| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Cell, Tumor, and Stem Cell Biology |
1 Baylor College of Medicine, Houston, Texas and 2 Fred Hutchinson Cancer Research Center, Seattle, Washington
Requests for reprints: David M. Spencer, Baylor College of Medicine, One Baylor Plaza/M929, Houston, TX 77030. Phone: 713-798-6475; Fax: 713-798-3033; E-mail: dspencer{at}bcm.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Approaches to noninvasive imaging differ by a number of key variables, such as spatial resolution, depth, imaging times, sensitivity, type of probe, cost, and others (reviewed in ref. 4). The highest spatial resolution is obtained with magnetic resonance imaging and computed tomography that can get close to 25 µm with long (>15 minutes) scan times. Molecular imaging using positron emission tomography (PET) and techniques based on detection of 
-rays produced by radiolabeled probes can provide resolution near 1 to 2 mm, but these methods require expensive equipment and relatively slow throughput. A new three-dimensional ultrasound microimaging technology has been developed with very good spatial resolution at low cost that has been used to image spontaneous models of prostate cancer (5). Alternatively, advances in charged coupled devices (CCD) have made optical imaging quite feasible for both fluorescence and especially bioluminescence imaging (BLI), where typically the oxidation of D-luciferin [D-(–)-2-(6'-hydroxy-2'-benzothiazolyl) thiazone-4-carboxylic acid] is observed following the Mg-ATP- and O2-dependent catalysis by firefly luciferase (6). The main advantages of BLI is sensitivity: Background is virtually absent in mammals imaged in light-tight specimen chambers attached to cryogenically cooled (
–105°C) CCD cameras. Another advantage of BLI is speed, as multiple animals can be scanned simultaneously in seconds to minutes in a single image acquisition with minimal postprocessing. Drawbacks include the partial opacity of highly vascularized tissue and melanin-pigmented skin, light scattering from tissue membranes and hair of most visible light, and planar two-dimensional images lacking depth information; however, newer "red-shifted" reporter proteins (i.e., producing light >600 nm) and other technologies may overcome these obstacles (7–10). Despite these limitations, imaging of as few as 100 tumor cells in the peritoneal cavity has been shown (10).
We previously reported development and characterization of a mouse model originally named EZC-prostate based on prostate epithelial-directed firefly (Photinus pyralis) luciferase and enhanced green fluorescent protein (EGFP) expression (11). The majority of expression was directed by prostate, with minor confounding signals coming from the testes and intestines. Longitudinal measurements showed strong androgen receptor (AR) responsiveness of luciferase reporter driven by the composite human kallikrein 2 (hK2) promoter, hk2-E3/P (12). In the present study, we describe two new prostate-directed expression systems that were used to build new luciferase reporter models, called EZC2-prostate and EZC3-prostate, based on the composite probasin promoter, ARR2PB (13), and a composite prostate-specific antigen promoter, PSA-E2/P (14), respectively. These new models direct even more exquisite prostate specificity than the original model, allowing for detection of distant soft-tissue metastasis when bred unto the TRAMP prostate cancer background (15, 16). Early detection of distant metastasis in living mice should greatly reduce the number of animals required in some studies. Moreover, localization of metastasis ex vivo should facilitate histologic analysis.
The use of AR-responsive reporter models also allows remote appraisal of AR activity during development and prostate tumor progression. Although AR levels are maintained or even elevated in the majority of cells within human prostate metastasis (17), AR levels typically drop in a subset of cells or poorly differentiated tumors, especially those displaying neuroendocrine markers (18–20). This observation has been used to suggest that AR signaling maintains the differentiated state of prostate epithelial cells, but loss of this signaling axis in the presence of other genetic and epigenetic events can drive dedifferentiation or reprogramming of cells in both human and animal models (reviewed in ref. 21). Consistent with this hypothesis, we observed a sharp decrease in reporter activity during cancer progression in bigenic models on both the TRAMP background and the JOCK1 model (22). Moreover, distant metastasis in bigenic EZC3-prostate x TRAMP (i.e., EZC-TRAMP) mice have low-level (but detectable) reporter signal, consistent with their prostatic origin.
Although palliative therapies can extend life span by several years, virtually all patients treated with surgical or medical androgen deprivation, combined with AR blockade, eventually succumb to an aggressive hormone-refractory prostate cancer (reviewed by ref. 23). Targets of androgen action include blockade of GnRH/LH-RH signaling (e.g., leuprolide, Abarelix), blockade of 5
-reductase that converts testosterone to the more potent androgen, 5-dihydrotestosterone (e.g., finasteride, dutasteride), nonsteroidal antiandrogens that act as AR antagonist (e.g., flutamide, bicalutamide), and selective AR modulators that likely prevent tissue-specific coactivator function (24, 25). To test the use of the EZC system in predicting response to hormone withdrawal therapy, we show that the GnRH antagonist, PPI-258, can almost completely eliminate prostate-directed reporter activity to levels comparable with that of castration (up to 3 orders of magnitude).
These new genetically engineered mouse models should accelerate the pace of drug development and our understanding of the role of AR signaling during prostate cancer progression.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
500 bp composite ARR2PB promoter originally from pSK/ARR2PB (13) or the 2.9 kb PSA-E2/P promoter (12), respectively, cloned upstream of the KCR intron in KBPA (26), which is, in turn, upstream of the modified firefly (Photinus sp.) luciferase gene from pGL3-Basic (Promega, Madison, WI) and a polyadenylic acid site from bovine growth hormone. Briefly, IRES-EGFP was released from plasmid phK2-E3/P-IRES-EGFP (11) by EcoRI digestion and religation to give phK2-E3/P-KBPA-Luc. To get pARR2PB-KBPA-Luc, the XbaI/SpeI ARR2PB fragment from pAdlox/ARR2PB-EGFP (27) was blunt ligated into the NotI/BamHI sites of phK2.E3/P-KBPA-Luc. Similarly, the NotI/XbaI (blunt) PSA-E2/P fragment from pSH1/PSA-E2/P-SEAP (27) was ligated into NotI/BamHI (blunt)–digested phK2.E3/P-KBPA-Luc to produce pPSA-E2/P-KBPA-Luc. Generation and screening of transgenic mice. The minimal eukaryotic-derived fragments were released from plasmids pARR2PB-KBPA-Luc and pPSA-E2/P-KBPA-Luc with BssHII and gel-purified with Gene Clean kit (Bio101, Vista, CA). The EZC2-prostate (ARR2PB-luc) and EZC3-prostate (PSA-E2/P-luc) transgenic mice were generated by microinjection of purified fragments into the male pronuclei of fertilized FVB oocytes by the Transgenic Core Lab at Baylor College of Medicine. EZC-prostate mice based on plasmid hK2-E3/P-Luc-IRES-EGFP were previously described (11). Mice were maintained in the Transgenic Mouse Facility, a pathogen-free environment, in compliance with Baylor College of Medicine policy. Mouse tail DNA was extracted using the DNeasy Tissue kit (Qiagen, Inc., Valencia, CA) and screened by PCR using primers for the firefly luciferase genes Luc-P1, 5'-AGCCAGCATGGAAGACGCCAAAAAC-3', and Luc-P2, 5'-ATCGCAGTATCCGGAATGATTTGATTGC-3'. DNA quality control primers for mouse casein: forward, 5'-GATGTGCTCCAGGCTAAAGTT-3', and reverse, 5'-AGAAACGGAATGTTGTGGAGT-3'.
Luciferase assays. To assay tissue-derived luciferase activity, animals were euthanized and dissected. Tissue specimens from the ventral, dorsolateral, and anterior prostate lobes were microdissected and stored temporarily on ice before homogenization in lysis buffer. Other tissue specimens, including testes, epididymis, seminal vesicle, penis, as well as brain, salivary glands, thymus, lung, heart, stomach, cecum, small intestine, large intestine, liver, adrenal gland, kidney, pancreas, and spleen were also analyzed to determine the tissue specificity of the EZC, EZC2, and EZC3 promoters. Tissues were homogenized with a PRO250 homogenizer (Pro Scientific, Inc., Oxford, CT) in 300 µL luciferase lysis buffer (Promega) containing 1/100 diluted protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN). Specimens were centrifuged at 8,000 rpm for 5 minutes and placed temporarily on ice. Luciferase activity of the cell lysates was measured with a TD 20/20 luminometer (Turner Designs, Inc., Mt. View, CA) and the protein concentration was determined using the detergent-compatible protein assay system (Bio-Rad, Hercules, CA) in a Beckman DU-640 spectrophotometer (Beckman Coulter, Fullerton, CA). The luminescence results are reported as relative light units per milligram of protein.
Histologic and immunohistochemical examination. The microdissected prostate parts and other tissues indicated above were fixed overnight in 10% formaldehyde and transferred to 1:1 formaldehyde/ethanol for 1 hour followed by transfer to 70% ethanol until processing. Tissues were processed in a series of increasing ethanol concentrations and embedded in paraffin wax. Five-micrometer sections were cut and stained with H&E. Immunostaining for firefly luciferase was also done. Briefly, sections were deparaffinized in xylene, rehydrated in decreasing ethanol from 100% to 80%, microwaved for 10 minutes in 10 mmol/L citrate buffer (pH 6.0) at 95°C to 99°C for antigen retrieval, and finally endogenous peroxidase activity was quenched with 3% hydrogen peroxide. Nonspecific binding was abolished with 10% Power Blocker (BioGenex, San Ramon, CA) for 10 minutes. Tissue sections were incubated overnight at 4°C with a 1:2,000 dilution of 10 mg/mL biotin-conjugated goat anti–firefly luciferase antibody (Abcam, Hartford, CT). After four PBS washes with 0.1% Tween 20 for 2 hours, sections were incubated with horseradish peroxidase–conjugated streptavidin using the Vectastin Elites ABC kit (Vector Laboratories, Inc., Burlingame, CA) for 45 minutes at room temperature. Peroxidase activity was revealed with 3',3'-diaminobenzidine (DAB) tetrahydrochlorate using a DAB kit (Vector Laboratories) according to the protocol provided. Finally, sections were washed in distilled water for terminating the reaction, counterstained with 1% methyl green, dehydrated, and mounted.
Imaging and quantification of bioluminescence data. Mice were anesthetized with a mixture of 1.5% isoflurane/air using an Inhalation Anesthesia System (VetEquip, Inc., Pleasant Hill, CA). D-Luciferin (Xenogen, Alameda, CA) was i.p. injected at 40 mg/kg mouse body weight (unless otherwise specified). Ten minutes after D-luciferin injection, mice were imaged with an IVIS Imaging System (Xenogen) with continuous 1% to 2% isoflurane exposure. Imaging variables were maintained for comparative analysis. Gray scale–reflected images and pseudocolorized images reflecting bioluminescence were superimposed and analyzed using the Living Image software (Xenogen). A region of interest (ROI) was manually selected over relevant regions of signal intensity. The area of the ROI was kept constant within experiments and the intensity was recorded as total photon counts per second per cm2 within a ROI. In some experiments, after imaging living mice, animals were euthanized and organs of interest were removed, arranged on black, bioluminescence-free paper, and ex vivo imaged within 30 minutes.
Castration of mice and administration of reagents. To study the effect of androgen ablation on EZC2-Prosate-driven firefly luciferase expression, 20-week-old male mice, were anesthetized with Rodent Combo Anesthetic III (37.6 mg/mL ketamine, 1.92 mg/mL xylazine, and 0.38 mg/mL acepromazine) at 240 mg/kg mouse weight, and cleaned around the scrotum with 70% ethanol. A 1-cm median incision was made at the tip of the scrotum. The testes lying in the sacs can be seen by placing pressure on the lower abdomen. A 5-mm incision was made into each sac, and the testes, epididymis, vas deferens, and spermatic blood vessels were pulled out. A single ligature was placed around the spermatic blood vessels and vas deferens. The testes and epididymis were removed by severing the blood vessels and vas deferens distal to the ligature. The remaining vas deferens was pushed back into the sac and the incision was sutured. The castrated mice were put on a heating pad until recovery and were given injectable Buprenex (buprenorphine hydrochloride; Reckitt Benckiser Healthcare UK, Ltd., England, United Kingdom). Sham-operated, age-matched males were used as controls. The orchiectomized mice were imaged biweekly until day 16 postcastration when animals were given testosterone pellets (10 mg/21-day release-Innovative Research of America, Sarasota, FL) for experimental groups and placebo for control groups. The mouse was shaved on dorsal side and cleaned with 70% ethanol. A 5-mm incision was made in the center of the back, the pellet was placed under the skin with forceps and the incision was clipped with surgical staples. The mice were measured biweekly and after 2 weeks animals were euthanized and imaged ex vivo as above. Mice treated with GnRH antagonist, PPI-258 (Praecis Pharma, Waltham, MA), were treated with a single s.c. 50 mg/kg dose (5 mg/mL 0.9% saline) of PPI-258-CMC, suspended in a time-release depot, designed to last for at least 4 weeks. For treatment of EZC-JOCK mice with chemical inducer of dimerization (CID), AP20187 was dissolved in 16.7% propanediol, 22.5% PEG400, 1.25% Tween 80, and injected i.p. biweekly at 2 mg/kg. Alternatively, carrier alone was injected.
Statistical analysis. Statistical significance was determined between indicated groups primarily using a nonpaired, two-sided Student's t test (Microsoft Excel 2004).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We were able to select new founder lines using both ARR2PB (one of three) and PSA-E2/P (two of four) that showed highly robust and prostate-specific expression, and have named them EZC2-prostate (EZC2) and EZC3-prostate (EZC3), respectively. Both new lines showed extremely high luminescence in the lower abdomen of living mice, requiring only seconds for imaging (Fig. 1A
). Dissection of these lines and ex vivo imaging of tissues confirmed that the major luminescence was derived from the relatively small (20 mg) ventral prostate lobes. Moreover, luciferase expression was almost undetectable in the intestines and all internal organs of both new lines (Fig. 1B). Additional relatively weak signals in the extremities of EZC3 mice are distant from the prostate and likely target tissues of prostate cancer metastasis in mice, such as draining lymph nodes, lungs, and liver. Thus, EZC2- and EZC3-prostate models reflect the highest overall signal intensity and tissue specificity ever seen in a prostate-directed bioluminescence model.
|
0.5%) came from a larger number of tissues, including dorsolateral prostate (7.8%), cecum (6.1%), anterior prostate (3.2%), large intestines (2.0%), testes (1.4%), spleen (1.4%), small intestines (1.1%), penis (0.5%), and thymus (0.4%; Fig. 2C). Thus, careful tissue analysis confirmed the robust expression and prostate specificity of EZC2 and EZC3.
|
100-fold) increase from 4 to 8 weeks (Fig. 3C). A somewhat more conservative increase was observed during puberty in EZC3-prostates (Supplementary Fig. S1), perhaps reflecting greater sensitivity of the PSA-E2/P promoter to lower testosterone levels. In EZC1 mice, the biggest temporal increase in reporter signal occurred earlier, between 3 and 4 weeks, although off-target reporter expression in their intestines may have partially confounded pubertal changes in AR activity during prostate development (data not shown). Thus, in all three of the EZC series of prostate models, AR activity in the prostate could be measured as a surrogate marker for prostate growth and development.
|
|
Imaging prostate cancer progression and metastasis with EZC-prostate models. One of the primary goals of generating organ-specific reporter mice is the ability to noninvasively image changes in the gland during normal development or pathologic states, such as neoplasia. To test whether the EZC-prostate model could be used to measure prostate hyperplasia and cancer progression. EZC1-, EZC2-, and EZC3-prostate mice were bred with two distinct animal models, the recently described inducible prostate intraepithelial neoplasia model, JOCK1 (22), and the well-characterized TRAMP model (15, 20). In the JOCK-1 model, administration of a lipid-permeable dimerizing drug leads to cross-linking and activation of a prostate-targeted isoform of FGFR-1 that carries cytoplasmic tyrosine kinase domains linked to tandem dimerizer drug-binding domains but lacks any extracellular native ligand-binding capacity. Signaling ensues within minutes of dimerizer delivery and proliferation is detectable within 24 hours. Hyperplasia is evident within 2 weeks of biweekly dimerizer administration; by 6 to 8 weeks, the ducts are filled with dysplastic cells, and by 24 weeks high-grade prostate intraepithelial neoplasia is widespread. Further, within 40 weeks of treatment, adenocarcinoma is reproducibly seen.3 Despite a large increase in cellular content, we did not detect a concomitant increase in reporter activity in bigenic mice treated with dimerizer drug for up to 40 weeks (Fig. 5A and B
). Histologic analysis showed a reduction in luciferase expression per cell following FGFR1-mediated hyperplasia and dysplasia (Supplementary Fig. S2). Ex vivo chemiluminescent imaging of 60-week-old EZC2-JOCK prostates (from above) also shows that AR activity is diminished per cell during FGFR1-stimulated progression, which increases total prostate volume, leading to an overall 5-fold decrease in reporter activity ex vivo (Fig. 5C).
|
|
|
20-fold greater per milligram of tissue than in the intestines (Fig. 2), reporter activity in distant metastasis in EZC1-TRAMP mice is actually weaker than that of intestinal tissue. However, lowering the pseudocolor threshold reveals reporter activity in the easily visible hypochromic metastatic tumors, consistent with prostate epithelial origin, and this bioluminescence follows the outline of the gross metastasis (Fig. 6E). Imaging tissues ex vivo can also reveal metastasis within tissues, such as in the enlarged spleen of this animal despite the absence of metastatic nodules on the surface. Therefore, although imaging of prostate metastasis in living mice may be challenging when AR-regulated promoters are used to express luciferase activity, prostate metastasis in distant organs can be easily detected ex vivo. As opposed to EZC1 and, to a lesser extent, EZC2 mice, the near absence of detectable extraprostatic reporter activity in the abdominal region of EZC3 mice led to the prediction that distant metastasis should be easily detectable in the EZC3-TRAMP model. Similar to EZC1- and EZC2-TRAMP, reporter activity from the lower abdomen of EZC3-TRAMP dropped with age in 40% (4 of 10) of animals imaged fortnightly until euthanasia [due to age (up to 32 weeks), distress, or prohibitive tumor size], consistent with prostate cancer progression (Fig. 7A and Supplementary Fig. S3A, C, E, and F). Moreover, in 50% of mice analyzed until euthanasia, we witnessed a large increase in disseminated reporter activity in living mice (Fig. 7A and Supplementary Fig. S3A, B, C, E, and F). Upon autopsy, we observed reporter activity in numerous tissues, including pancreas (10 of 10) para-aortic (prostate draining) and mesenteric lymph nodes (6 of 10), kidneys (3 of 10), adrenal glands (2 of 10), liver (2 of 10), spleen (2 of 10), and salivary glands (1 of 10). Histologic analysis confirmed that putative tumors imaged in living mice were present in suspect organs (Fig. 7C-F). Thus, the EZC3-prostate model can be used to identify prostate metastasis in living mice and to localize metastasis to specific organ sites ex vivo.
In vivo imaging of AR activity. The three EZC-prostate models were based on highly AR-responsive promoters, leading to the prediction that they should permit remote determination of AR activity in living mice. To test this hypothesis, we compared in vivo reporter activity between castrated and intact mice. Circulating DHT levels in intact mice varied widely (from 50 to 3,000 pg/mL), but dropped below detection (<5 pg/mL) in all castrated mice (data not shown). Although there was some variability in signal intensity in intact mice, likely due to these large variations in circulating androgen levels in cohabitating male animals, the luminescence of castrate mice were quite reproducible and dropped rapidly (3-4 days) to 10% of total intact levels in living EZC2-prostate mice (Fig. 8A and C
). Ectopic delivery of testosterone (10 mg/pellet/21 days) led to recovery of intact reporter levels over a 3-week period. The lengthened time required for reporter rebound versus suppression was likely due to the additional time needed for restoring prostate cellularity. Addition of testosterone pellets to intact mice led to only a slight increase in average reporter levels (Fig. 8A), likely reflecting intact homeostatic regulation of circulating testosterone levels in intact mice. Thus, EZC-prostate mice permit high-throughput appraisal of AR activity in mouse prostates, which could be used to facilitate the characterization and development of drugs that target the AR axis.
|
10% of untreated mice (Fig. 8B and C). As above, addition of testosterone during GnRH blockade led to reporter rebound, consistent with the mechanism for GnRH antagonist function, which targets leutenizing hormone production in the pituitary gland. Thus, the EZC2-prostate model should provide a convenient model for testing some classes of new drugs targeting the AR axis. Because residual androgen-independent signaling in some tissues could, in principle, limit the total drop in reporter activity, we analyzed tissues ex vivo in intact and castrated mice (Fig. 8D). Surprisingly, the drop in reporter activity in prostates following castration was >1,000-fold, much larger than the measured change in intact mice when chemiluminescence over the entire lower abdomen was integrated. As described above, weak reporter activity was also detected in intestines, which was not subject to the effects of castration. Thus, narrowing the region-of-interest for measuring reporter activity during analysis to just over the prostate should further improve the sensitivity of the EZC2-prostate model for remote detection of AR activity.
In EZC3-prostate mice, castrate levels of total lower abdomen reporter activity in living mice were 3% of intact levels (Fig. 9A and B
), suggesting that this model may provide even greater AR sensitivity than EZ2 mice. Interestingly, the drop in reporter activity from prostate tissue following castration was not as great as in EZC2-prostate mice (65-fold versus 2,500-fold), but the near absence of intestinal reporter activity apparently more than compensates for residual prostate-localized signal (Fig. 9C). Peripheral signals in feet and jaw are well removed from prostate-derived signals. Thus, the EZC3-prostate model should be ideal for testing the efficacy of some classes of AR-targeted drugs.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Other methodologies are also unlikely to have the throughput of optical imaging, as in micro-PET, micro–computed tomography, micro–single-photon emission computed tomography, micro–magnetic resonance imaging, and ultrasound, animals are scanned and imaged individually by skilled personnel and large computer files are required to reconstruct and quantify images in three dimensions; however, continuous improvements in computer power will obviate this concern. Similarly, however, it is likely that reengineering of optical imaging stations to accommodate multiangle imaging can greatly improve linearity and dimensionality, coincident with engineering of luciferase and fluorescent reporters with favorable red-shifted emission profiles that better penetrate mammalian tissue. Therefore, despite the allure of optical imaging, the ultimate choice of imaging technologies is application and budget driven.
Although inevitable improvements in imaging will further expand the use of the EZC-prostate models described here, there are already several current applications of these prostate reporter mice. First, we were able to successfully compare three distinct (12–14), robust prostate-specific promoters in an unbiased in vivo setting. Due to the strength of all three AR-responsive composite promoters used in this study, we were able to easily image mice in 10 to 30 seconds. Weaker tissue-specific promoters can require additional manipulations, such as two-step transcriptional amplification (29), to amplify reporter activity to convenient levels for imaging; however, prostate-directed luciferase activity based on two-step transcriptional amplification technology seems to have much lower tissue specificity than the EZC-prostate models (30). In the composite promoters used for the EZC models, multimerization of AREs was largely responsible for magnifying the transcriptional rates while maintaining extremely high (>99%) tissue specificity; however, each line displayed unique characteristics. This comparison indicated that the composite PSA-E2/P and ARR2PB promoters were probably somewhat (2-fold) more potent and prostate specific than hK2-E3/P, whereas hK2-E3/P seemed to be less susceptible to position-effect variegation (i.e., qualitatively similar expression patterns in six of six founders). All promoters were highly AR dependent with dramatic drops in prostate reporter expression up to 2 or 3 orders of magnitude in the EZC3-prostate (PSA) and EZC2-prostate (ARR2PB) models, respectively. At the same time, low-level AR-independent expression in the intestines of EZC1-prostate (hK2) and EZC2-prostate mice make the EZC3-prostate model likely better for appraising AR activity. Although qualitatively similar data is likely to occur in gene therapy vectors (12, 27), adjacent promoter elements, such as those found in viral vectors, can modulate tissue specificity (31).
We were also able to detect relatively small metastasis in TRAMP mice when bred onto the EZC3-prostate background, facilitated by the extremely low background expression outside of the urogenital tract. Low-level signals from the intestines of EZC1- and EZC2-prostate mice make detection of distant metastasis more challenging in these models as stochastic movements of the gastrointestinal tract to surface proximal and distal positions can reduce signal-to-noise ratio. Although all three promoters may be useful for prostate epithelial-specific expression of most transgenes, the extreme sensitivity of BLI can cause confounding signals from even minute levels of luciferase that are orders of magnitude lower than in the prostate, especially if the nontargeted expression emanates from surface-proximal or relatively large tissues. Thus, the unusually high prostate specificity of the EZC3-prostate model makes it ideal for identifying early metastasis in genetically engineered mice predisposed to prostate cancer.
Background bioluminescence is much less of a factor when applying BLI to the tracking of adoptively transferred luciferase-transfected cells. In numerous reports, tumor cell growth and metastasis were accurately measured and identified with BLI, although in vivo imaging was less sensitive than ex vivo imaging as expected (32). Further, response of tumor cells to traditional therapies or immunotherapy has been accurately monitored in vivo as well as the tracking of leukocytes to neoplastic tissue (33). Despite the extra challenges of transgenic models, a number of reports indicate their broad use (reviewed in ref. 8).
Due to both the exquisite AR sensitivity of both the three composite promoters used in this study and the involution of the prostate that occurs following medical or chemical castration, the EZC-prostate models are also useful for monitoring AR activity in the prostate in living mice. As proof-of-principle, we were able to show a drop in reporter activity up to 2 orders of magnitude following castration in living EZC2- and EZC3-prostate mice. Similar results were achieved when GnRH antagonist, PPI-258, was added to EZC2-prostate mice. This model should permit comparisons of not only distinct reagents but also of dosing schedules and routes of injection. Of course, there are always concerns about the applicability of animal studies to humans, but due to high conservation of the AR axis, these mice should be appropriate for preclinical evaluation of other classes of AR-blocking compounds. The high magnitude of the inhibition following castration was somewhat surprising as adrenal-derived androgens comprise up to 5% of circulating testosterone and have been shown to activate the AR (reviewed in ref. 25). Nevertheless, the combination of prostate involution and promoter inhibition following castration or GnRH antagonist administration almost completely inhibited prostate-derived luciferase in all three models. It will be interesting to evaluate other classes of drugs, such as 5-reductase inhibitors, nonsteroidal antiandrogens, and selective AR modulators using these mice. Similarly, these models should be useful for elucidating the potential roles of numerous transcriptional coactivators associated with AR signaling in normal or prostate cancer tissue (34, 35).
The recent development of genetically engineered mouse modeling prostate cancer has stimulated our understanding of disease progression and has been a platform for treatment development. Further developments in imaging technologies will synergize with that work. These new EZC-prostate models should not only expand the use of transgenic prostate cancer mouse models, but will also be useful for drug and treatment development and a better understanding of the role(s) of the AR axis during prostate development and disease progression.
|
Acknowledgments |
|---|
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 K. Slawin (Baylor College of Medicine) for helpful discussions, J.J. Armstrong, E. Nikitina, and R. Gangula (Baylor College of Medicine) for technical assistance, and W. Westlin (Praecis Pharma, Waltham, MA) for providing PPI-258-CMC.
|
Footnotes |
|---|
Current address for X. Xie: University of Texas M.D. Anderson Cancer Center, Houston, TX.
3 D.M. Spencer, E. Nikitina, and V. Acevedo, unpublished results. 
Received 11/ 3/05. Revised 3/13/06. Accepted 4/11/06.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S Kerkhofs, S Denayer, A Haelens, and F Claessens Androgen receptor knockout and knock-in mouse models J. Mol. Endocrinol., January 1, 2009; 42(1): 11 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Ellwood-Yen, J. Wongvipat, and C. Sawyers Transgenic Mouse Model for Rapid Pharmacodynamic Evaluation of Antiandrogens Cancer Res., November 1, 2006; 66(21): 10513 - 10516. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
) or castrate (day 0, 
) EZC2-prostate mice (n = 10) was measured biweekly. Testosterone (10 mg/21 days) tablets were administered s.c. at day 16 in intact (
) or castrate (
) animals (n = 5). Points, mean; bars, SD. (Error bars were removed from intact mice for clarity). B, total luminescence of intact mice (n = 10) treated at day 11 with GnRH antagonist s.c. (
, n = 5). C, in vivo imaging of luciferase activity in two representative intact, castrate (day 15 following castration) or GnRH antagonist-treated EZC2-prostate mice (day 11 of GnRH treatment). D, ex vivo imaging of liver, intestines (cecum and colon only), and prostates of intact and castrate [5 weeks (first two), 11 weeks (last two)] mice. Average ± SD of photon flux (p/s/cm2/sr) is shown to the right. Note that prostate images were measured separately to avoid reflected light and are shown with a 10-fold increased pseudocolor scale. *, P < 0.05.