLI intensity originating from the region of interest for this and some other mice, including two Cre-negative littermate controls is graphically presented in Fig. 3B. In general, the signal increased with mouse age, although occasional decreases in BLI intensity followed by an increase were also apparent. For the cPten–/–L mouse no. 1470, the BLI increased at least 5-fold over the course of 43 weeks, whereas that for no. 1425 increased 50-fold during the 30 weeks of observation. There was, however, a 2-fold difference in the maximal signal strength attained between these animals. In other mice, the signal intensities were lower, although an increase with age was still quite evident. Considering that the spatial resolution of BLI is limited, the fact that tumor sizes varied from animal to animal, and that several other factors could influence BLI intensity, the consistent finding of a temporal increase in BLI in individual cPten–/–L mice was very encouraging, especially when littermate controls, monitored under analogous conditions virtually lacked detectable BLI throughout the course of the observation. It is clear that this technology will allow the evaluation of tumor growth in individual animals, each one serving as its own control. Prostate tissues of animals with BLI that were sacrificed between 12 and 52 weeks of age were confirmed histopathologically to harbor adenocarcinomas, consistent with previous reports (12, 16, 17).


Figure 3
View larger version (57K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Longitudinal BLI monitoring of living cPten–/–L animals. A, serial images of an individual male (no. 1470) over a period exceeding 1 y. B, BLI signal measurements for a cohort of four cPten–/–L (solid lines) and two control (dotted lines) animals. C, BLI images of a cPten–/–L mouse (no. 1588) at times before and after castration that was carried out at ~5 mo of age. D, BLI monitoring of a cohort of several cPten–/–L animals (solid lines) and four controls (dotted lines) castrated at a similar age. Vertical dotted line, the castration time point. Signal intensity was measured for regions of interest drawn around the lower abdomen.

 
To investigate the effect of androgen depletion on the growth of prostate cancer in the cPten–/–L model, a cohort of recombined male mice were castrated at ~5 months of age after the BLI signal had been recorded from the prostate region. Imaging was then done at 2 to 3 weeks postcastration, and, in all cases, a sharp decline in BLI was evident. This was illustrated by a set of serial images from a mouse in Fig. 3C, and by intensity curves generated from this mouse and others, including four different littermate controls in Fig. 3D. Following a period of latency varying from 7 to 28 weeks postcastration, the signal from the prostate began to increase again indicating the emergence of ADI-proliferating lesions. This was very encouraging because it showed that the growth of a recurrent tumor, like that of the primary tumor, could be followed using BLI in living animals.

In some older cPten–/–L mice, we also noted a variable signal in the thoracic area. These mice were mostly 52 weeks of age or more, although a few were as young as 28 weeks. The BLI in the upper body region, seen in both intact and castrated older mice (Fig. 3A and D), was not detected in the littermate controls lacking the Cre gene. When we imaged various isolated tissues from some of these animals by ex vivo BLI, we detected intensities higher than the background in the heart and the rib cage. The observation in the rib cage led us to search for potential skeletal metastases using histopathologic analyses of bone marrow and bone tissues.

However, we did not detect cancer cells in the bone structures in question (data not shown). The possibility that Cre-mediated recombination in inflammatory cells or other cell types could have occurred in older animals remains to be examined. Although BLI from nonrecombined tissues in the luciferase reporter line was described to be low (18), in consideration of the mixed genetic background of the cPten–/–L mice, we tested whether leaky expression of the reporter could be a confounding factor in our analysis. When tissues (lung, liver, spleen, body muscle, kidney, brain, and testes) from Ptenfloxed/floxedL or Ptenfloxed/wild-typeL allelic mice without the Cre gene were examined for luciferase transcripts by reverse transcription-PCR, readily detectable expression was found in the muscle (Fig. 4A ). Thus, it seems that the ß-actin promoter-driven luciferase reporter has a floxed polyadenylation sequence before the luciferase open reading frame could undergo detectable read-through transcription in some tissues. The reason for this effect remains unclear but could be related to rearrangements that occur in the cancatemeric transgene or to different insertion sites during segregation (17). However, this was most likely not a significant problem in altering BLI specificity because signals from the Ptenfloxed/floxedL mice were used to establish the baseline.


Figure 4
View larger version (42K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Other characteristics of the models. A, tissues from two nonrecombined males, one Ptenfloxed/wild-typeL (no. 1382), and the other Ptenfloxed/floxedL (no. 1410), both lacking the Cre gene, were analyzed by reverse transcription-PCR for potential leakiness in luciferase gene transcription. Of the various tissues examined, muscle displayed a readily detectable PCR band (lanes 2 and 4) that was not present when the reverse transcription step was omitted (lanes 1 and 3). B and D, detection of metastatic cells in the subcapsular sinus of a lumbar lymph node from a cPten–/–L mouse. Cells with nuclear staining of AR (B) and cytoplasmic staining of CK8 (D) were detected. The ABC Elite method was used with hematoxylin counterstaining (bar, 10 µm). C, illustration of fluorescence-based sorting of cells from the neoplastic prostate of a cPten–/–G mouse (no. 118) relative to the prostate of a littermate control (no. 112).

 
Identification and phenotypic characterization of cancer cells in the models. The regional lymph nodes from the cPten–/–L mice that were found to be positive following BLI were examined by histopathology. Metastasis of prostate cancer cells into these lymph nodes was evident when tissue sections were examined by immunohistochemistry. This was illustrated by the detection of metastatic cells within a lumbar lymph node that stained positive for AR and CK8 (Fig. 4B and D). With regard to the cPten–/–G model, we attempted to sort cells from primary tumors on the basis of EGFP fluorescence. These results relative to the corresponding control materials are depicted in Fig. 4C. Although there was no evidence of EGFP-expressing cells in the control prostate, the neoplastic prostate indicated 5% of the total cell population to be EGFP-expressing. Further analyses of four other such pairs indicated a similar ability to sort ~4% to 6% of EGFP-positive cells. The extent of homogeneity and the phenotypic characteristics of the sorted cells, however, remains to be determined.

Using the cPten–/–L model, we wished to determine if the distribution of phenotypically distinct populations of prostate epithelial cells might be altered in the recurrent cancer relative to the primary cancer. We used CK8 and AR as markers for the secretory luminal epithelial cells, CK5 for basal cells, synaptophysin for neuroendocrine cells, and Ki67 as a marker for proliferation. The results of such analyses with three different pairs of tumors were similar. Representative photomicrographs are shown in Fig. 5 . Epithelial cells of the normal prostate and primary tumors had positive nuclear staining for AR. In contrast, epithelial cells from recurrent tumor displayed mostly a diffuse cytoplasmic AR staining (Fig. 5A). A majority of the atypical cells in the primary tumor and the recurrent tumor were CK8-positive, and basal cells, marked by CK5 expression and separated from the stroma by the basement membrane in the normal prostate, were commonly found inside both the primary and recurrent tumors. Although only a few cells in the normal epithelium expressed synaptophysin, many cells in primary tumors and many more in recurrent tumors expressed this neuroendocrine marker. These cells were arranged both individually and in groups, and in one case, substituted the entire thickness of the epithelium. Their neuroendocrine differentiation was confirmed by using an additional antibody against synaptophysin as well as by detection of Chromogranin A. The characteristics of the synaptophysin-expressing cells in the primary tumor were further examined by immunofluorescence analysis (Fig. 5B). Interestingly, synaptophysin and AR colocalized in some (arrowhead) but not all (arrow) cells. Similarly, colocalization of synaptophysin and Ki67 were detected in some (arrowhead) but not all (arrow) cells and no colocalization of CK5 and synaptophysin was observed.


Figure 5
View larger version (56K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. Comparative immunohistochemical analysis of epithelial cells from the normal prostate and primary and recurrent tumors. A, staining for H&E, AR, CK8, CK5, and synaptophysin on serial sections of the normal (control) and neoplastic (primary tumor and recurrent tumor) epithelium of the dorsolateral prostatic lobes. Arrows, areas enlarged in the insets. Note that prior to castration, nuclear AR was detected in many of the normal prostatic cells and in virtually all neoplastic cells. After castration, the expression of AR was reduced and mainly cytoplasmic. The majority of neoplastic cells were CK8-positive in both primary and recurrent tumors. The number of CK5-positive cells was increased but not equally in all glands. Individual CK5 cells and their clusters were located within both basal and luminal layers (arrow). Synaptophysin-positive cells were very rare in the distal part of the normal prostate, and punctate synaptophysin staining identifies nerve terminals (arrow). Primary tumors have an increased number of synaptophysin-positive cells (arrow) and their population further increases in castrated mice (arrow). Bar, low magnification (100 µm); high magnification (33 µm). B, characterization of synaptophysin-positive cells by coimmunofluorescence. Top, colocalization of synaptophysin (red) and AR (green) in some (arrowhead) but not all (arrow) cells. Middle, colocalization of synaptophysin (red) and Ki67 (green) in some (arrowhead) but not all (arrow) cells. Bottom, absence of colocalization of CK5 (green) and synaptophysin (red, arrows). DAPI counterstaining (blue) was used for multiple immunofluorescence. Bar, 50 µm.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
BLI of prostate cancer in xenograft and spontaneous tumor mouse models has been described (2, 5, 28–30). In the spontaneous models reported to date, the detection of BLI was based on prostate epithelium–specific and androgen-responsive promoter-driven expression of a luciferase reporter gene (5, 30). Our approach has been different in the sense that the prostate epithelial cells in which Pten alleles were inactivated by Cre/loxP were also targeted by the same mechanism for activation of the reporter gene, either luciferase or EGFP. The initial recombination processes require Cre expression, which is regulated by androgen. However, once the genes undergo recombination, the affected cells or their progeny become independent of androgen because neither the Pten allele nor the reporter transgene requires androgen for expression. Thus, the models are ideal for monitoring growth, regression, or relapse of the cancer irrespective of hormonal manipulations, such as androgen deprivation. In fact, we clearly show in the cPten–/–L model that BLI can register the growth kinetics of the prostate cancer before and after castration, and most significantly, demarcate the time points when the ADI recurrent disease emerges. The capacity to ascertain the growth of the recurrent cancer noninvasively is an achievement that, to our knowledge, has not been reported before. This has many implications. We have only presented a snapshot of how the model could be exploited to delineate the characteristics of recurrent cancer from those of the primary tumor. One of the common features of recurrent prostate cancers is neuroendocrine differentiation. Human prostate adenocarcinomas exhibit at least focal positivity for neuroendocrine markers ranging from 30% to 100% of the cases (31–35). Neuroendocrine differentiation has been reported to increase in advanced tumors and ADI tumors (36–44). However, some other studies have led to findings that differed from such a correlation. Thus, the role of focal neuroendocrine differentiation in prostate cancer and its connection with the development of ADI neoplasms remains controversial. Our results indicate that in addition to well-established hyperplasia of CK5-positive basal cell compartment (45), Pten inactivation leads to the expansion of cells with neuroendocrine differentiation. The degree of hyperplasia of such differentiated cells increases after castration, which may parallel the appearance of the neuroendocrine phenotype in human ADI tumors. Given that not all synaptophysin-positive cells express AR and none express CK5, it is possible that those cells represent an expanding pool of luminal/neuroendocrine progenitor cells. With the availability of distinctive markers for prostate stem cells, the present model could also allow us to make a critical analysis of the hypothesis that recurrent cancer has an origin in cancer stem cells. This type of investigation will be facilitated by timing the collection of tumors at specific growth or regrowth points, an advantage that is provided by the model.

We have also identified certain limitations of the cPten–/–L model. First, there is detectable transcription from the luciferase reporter downstream of the poly-A termination signal in nonrecombined tissues, particularly muscle. Because the reporter transgene is driven by the ß-actin promoter (18), it is not unexpected that such read-through transcripts should turn out to be relatively more frequent in muscle tissue where the promoter should be most active. Initially, when we administered luciferin by i.p. injection, BLI signals were detected around the site of injection in both cPten–/–L and Ptenfloxed/floxedL controls. We speculate that such activity is related, at least in part, to the above-noted leakiness. We were, however, able to greatly minimize this problem by administering luciferin via the tail vein.

Another phenomenon that cannot be readily explained is the increasingly strong BLI that became apparent with time in the thoracic area of some older experimental animals but not in the control animals. We were particularly interested in the signal that is emitted from the rib cage. From extensive scrutiny of the associated skeletal structures, we determined that bone metastasis, if any, remains below the level of our detection. In spite of these limitations, the model seems to serve efficiently when BLI is used for detection, growth, atrophy, or relapse of prostate cancer at the primary organ site.

Although the cPten–/–L mice are most useful for tumor studies in live animals, cPten–/–G mice, on the other hand, provide a means to highlight and localize primary tumors at necropsy. Although it remains to be shown, the anticipated spontaneous EGFP expression in recurrent cancer should be similarly valuable. The ability to isolate fluorescent cell populations from primary or recurrent cancers should provide excellent resources for studies to derive clues that may distinguish cellular and molecular properties of these cancer cells of varied origin. Additionally, isolation of such cell types should facilitate the generation of cell lines for the study of signaling aberrations. The refined preclinical models that we have developed should help to uncover the underlying mechanisms of prostate cancer progression and serve as convenient study systems to validate tests used for diagnosis, prevention, or treatment.


   Acknowledgments
 
Grant support: NIH grants RO1 CA59705 and RO1 CA113392 (P. Roy-Burman), and in part, by California Institute for Regenerative Medicine Training grant T1-00004 (C-P. Liao), and NIH grants RO1 CA96823 and K26 RR1759 (A. Nikitin).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Hong Wu of the University of California Los Angeles for providing the mouse strain with the floxed Pten allele, David Boyko of Xenogen Corporation for the approval of the use of the luciferase reporter mouse strain for our research, David Hinton for advice about confocal microscopy, Peter Nichols and Sue Ellen Martin for histopathology, Robert Maxson for critically reading the manuscript, and all members of the Roy-Burman laboratory for assistance in various aspects of the work.


   Footnotes
 
Note: C-P. Liao and C. Zhong contributed equally to this work.

7 http://www.eulep.org/Necropsy_of_the_Mouse/printable.php Back

Received 2/19/07. Revised 4/16/07. Accepted 5/16/07.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Massoud TF, Gambhir SS. Molecular imaging in living subjects: seeing fundamental biological processes in a new light. Genes Dev 2003;17:545–80.[Free Full Text]
  2. Adams JY, Johnson M, Sato M, et al. Visualization of advanced human prostate cancer lesions in living mice by a targeted gene transfer vector and optical imaging. Nat Med 2002;8:891–7.[Medline]
  3. Lyons SK. Advances in imaging mouse tumour models in vivo. J Pathol 2005;205:194–205.[CrossRef][Medline]
  4. Hsieh CL, Xie Z, Liu ZY, et al. A luciferase transgenic mouse model: visualization of prostate development and its androgen responsiveness in live animals. J Mol Endocrinol 2005;35:293–304.[Abstract/Free Full Text]
  5. Lyons SK, Lim E, Clermont AO, et al. Noninvasive bioluminescence imaging of normal and spontaneously transformed prostate tissue in mice. Cancer Res 2006;66:4701–7.[Abstract/Free Full Text]
  6. Ellwood-Yen K, Wongvipat J, Sawyers C. Transgenic mouse model for rapid pharmacodynamic evaluation of antiandrogens. Cancer Res 2006;66:10513–6.[Abstract/Free Full Text]
  7. Maddison LA, Nahm H, DeMayo F, Greenberg NM. Prostate specific expression of Cre recombinase in transgenic mice. Genesis 2000;26:154–6.[CrossRef][Medline]
  8. Wu X, Wu J, Huang J, et al. Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation. Mech Dev 2001;101:61–9.[CrossRef][Medline]
  9. Abdulkadir SA, Magee JA, Peters TJ, et al. Conditional loss of Nkx3.1 in adult mice induces prostatic intraepithelial neoplasia. Mol Cell Biol 2002;22:1495–503.[Abstract/Free Full Text]
  10. Ma X, Ziel-van der Made AC, Autar B, et al. Targeted biallelic inactivation of Pten in the mouse prostate leads to prostate cancer accompanied by increased epithelial cell proliferation but not by reduced apoptosis. Cancer Res 2005;65:5730–9.[Abstract/Free Full Text]
  11. Huang J, Powell WC, Khodavirdi AC, et al. Prostatic intraepithelial neoplasia in mice with conditional disruption of the retinoid X receptor {alpha} allele in the prostate epithelium. Cancer Res 2002;62:4812–9.[Abstract/Free Full Text]
  12. Wang S, Gao J, Lei Q, et al. Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer Cell 2003;4:209–21.[CrossRef][Medline]
  13. Trotman LC, Niki M, Dotan ZA, et al. Pten dose dictates cancer progression in the prostate. PLoS Biol 2003;1:385–96.
  14. Chen Z, Trotman LC, Shaffer D, et al. Crucial role of p53-dependent cellular senescence in suppression of Pten-deficient tumorigenesis. Nature 2005;436:725–30.[CrossRef][Medline]
  15. Zhou Z, Flesken-Nikitin A, Corney DC, et al. Synergy of p53 and Rb deficiency in a conditional mouse model for metastatic prostate cancer. Cancer Res 2006;66:7889–98.[Abstract/Free Full Text]
  16. Zhong C, Saribekyan G, Liao CP, Cohen MB, Roy-Burman P. Cooperation between FGF8b overexpression and PTEN deficiency in prostate tumorigenesis. Cancer Res 2006;66:2188–94.[Abstract/Free Full Text]
  17. Khodavirdi AC, Song Z, Yang S, et al. Increased expression of osteopontin contributes to the progression of prostate cancer. Cancer Res 2006;66:883–8.[Abstract/Free Full Text]
  18. Lyons SK, Meuwissen R, Krimpenfort P, Berns A. The generation of a conditional reporter that enables bioluminescence imaging of Cre/loxP-dependent tumorigenesis in mice. Cancer Res 2003;63:7042–6.[Abstract/Free Full Text]
  19. Mao X, Fujiwara Y, Chapdelaine A, Yang H, Orkin SH. Activation of EGFP expression by Cre-mediated excision in a new ROSA26 reporter mouse strain. Blood 2001;97:324–6.[Abstract/Free Full Text]
  20. Roy-Burman P, Tindall DJ, Robins DM, et al. Androgens and prostate cancer: are the descriptors valid? Cancer Biol Ther 2005;4:4–5.[Medline]
  21. Zhang J, Thomas TZ, Kasper S, Matusik RJ. A small composite probasin promoter confers high levels of prostate-specific gene expression through regulation by androgens and glucocorticoids in vitro and in vivo. Endocrinology 2000;141:4698–710.[Abstract/Free Full Text]
  22. Lesche R, Groszer M, Gao J, et al. Cre/loxP-mediated inactivation of the murine Pten tumor suppressor gene. Genesis 2002;32:148–9.[CrossRef][Medline]
  23. Shariatmadari R, Sipila PP, Huhtaniemi IT, Poutanen M. Improved technique for detection of enhanced green fluorescent protein in transgenic mice. Biotechniques 2001;30:1282–5.[Medline]
  24. Vooijs M, Jonkers J, Berns A. A highly efficient ligand-regulated Cre recombinase mouse line shows that LoxP recombination is position dependent. EMBO Rep 2001;2:292–7.[CrossRef][Medline]
  25. Rice BW, Cable MD, Nelson MB. In vivo imaging of light-emitting probes. J Biomed Opt 2001;6:432–40.[CrossRef][Medline]
  26. El Hilali N, Rubio N, Martinez-Villacampa M, Blanco J. Combined noninvasive imaging and luminometric quantification of luciferase-labeled human prostate tumors and metastases. Lab Invest 2002;82:1563–71.[Medline]
  27. Rubio N, Villacampa MM, El Hilali N, Blanco J. Metastatic burden in nude mice organs measured using prostate tumor PC-3 cells expressing the luciferase gene as a quantifiable tumor cell marker. Prostate 2000;44:133–43.[CrossRef][Medline]
  28. Jenkins DE, Oei Y, Hornig YS, et al. Bioluminescent imaging (BLI) to improve and refine traditional murine models of tumor growth and metastasis. Clin Exp Metastasis 2003;20:733–44.[CrossRef][Medline]
  29. Scatena CD, Hepner MA, Oei YA, et al. Imaging of bioluminescent LNCaP-luc-M6 tumors: a new animal model for the study of metastatic human prostate cancer. Prostate 2004;59:292–303.[CrossRef][Medline]
  30. Xie X, Luo Z, Slawin KM, Spencer DM. The EZC-prostate model: noninvasive prostate imaging in living mice. Mol Endocrinol 2004;18:722–32.[Abstract/Free Full Text]
  31. Abrahamsson PA. Neuroendocrine differentiation in prostatic carcinoma. Prostate 1999;39:135–48.[CrossRef][Medline]
  32. Abrahamsson PA, di Sant'Agnese PA. Neuroendocrine cells in the human prostate gland. J Androl 1993;14:307–9.[Free Full Text]
  33. di Sant'Agnese PA. Neuroendocrine differentiation in carcinoma of the prostate. Diagnostic, prognostic, and therapeutic implications. Cancer 1992;70:254–68.[CrossRef][Medline]
  34. di Sant'Agnese PA. Neuroendocrine differentiation in human prostatic carcinoma. Hum Pathol 1992;23:287–96.[CrossRef][Medline]
  35. Young RH, Srigley JR, Amin MB, Ulbright TM, Cubilla AL. Tumors of the prostate gland, seminal vesicles, male urethra, and penis. Washington: Armed Forces Institute of Pathology; 2000.
  36. Jongsma J, Oomen MH, Noordzij MA, et al. Androgen deprivation of the PC-310 [correction of prohormone convertase-310] human prostate cancer model system induces neuroendocrine differentiation. Cancer Res 2000;60:741–8.[Abstract/Free Full Text]
  37. di Sant'Agnese PA. Neuroendocrine differentiation in prostatic carcinoma: an update on recent developments. Ann Oncol 2001;12 Suppl 2:S135–40.[Abstract]
  38. Ito T, Yamamoto S, Ohno Y, et al. Up-regulation of neuroendocrine differentiation in prostate cancer after androgen deprivation therapy, degree and androgen independence. Oncol Rep 2001;8:1221–4.[Medline]
  39. Bonkhoff H, Stein U, Remberger K. Endocrine-paracrine cell types in the prostate and prostatic adenocarcinoma are postmitotic cells. Hum Pathol 1995;26:167–70.[CrossRef][Medline]
  40. Bonkhoff H, Wernert N, Dhom G, Remberger K. Relation of endocrine-paracrine cells to cell proliferation in normal, hyperplastic, and neoplastic human prostate. Prostate 1991;19:91–8.[Medline]
  41. Weinstein MH, Partin AW, Veltri RW, Epstein JI. Neuroendocrine differentiation in prostate cancer: enhanced prediction of progression after radical prostatectomy. Hum Pathol 1996;27:683–7.[CrossRef][Medline]
  42. Vashchenko N, Abrahamsson PA. Neuroendocrine differentiation in prostate cancer: implications for new treatment modalities. Eur Urol 2005;47:147–55.[CrossRef][Medline]
  43. Hoosein NM, Logothetis CJ, Chung LW. Differential effects of peptide hormones bombesin, vasoactive intestinal polypeptide and somatostatin analog RC-160 on the invasive capacity of human prostatic carcinoma cells. J Urol 1993;149:1209–13.[Medline]
  44. Huang J, Yao JL, di Sant'Agnese PA, Yang Q, Bourne PA, Na Y. Immunohistochemical characterization of neuroendocrine cells in prostate cancer. Prostate 2006;66:1399–406.[CrossRef][Medline]
  45. Wang S, Garcia AJ, Wu M, Lawson DA, Witte ON, Wu H. Pten deletion leads to the expansion of a prostatic stem/progenitor cell subpopulation and tumor initiation. Proc Natl Acad Sci U S A 2006;103:1480–5.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
Y. Fu, S. Wey, M. Wang, R. Ye, C.-P. Liao, P. Roy-Burman, and A. S. Lee
Pten null prostate tumorigenesis and AKT activation are blocked by targeted knockout of ER chaperone GRP78/BiP in prostate epithelium
PNAS, December 9, 2008; 105(49): 19444 - 19449.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
A. Pichler, J. L. Prior, G. D. Luker, and D. Piwnica-Worms
Generation of a highly inducible Gal4->Fluc universal reporter mouse for in vivo bioluminescence imaging
PNAS, October 14, 2008; 105(41): 15932 - 15937.
[Abstract] [Full Text] [PDF]

Home page
 Molecular Cancer TherapeuticsHome page
A. Tomioka, M. Tanaka, M. A. De Velasco, S. Anai, S. Takada, T. Kushibiki, Y. Tabata, C. J. Rosser, H. Uemura, and Y. Hirao
Delivery of PTEN via a novel gene microcapsule sensitizes prostate cancer cells to irradiation
Mol. Cancer Ther., July 1, 2008; 7(7): 1864 - 1870.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
S. Yang, L. K. Pham, C.-P. Liao, B. Frenkel, A. H. Reddi, and P. Roy-Burman
A Novel Bone Morphogenetic Protein Signaling in Heterotypic Cell Interactions in Prostate Cancer
Cancer Res., January 1, 2008; 68(1): 198 - 205.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liao, C.-P.
Right arrow Articles by Roy-Burman, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liao, C.-P.
Right arrow Articles by Roy-Burman, P.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online