Real-time analysis of gene expression in experimental tumor models represents a major tool to document disease biology and evaluate disease treatment. However, monitoring gene regulation in vivo still is an emerging field, and thus far it has not been linked to long-term tumor growth and disease outcome. In this report, we describe the development and validation of a fluorescence-based gene expression model driven by the promoter of the cyclin-dependent kinase inhibitor p21waf1,cip1. The latter is a key regulator of tumor cell proliferation and a major determinant in the response to many anticancer agents such as histone deacetylase inhibitors. In response to histone deacetylase inhibitors, induction of fluorescence in A2780 ovarian tumors could be monitored in living mice in a noninvasive real-time manner using whole-body imaging. Single p.o. administration of the histone deacetylase inhibitor MS-275 significantly induces tumor fluorescence in a time- and dose-dependent manner, which accurately predicted long-term antitumoral efficacy in individual mice following extended treatment. These findings illustrate that this technology allows monitoring of the biological response induced by treatment with histone deacetylase inhibitors. In addition to providing experimental pharmacokinetic/pharmacodynamic markers for investigational drugs, this model provides insight into the kinetics of in vivo regulation of transcription, which plays a key role in causing and maintaining the uncontrolled proliferation of tumor tissue. [Mol Cancer Ther 2006;5(9):2317–24]

Transcriptional regulators are frequently altered in tumors and play a causal role in maintaining malignant phenotype (1). A well-known example is the tumor suppressor p53, which controls cellular response to radiation and chemotherapeutic agents and is inactivated in the majority of human tumors (2, 3). One of the main target genes of p53 is the cyclin-dependent kinase inhibitor p21waf1,cip1, which is crucial for G1 checkpoint control and for maintaining the G2 checkpoint in response to DNA damage (47). In addition, p21waf1,cip1 protein plays an important role in cell senescence and differentiation (8). p21waf1,cip1 is highly regulated at the transcriptional level and its expression is induced by a broad panel of upstream signaling pathways. p53 activates p21waf1,cip1 transcription through direct binding to Sp1, a constitutive transcriptional activator that binds to the GC-boxes in the proximal region of the p21waf1,cip1 promoter. p53 binding results in the dissociation of histone deacetylase 1 (HDAC1), which is responsible for the constitutive suppression of the p21waf1,cip1 gene (9). Histone deacetylase–induced transcriptional silencing is a key process in tumor growth (10) and several histone deacetylase inhibitors have shown promising activity in early clinical development (11).

As p21waf1,cip1 is regulated by a large number of signaling pathways and is also a key determinant in the biological outcome after triggering these signaling pathways, an experimental model visualizing p21waf1,cip1 gene expression in a real-time manner would be of value in predicting the long-term antitumor effect of investigational agents. In addition, such a model would allow rapid identification of novel antineoplastic agents in general, a process that is currently hampered by the limited throughput of mouse xenograft tumor studies, requiring long-term studies and the necessity for extensive compound synthesis. For this reason, many promising early compounds are not evaluated in vivo and potentially numerous promising new therapeutic agents are abandoned.

We have developed an experimental tumor model that allows noninvasive real-time analysis of signal transduction and gene expression in nude mice bearing human xenografts using whole-body imaging technology (12). Human A2780 ovarian carcinoma cells have been engineered with a reporter gene construct encoding the fluorescent ZsGreen protein, of which the expression is under the control of the p21waf1,cip1 promoter. When s.c. grafted to athymic nude mice, A2780-p21waf1,cip1ZsGreen ovarian xenografts allow noninvasive real-time imaging of gene expression regulation in living tumor-bearing mice. ZsGreen fluorescence fully parallels regulation of the endogenous p21waf1,cip1 protein, and induction of fluorescence in vivo was found to accurately predict long-term antitumor activity of the HDAC inhibitor MS-275 in individual mice. Induction of the p21waf1,cip1 promoter in vivo was observed in both the center and periphery of the experimental tumors, indicating an exposure and biological response in the entire tumor. Our data therefore show that this model allows the analysis of the mechanism of action of a selected drug in vivo and noninvasive real-time evaluation of the pharmacodynamic response to HDAC inhibitors in animal models.

Materials

Human A2780 ovarian carcinoma cells (American Type Culture Collection, Manassas, VA) were maintained as monolayers in RPMI 1640 supplemented with 10% FCS, 2 mmol/L l-glutamine, and gentamicin at 37°C in a humidified 5% CO2 incubator. All cell culture reagents were purchased from Life Technologies, Inc. (Gaithersburg, MD). Trichostatin A was obtained from Calbiochem (San Diego, CA). Cell proliferation was measured by a classic 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide assay (Serva, Heidelberg, Germany).

Production of the 1,300-kb p21waf1,cip1 Promoter-ZsGreen Construct

To generate an HDAC inhibitor–responsive p21waf1,cip1 promoter construct, genomic DNA was extracted from A2780 cells and used as template for nested PCR isolation of a 1.3-kb fragment containing the −1,300 to +88 region of the p21waf1,cip1 promoter relative to the TATA box. The first amplification was done with the oligonucleotide pair GAGGGCGCGGTGCTTGG and TGCCGCCGCTCTCTCACC. The resulting 4.5-kb fragment was reamplified with the oligos TCGGGTACCGAGGGCGCGGTGCTTGG and ATACTCGAGTGCCGCCGCTCTCTCACC and subsequently with the oligos TCGGGTACCGGTAGATGGGAGCGGATAGACACATC and ATACTCGAGTGCCGCCGCTCTCTCACC. The luciferase reporter was removed from the pGL3-basic plasmid (Promega, Leiden, the Netherlands) and replaced by the ZsGreen reporter (pZsGreen1-N1 plasmid, Clontech, Mountain View, CA) at KpnI and XbaI restriction sites. pGL3-basic-ZsGreen-1300 was constructed via insertion of the above-mentioned 1.3-kb fragment of the human p21waf1,cip1 promoter region into pGL3-basic-ZsGreen at the XhoI and KpnI sites. All restriction enzymes were purchased from Boehringer Mannheim (Mannheim, Germany).

Generation of Stable Cell Lines

A2780 cells were seeded at a density of 2 × 105 per 10-cm diameter cell culture dish, grown for 24 hours, and transfected with 2 μg of pGL3-basic-ZsGreen-1300 and 0.2 μg of pSV2neo resistance vector using Lipofectamine 2000 (Invitrogen, Brussels, Belgium). Single-cell clones were obtained in medium containing G418 (450 μg/mL) and tested for detectable basal level of fluorescence to allow for detection of tumor mass in the absence of stimulation. All further studies were done with A2780-p21waf1,cip1ZsGreen clone 5 (A2780-p21waf1,cip1ZsGreen).

Fluorescence Detection of p21waf1,cip1 Promoter Activity in Cell-Based Assays

A2780-p21waf1,cip1ZsGreen cells were seeded at 10,000 per well into 96-well multititer plates, grown for 24 hours, and treated for an additional 24 hours with the indicated HDAC inhibitors. Subsequently, cells were fixed with 4% paraformaldehyde for 30 minutes and counterstained with Hoechst dye. p21waf1,cip1 promoter activation leading to ZsGreen protein production (and thus fluorescence signal) was monitored by the Ascent Fluoroskan (Thermo Labsystems, Brussels, Belgium) by fluorescence-activated cell sorting or with a fluorescent microscope [Axiovert 135 (Zeiss, Jena, Germany), equipped with a green fluorescence (excitation, 450–490 nm; emission, 510 nm) filter set, supported by MagnaFire software (Optronics, Goleta, CA)]. Ascent Fluoroskan data are presented as mean ± variation of two independent experiments. IC50 values were calculated by nonlinear regression analysis using SigmaPlot 4.01 software. For fluorescence-activated cell sorting analysis of fluorescence, A2780-p21waf1,cip1ZsGreen cells were seeded at a density of 3 × 105 per well in a six-well cell culture plate. Cells were treated with DMSO or 10−7 mol/L trichostatin A. After 24 hours, cells were collected and resuspended in 0.5-mL CellScrub Buffer (Gene Therapy Systems, Inc., San Diego, CA) at 4°C. Cells were subsequently analyzed on a FACScan (BD Biosciences, Erembodegem, Belgium) for ZsGreen fluorescence (detection filter set at 530 nm). Cell aggregates were filtered out and 10,000 events were analyzed. The effect of HDAC inhibitors on endogenous p21waf1,cip1 protein expression was measured with a WAF1 ELISA (Oncogene, Cambridge, MA) according to the prescription of the manufacturer. Data are presented as mean ± variation of two independent experiments.

In vivo Evaluation of Fluorescence Induction

Athymic male NMRI nu/nu mice were purchased from Janvier (Le Genest St.-Isle, France) and were treated according to the ethical guidelines prescribed by United Kingdom Coordinating Committee on Cancer Research. The A2780-p21waf1,cip1ZsGreen cells were injected s.c. (107/200 μL) into the inguinal region of nude mice and caliper-measurable tumors were obtained after 12 days. Subcutaneous. fluorescent tumors could be detected through the skin of nude mice and measured using an automated whole-body imaging system (12). From day 12, mice were treated p.o. once daily for 1, 2, 3, or 4 days with solvent or the indicated dose of MS-275 (10 animals per group). Tumors were evaluated for fluorescence by the automated whole-body imaging system [fluorescent stereomicroscope type Olympus SZX12 equipped with a green fluorescence filter and coupled to a CCD camera type JAI CV-M90 controlled by a software package based on the IMAQ Vision Software from National Instruments (Austin, TX)]. Median and average value of tumor fluorescence were calculated by locating the tumor by creating a gray-level image followed by calculating a binary image. The light intensity of all pixels within this area was normalized between 0 (black) and 1 (white). Median light intensity of the tumor was normalized for tumor area.

Fluorescence Microscopy

To evaluate the distribution of ZsGreen protein expression, A2780-p21waf1,cip1ZsGreen tumors were excised, mounted in Peel-A-Way molds (Polysciences, Warrington, PA) with optimum cutting temperature (Sakura, Torrance, CA), and then frozen in liquid nitrogen. Cryosections of 10-μm thickness were mounted on glass slides and air-dried for 30 minutes at room temperature. Within 1 hour after drying, ZsGreen fluorescence was evaluated with an Axioplan 2 (Zeiss) equipped with Epiplan-Neofluar objectives and an Axiocam camera. Images taken with the Axiocam camera from peripheral and central areas of tumor sections from control and treated groups were stored as TIF files and imported in LSM-510 software (Zeiss). To enhance visualization of the distribution of the ZsGreen protein, the imported TIF files were viewed as pseudo three-dimensional pixel intensity maps. For three-dimensional colocalization with tumor vasculature, tumors were collected by transcardial perfusion fixation with 4% paraformaldehyde. Actin and endothelium of blood vessels were visualized by Bodipy 558/568 phalloidin staining and CD-31 (Cy3) immunofluorescent labeling, respectively. Mounted samples were observed with the LSM510 laser scanning microscope.

In vivo Antitumoral Studies

MS-275 was synthesized in-house and formulated at 1 mg/mL in 20% hydroxypropyl-β-cyclodextrin (final pH 8.5) as an injectable solution. A2780-p21waf1,cip1ZsGreen cells were injected s.c. (107/200 μL) into the inguinal region of nude mice (athymic male NMRI nu/nu mice; Janvier). From day 4, mice were dosed p.o. daily during 35 days (q.d. × 35, p.o.) with MS-275 (10 animals per group, 0.5 mL/mouse). Tumor size was measured with a caliper. Tumor volume was determined by the formula (a2 × b) / 2, in which a represents the width and b the length.

p21waf1,cip1ZsGreen Reporter Gene Model Predicts the Biological Effect of Histone Deacetylase Inhibitors in Human A2780 Ovarian Cancer Cells

To establish a model enabling real-time visualization of p21waf1,cip1 gene transcription, human A2780 ovarian carcinoma cells were stably transfected with a p21waf1,cip1 promoter fragment (−1,300 to +88 bp) regulating the expression of ZsGreen fluorescent protein (see Materials and Methods). This promoter region contains key cis-acting elements that are required for p21waf1,cip1 regulation, including Sp1 sites that mediate its activation by HDAC inhibitors (13, 14).

To validate this engineered cell model, ZsGreen protein expression, and therefore fluorescence signal, needs to reflect the regulation of endogenous p21waf1,cip1 in cell-based assays. To study p21waf1,cip1 transcriptional regulation, we employed HDAC inhibition, which results in derepression of the p21waf1,cip1 promoter, thereby causing cycle arrest (15). As illustrated in Fig. 1, the majority of the A2780-p21waf1,cip1ZsGreen cell population has a low basal fluorescence (i.e., below 102 relative fluorescence units). The HDAC inhibitor trichostatin A strongly increased median fluorescence to >103 relative fluorescence units. To determine to what extent induction of fluorescence predicts the antiproliferative activity, a number of HDAC inhibitors that are currently in phase I and II clinical trials were evaluated [i.e., MS-275 (Berlex/Schering, Berlin, Germany), suberoylanilide hydroxamic acid (SAHA; Merck, Darmstadt, Germany), and LAQ-824 (Novartis, Basel, Switzerland)]. A2780-p21waf1,cip1ZsGreen ovarian carcinoma cells were treated with increasing concentrations of HDAC inhibitors as indicated in Materials and Methods. As shown in Fig. 2A, SAHA and MS-275 induced fluorescence in A2780-p21waf1,cip1ZsGreen cells by 3.6- and 4.3-fold, respectively, at concentrations ranging from 3 to 10 μmol/L, whereas LAQ-824 and trichostatin A were more potent and induced fluorescence 5.5- and 5.2-fold at 0.3 and 1 μmol/L, respectively. This closely reflected the endogenous p21waf1,cip1 protein induction by these agents in A2780-p21waf1,cip1ZsGreen cells (Fig. 2B). Again, LAQ-824 and trichostatin A were most potent at 0.3 and 1 μmol/L, respectively, and induced p21waf1,cip1 protein levels to a comparable extent as the observed increase in ZsGreen fluorescence (5.5- and 5.7-fold, respectively). SAHA and MS-275 maximally induced endogenous p21waf1,cip1 protein expression at 10 μmol/L (4.1- and 6.1-fold, respectively). Similar induction of endogenous p21waf1,cip1 protein expression was observed in parental nontransfected A2780 cells (data not shown). The decrease in fluorescence and p21waf1,cip1 protein levels at higher doses was due to cell death. The potency of these agents vis-à-vis p21waf1,cip1 induction was paralleled by their antiproliferative activity as illustrated in Fig. 2C. Cell proliferation after 4 days was inhibited with IC50 values of 4.2 and 3.4 μmol/L for SAHA and MS-275, respectively, which is a similar potency at which an induction of p21waf1,cip1 and ZsGreen protein expression was observed after 24 hours. LAQ-824 and trichostatin A inhibited cell proliferation at 0.15 and 0.21 μmol/L (IC50), respectively, whereas ZsGreen protein was induced at 0.3 and 1 μmol/L, respectively. Summarized, these data reveal that fluorescence induction in A2780-p21waf1,cip1ZsGreen cells predicts the antiproliferative potency of HDAC inhibitors in cell-based assays.

Figure 1.

Induction of fluorescence in A2780-p21waf,cip1ZsGreen cells treated by trichostatin A. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were incubated for 24 h with 100 nmol/L of trichostatin A (TSA). ZsGreen protein production was monitored by fluorescence-activated cell sorting analysis. The autofluorescence peak of the parental A2780 cell line was located between 1 and 10 relative fluorescence units. Insets, trichostatin A–induced fluorescence monitored by fluorescent microscope.

Figure 1.

Induction of fluorescence in A2780-p21waf,cip1ZsGreen cells treated by trichostatin A. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were incubated for 24 h with 100 nmol/L of trichostatin A (TSA). ZsGreen protein production was monitored by fluorescence-activated cell sorting analysis. The autofluorescence peak of the parental A2780 cell line was located between 1 and 10 relative fluorescence units. Insets, trichostatin A–induced fluorescence monitored by fluorescent microscope.

Close modal
Figure 2.

Induction of fluorescence in A2780-p21waf,cip1ZsGreen cells predicts biological effect of histone deacetylase inhibitors in vitro. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were treated for 24 h with either the vehicle (0.1% DMSO) or the indicated concentrations of various HDAC inhibitors [SAHA (×), MS-275 (▴), LAQ-824 (▪), and trichostatin A (•)]. A, dose-dependent induction of fluorescence signal. B, dose-dependent induction of p21waf1,cip1 protein expression as measured with a p21 ELISA. C, dose-dependent inhibition of cell proliferation after 4 d of incubation. Points, cell viability expressed as average percent of control for three independent experiments; bars, SD.

Figure 2.

Induction of fluorescence in A2780-p21waf,cip1ZsGreen cells predicts biological effect of histone deacetylase inhibitors in vitro. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were treated for 24 h with either the vehicle (0.1% DMSO) or the indicated concentrations of various HDAC inhibitors [SAHA (×), MS-275 (▴), LAQ-824 (▪), and trichostatin A (•)]. A, dose-dependent induction of fluorescence signal. B, dose-dependent induction of p21waf1,cip1 protein expression as measured with a p21 ELISA. C, dose-dependent inhibition of cell proliferation after 4 d of incubation. Points, cell viability expressed as average percent of control for three independent experiments; bars, SD.

Close modal

HDAC Inhibitors Induce Fluorescence in A2780-p21waf1,cip1ZsGreen Xenografts in Nude Mice

To investigate whether the p21waf1,cip1ZsGreen model could be used to monitor the biological effect of HDAC inhibitors in tumor-bearing mice, studies were done with MS-275. Previous work had shown that MS-275 inhibits A2780-p21waf1,cip1ZsGreen tumor growth when cells were grafted s.c. in immunodeficient mice (data not shown). In those studies, mice were treated p.o. with increasing doses of MS-275 once daily for 21 days starting 4 days after injection of cells. At the maximum tolerated dose (20 mpk), tumor growth was inhibited by 83% whereas treatment at 10 mpk resulted in 58% inhibition. Tumor growth inhibition was statistically significant starting on day 14 at both doses, and on day 24 the inhibition of tumor growth at 20 mpk was significantly higher compared with 10 mpk (P < 0.05). At 5 mpk, tumor inhibition on day 24 was marginal (31% inhibition; P = 0.08, one sided), whereas 2.5 mpk was an inactive dose.

As shown in Fig. 3, MS-275 at biologically active nontoxic doses (i.e., 5, 10, and 20 mpk) clearly induced tumor fluorescence in vivo (Fig. 3A). Time course analysis of this induction using automated whole-body imaging of tumor-bearing living animals revealed that a significant induction of fluorescence in the tumor could be detected already after one single administration of MS-275 (Fig. 3B). Dosing at 5, 10, and 20 mpk resulted in a significant increase (P < 0.05) in median fluorescence from 0.25 to 0.57, 0.63, and 0.75 absorbance units, respectively. Median fluorescence intensity did not alter at the 5- and 10-mpk dose when treating the mice for 2 or 3 subsequent days. In contrast, at the most active dose (20 mpk), a strong increase in the extent of fluorescence was observed when comparing dosing once (0.75 absorbance unit) with dosing for 2 and 3 days (Fig. 3C, 1.01 and 2.44 absorbance units, respectively). The observed increased p21waf1,cip1 promoter activity, when comparing dosing once to dosing for 3 consecutive days, was statistically significant with the Wilcoxon-Mann-Whitney test (P < 0.05). These data illustrate that whole-body imaging allows the monitoring of biological responses in the tumor as a function of time. Because a significant response was observed at biologically active (antitumor) doses only (i.e., not at 2.5 mpk; Fig. 3B) and the extent of the response paralleled the known antitumor efficacy of MS-275, this suggested that the p21waf1,cip1ZsGreen tumor model can be used to evaluate HDAC inhibitor activity in vivo in a short time frame (a few days compared with a few weeks for classic antitumor studies).

Figure 3.

HDAC inhibitor MS-275 induces fluorescence in A2780-p21waf1,cip1ZsGreen xenografts in nude mice. Human A2780-p21waf1,cip1ZsGreen ovarian tumors cells were injected s.c. into the inguinal region of male athymic nu/nu CD-1 mice. When palpable tumors were obtained, mice were treated p.o. once daily for 3 d (q.d. × 3, p.o.) with the indicated dose of MS-275 or vehicle (20% hydroxypropyl-β-cyclodextrin). Tumors were evaluated for fluorescence every day as described in Materials and Methods. A, fluorescent signal measured in dissected tumors collected 24 h following the last treatment for five individual mice. B, dose- and time-dependent increase in fluorescence using whole-body imaging analysis for individual tumors (10 mice per group). Columns, median values. C, noninvasive real-time fluorescence using whole-body imaging. Tumors shown are those closest to the median of the group (B).

Figure 3.

HDAC inhibitor MS-275 induces fluorescence in A2780-p21waf1,cip1ZsGreen xenografts in nude mice. Human A2780-p21waf1,cip1ZsGreen ovarian tumors cells were injected s.c. into the inguinal region of male athymic nu/nu CD-1 mice. When palpable tumors were obtained, mice were treated p.o. once daily for 3 d (q.d. × 3, p.o.) with the indicated dose of MS-275 or vehicle (20% hydroxypropyl-β-cyclodextrin). Tumors were evaluated for fluorescence every day as described in Materials and Methods. A, fluorescent signal measured in dissected tumors collected 24 h following the last treatment for five individual mice. B, dose- and time-dependent increase in fluorescence using whole-body imaging analysis for individual tumors (10 mice per group). Columns, median values. C, noninvasive real-time fluorescence using whole-body imaging. Tumors shown are those closest to the median of the group (B).

Close modal

When doing long-term in vivo studies with antitumor agents, a large variation in the biological response of individual mice is generally observed, requiring the use of large numbers of animals. Because significant variability was also found in the short-term induction of p21waf1,cip1-ZsGreen, we sought to investigate whether ZsGreen fluorescence correlated with tumor growth for individual mice. As illustrated in Fig. 4, mice showing high tumor fluorescence also had a relatively low final tumor weight, whereas poor responders (e.g., in the 10- and 15-mpk treatment groups) had higher final tumor weights. In summary, this shows that the p21waf1,cip1-ZsGreen fluorescence induction in tumors predicts the extent of response to the treatment with the HDAC inhibitor MS-275 in individual mice.

Figure 4.

Induction of fluorescence in p21waf,cip1 promoter- ZsGreen tumor model predicts antitumor effect of MS-275 in nude mice. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were injected s.c. into the inguinal region of male athymic nu/nu CD-1 mice, and from day 4, animals were treated p.o. once daily with either vehicle (control group, 20% hydroxypropyl-β-cyclodextrin) or HDAC inhibitor, MS-275, at the indicated doses. Tumor weight and fluorescence of individual dissected tumors were evaluated on day 28 using the automated whole-body imaging system. Linear regression showed a correlation coefficient of r = 0.79.

Figure 4.

Induction of fluorescence in p21waf,cip1 promoter- ZsGreen tumor model predicts antitumor effect of MS-275 in nude mice. Human A2780-p21waf1,cip1ZsGreen ovarian tumor cells were injected s.c. into the inguinal region of male athymic nu/nu CD-1 mice, and from day 4, animals were treated p.o. once daily with either vehicle (control group, 20% hydroxypropyl-β-cyclodextrin) or HDAC inhibitor, MS-275, at the indicated doses. Tumor weight and fluorescence of individual dissected tumors were evaluated on day 28 using the automated whole-body imaging system. Linear regression showed a correlation coefficient of r = 0.79.

Close modal

Localized Response to HDAC Inhibitors within A2780 Tumor Tissue In vivo

In addition to predicting long-term antitumor activity, the p21waf1,cip1ZsGreen model may also be of value to study the distribution of the biological response within the tumor tissue. To evaluate the tissue distribution of the response to MS-275, a more detailed analysis of the fluorescent signal distribution pattern was done using fluorescence analysis of tumor cryosections. A distinct difference in expression of ZsGreen was observed. Fluorescence was much stronger in treated tumors than in controls (Fig. 5). On control samples, only a few areas of high fluorescence intensity were observed (Fig. 5A and C). In treated tumors, ZsGreen-fluorescence was not uniform throughout the tissue. Generally, focal spots of high fluorescence were surrounded by areas with weaker intensity. The areas of these high intensity spots contained from a few cells to hundreds of cells and were found not to colocalize with tumor vasculature. Analysis of sections stained for β-actin showed that ZsGreen-labeling was localized in the cytoplasm of the tumor cells (Supplementary Fig. S6).5

5

Supplementary material for this article is available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).

There was no difference in fluorescence between the peripheral and central parts of the tumor (Fig. 5B and D), indicating exposure of the drug and/or a biological response throughout the entire tumor.

Figure 5.

Localization of biological response to MS-275 in A2780-p21waf1,cip1ZsGreen tumors. Distribution of ZsGreen fluorescent signal in unfixed dry cryosections from control nude mice (A and C) or animals treated with MS-275 at 20 mg/kg (q.d., p.o.) for 4 d (B and D). To reveal the distribution of ZsGreen fluorescence in the tissue, fluorescent images (green) were taken of both the peripheral (A and B) and central (C and D) areas of the tumors and were merged with images taken with reflected light microscopy. Reflected light (light blue) allows detection of cellular borders and interstitial tissue. Top and bottom, pseudo three-dimensional pixel intensity maps of the ZsGreen fluorescent signal from each image.

Figure 5.

Localization of biological response to MS-275 in A2780-p21waf1,cip1ZsGreen tumors. Distribution of ZsGreen fluorescent signal in unfixed dry cryosections from control nude mice (A and C) or animals treated with MS-275 at 20 mg/kg (q.d., p.o.) for 4 d (B and D). To reveal the distribution of ZsGreen fluorescence in the tissue, fluorescent images (green) were taken of both the peripheral (A and B) and central (C and D) areas of the tumors and were merged with images taken with reflected light microscopy. Reflected light (light blue) allows detection of cellular borders and interstitial tissue. Top and bottom, pseudo three-dimensional pixel intensity maps of the ZsGreen fluorescent signal from each image.

Close modal

Technologies that allow real-time molecular imaging of gene expression offer the potential for increasing our understanding of disease biology, evaluating disease treatment in vivo, and visualizing the effects of drugs at the cellular level in living animals (16). This report shows that A2780-p21waf1,cip1ZsGreen ovarian xenografts allow noninvasive real-time assessment of the regulation of gene expression in living tumor-bearing mice. ZsGreen fluorescence fully parallels regulation of the endogenous p21waf1,cip1 protein, and induction of fluorescence in vivo accurately predicted the long-term antitumoral activity of HDAC inhibitors in individual mice. Because p21waf1,cip1 is commonly regulated by anticancer agents, such as azacytidine, taxanes, anthracyclins, and cisplatin, this model is not limited to the evaluation of HDAC inhibitors but has broader application. Finally, induction of p21waf1,cip1 promoter activity in vivo was observed in both the central and peripheral regions of the treated tumors, indicating a biological response in the entire tumor.

Although whole-body imaging of highly fluorescent xenografts is commonly used to monitor tumor growth over time (17), real-time monitoring of gene regulation in vivo in response to pharmaceutical agents represents an emerging field. Noninvasive in vivo imaging has been described for a number of signal transduction processes such as protease activity, protein-protein interactions, and protein degradation (1821). For example, Bremer and Tung (18) described visualization of the inhibition of matrix metalloproteinase protease activity in intact tumor-bearing mice with Prinomastat (AG3340). This involved the delivery of fluorescent probes of which the properties change after being bound or cleaved, such as in the case of fluorescence recovery after photobleaching (22) and imaging using autoquenched near-IR fluorescence (23). In vivo analysis of protein degradation was shown by Luker et al. (20) by monitoring overall proteasome activity by imaging turnover of a labile luciferase protein, whereas Zhang et al. (21) monitored cyclin-dependent kinase-2 activity in vivo by imaging the turnover of a p27-luciferase fusion protein. Recently, Qian et al. (24) described reactivation of the retinoic acid receptor β2 promoter by MS-275 and retinoids in PC3 prostate xenografts using bioluminescence. However, in all these studies, a link was not established between the signal transduction response and final therapeutic outcome.

To monitor real-time gene expression regulation in vivo, fluorescent protein expression was used as a reporter because of its compatibility with unanesthesized animals and because it does not require the delivery of a substrate to the tumor before imaging. This second aspect is important when studying in vivo gene responses to pharmaceutical agents because a substrate may nonspecifically modulate gene expression or might interfere with the pharmacokinetics or pharmacodynamics of the investigational drug (16). This issue is illustrated by the work of Doubrovin et al. (25). The enzyme herpes simplex virus-1 thymidine kinase was expressed under the control of a p53-responsive element, enabling positron emission tomography imaging of gene expression after administration of the thymidine kinase substrate [124I]-2′-fluoro-2′-deoxy-1β-d-arabinofuranosyl-5-iodouracil. Although a significant increase in radioactivity was observed in [N,N′-bis(2-chloroethyl)-N-nitrosourea]–treated rats as compared with control animals, interpretation of these data was hampered by a nonspecific response to [N,N′-bis(2-chloroethyl)-N-nitrosourea] in nontransgene tissue. This was linked to a changed pharmacokinetic profile of the [124I]-2′-fluoro-2′-deoxy-1β-d-arabinofuranosyl-5-iodouracil substrate, presumably due to the nonspecific toxic affects of [N,N′-bis(2-chloroethyl)-N-nitrosourea]. In addition to the risk of changing pharmacokinetic properties when coadministrating both an activator of signal transduction and a positron emission tomography substrate, the radioactive substrates needed for positron emission tomography technology may potentially also induce DNA damage/stress pathways and thereby activate p53 signaling. These data clearly illustrate the importance of nonsubstrate dependent assays when evaluating antitumoral therapeutic agents. These drawbacks were circumvented by engineering the stable A2780 ovarian carcinoma cell line in which the p21waf1,cip1 promoter (−1,300 to +88 bp) regulates the expression of ZsGreen fluorescence.

In this article, we have used HDAC inhibitors to induce the expression of the cyclin-dependent kinase inhibitor p21waf1,cip1. A causal relationship between this p21waf1,cip1 up-regulation and the antiproliferative effects of HDAC inhibitors has previously been established by studies showing that p21waf1,cip1-deficient HCT-116 human colon carcinoma cells have increased resistance to these agents (15, 26). HDAC inhibitors activate p21waf1,cip1 gene expression by dissociating the HDAC repressor complex from the transcriptional activator Sp1, which drives p21waf1,cip1 transcriptional activity. ZsGreen fluorescence was found to fully parallel the regulation of the endogenous p21waf1,cip1 protein by a panel of HDAC inhibitors in vitro. This is in agreement with other reports showing that trichostatin A, MS-275, and SAHA induce transactivation of the murine p21waf1,cip1 promoter through the Sp1 sites located within the −166 to +40 bp region relative to the TATA box (12, 14), a region contained within our p21waf1,cip1 −1,300 to +88 bp promoter construct. In agreement with the key role of p21waf1,cip1 in the antitumor effects of many neoplastic agents, induction of p21waf1,cip1 promoter–driven fluorescence in vivo was found to properly predict the long-term antitumor activity of anticancer agents. Furthermore, kinetic analysis of the fluorescence after several days of dosing indicated accumulation of response after 3 days of dosing for MS-275 at 20 mpk, but not at 10 mpk. This is in agreement with the higher antitumor activity of this agent at 20 mpk. Such a cumulative effect at higher doses may point to accumulation of the compound in tumor tissue over time or changed pharmacokinetic behavior of the agent after repeated dosing. Although the pharmacokinetic profile of MS-275 in mice is not known, early data from clinical trials in human show a long half-life for this agent (29.9 hours) and an increase in exposure at the second dose in four of six patients (27). Interestingly, the fluorescence response was found to be predictive for tumor growth in individual mice. Typically, the response of individual animals to neoplastic agents shows a high variability, which is why large numbers of animals are used for each dose group in this kind of study. Pharmacokinetic analysis showed comparable levels of drug in the tumor tissue between individual animals. Therefore, our data indicate that the high interindividual variation in tumor growth inhibition is dictated by the absence of the biological response to these agents and is not due to lack of activation of HDAC regulated genes.

In summary, our data indicate that the p21waf1,cip1 promoter fluorescent xenografts provide a powerful tool that allows rapid assessment of in vivo antitumoral activity. This facilitates earlier evaluation of the compounds in the drug discovery process without the need for time- and reagent-consuming pharmacokinetic/pharmacodynamic studies, thereby accelerating significantly the discovery and development of novel antitumor agents.

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 Dr. Jorge Vialard for critical reading of the manuscript.

1
Kaelin WG, Jr. Choosing anticancer drug targets in the postgenomic era.
J Clin Invest
1999
;
104
:
1503
–6.
2
Fisher DE. Turning p53 on or off: either way may treat cancer.
Drug Resist Updat
2000
;
3
:
77
–9.
3
Bunz F, Hwang PM, Torrance C, et al. Disruption of p53 in human cancer cells alters the responses to therapeutic agents.
J Clin Invest
1999
;
104
:
263
–9.
4
Bunz F, Dutriaux A, Lengauer C, et al. Requirement for p53 and p21 to sustain G2 arrest after DNA damage.
Science
1998
;
282
:
1497
–501.
5
Deng C, Zhang P, Harper JW, Elledge SJ, Leder P. Mice lacking p21CIP1/WAF1 undergo normal development, but are defective in G1 checkpoint control.
Cell
1995
;
82
:
675
–84.
6
Cheng M, Olivier P, Diehl JA, et al. The p21Cip1 and p27Kip1 CDK “inhibitors” are essential activators of cyclin D-dependent kinases in murine fibroblasts.
EMBO J
1999
;
18
:
1571
–83.
7
Martín-Caballero J, Flores JM, García-Palencia P, Serrano M. Tumor susceptibility of p21Waf1/Cip1-deficient mice.
Cancer Res
2001
;
61
:
6234
–8.
8
Dotto GP. p21WAF1/Cip1: more than a break to the cell cycle?
Biochim Biophys Acta
2000
;
1471
:
M43
–56.
9
Lagger G, Doetzlhofer A, Schuettengruber B. The tumor suppressor p53 and histone deacetylase 1 are antagonistic regulators of the cyclin-dependent kinase inhibitor p21/WAF1/CIP1 gene.
Mol Cell Biol
2003
;
23
:
2669
–79.
10
Cress WD, Seto E. Histone deacetylase, transcriptional control, and cancer.
J Cell Physiol
2000
;
184
:
1
–16.
11
Marks PA, Rifkind RA, Richon VM, Breslow R, Miller T, Kelly WK. Histone deacetylase and cancer: causes and therapies.
Nat Rev
2001
;
1
:
194
–202.
12
Bakker A, Floren W, Voeten J, et al. Automation of whole body imaging of green fluorescent protein-expressing tumors in living animals.
GIT Imaging & Microscopy
2001
;
3
:
52
–4.
13
Xiao H, Hasegawa T, Isobe K. Both Sp1 and Sp3 are responsible for p21waf1 promoter activity induced by histone deacetylase inhibitor in NIH3T3 cells.
J Cell Biochem
1999
;
73
:
291
–302.
14
Huang L, Sowa Y, Sakai T, Pardee AB. Activation of the p21WAF1/CIP1 promoter independent of p53 by the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) through the Sp1 sites.
Oncogene
2000
;
19
:
5712
–9.
15
Archer SY, Meng S, Shei A, Hodin RA. p21WAF1 is required for butyrate-mediated growth inhibition of human colon cancer cells.
Proc Natl Acad Sci U S A
1998
;
95
:
6791
–6.
16
Massoud TF, Gambhir SS. Molecular imaging in living subjects: seeing fundamental biological processes in a new light.
Genes Dev
2003
;
17
:
545
–80.
17
Bouvet M, Wang J, Nardin SR, et al. Real-time optical imaging of primary tumor growth and multiple metastatic events in a pancreatic cancer orthotopic model.
Cancer Res
2002
;
62
:
1534
–40.
18
Bremer C, Tung C-H, Weissleder R. In vivo molecular target assessment of matrix metalloproteinase inhibition.
Nat Med
2001
;
7
:
743
–8.
19
Paulmurugan R, Massoud TF, Huang J, Gambhir SS. Molecular imaging of drug-modulated protein-protein interactions in living subjects.
Cancer Res
2004
;
64
:
2113
–9.
20
Luker GD, Pica CM, Song J, Luker KE, Piwnica-Worms D. Imaging 26S proteasome activity and inhibition in living mice.
Nat Med
2003
;
9
:
969
–73.
21
Zhang G-J, Safran M, Wei W, et al. Bioluminescent imaging of cdk2 inhibition in vivo.
Nat Med
2004
;
10
:
643
–8.
22
Berk DA, Yuan F, Leunig M, Jain RK. Direct in vivo measurement of targeted binding in a human tumor xenograft.
Proc Natl Acad Sci U S A
1997
;
94
:
1785
–90.
23
Weissleder R, Tung C-H, Mahmood U, Bodganov A, Jr. In vivo imaging of tumors with protease-activated near-infrared fluorescent probes.
Nat Biotechnol
1999
;
17
:
375
–8.
24
Qian DZ, Ren M, Wei Y, et al. In vivo imaging of retinoic acid receptor β2 transcriptional activation by the histone deacetylase inhibitor MS-275 in retinoid-resistant prostate cancer cells.
Prostate
2005
;
64
:
20
–8.
25
Doubrovin M, Ponomarev V, Beresten T, et al. Imaging transcriptional regulation of p53-dependent genes with positron emission tomography in vivo.
Proc Natl Acad Sci U S A
2001
;
98
:
9300
–5.
26
Kim YB, Ki SW, Yoshida M, Horinouchi S. Mechanism of cell cycle arrest caused by histone deacetylase inhibitors in human carcinoma cells.
J Antibiot
2000
;
53
:
1191
–200.
27
Gojo I, Karp JE, Mann D, et al. Phase I study of histone deactylase inhibitor (HDI) MS-275 in adults with refractory or relapsed hematologic malignancies [abstract no 2198]. Proc Am Soc Hem 2002.