Nature Medicine10, 1257 - 1260 (2004)
Published online: 24 October 2004; | doi:10.1038/nm1120
Dissecting tumor maintenance requirements using bioluminescence imaging of cell proliferation in a mouse glioma model
Lene Uhrbom1, 2, 3, Edward Nerio2, 3
& Eric C Holland2
1 Uppsala University, Department of Genetics and Pathology, Rudbeck Laboratory, SE-75185 Uppsala, Sweden.
2 Memorial Sloan-Kettering Cancer Center, Departments of Surgery (Neurosurgery), Neurology and Cancer Biology and Genetics, RRL 917B, 1275 York Avenue, New York, New York 10021, USA.
Bioluminescence imaging has previously been used to monitor the formation of grafted tumors in vivo and measure cell number during tumor progression and response to therapy. The development and optimization of successful cancer therapy strategies may well require detailed and specific assessment of biological processes in response to mechanistic intervention. Here, we use bioluminescence imaging to monitor the cell cycle in a genetically engineered, histologically accurate model of glioma in vivo. In these platelet-derived growth factor (PDGF)-driven oligodendrogliomas, G1 cell-cycle arrest is generated by blockade of either the PDGF receptor or mTOR using small-molecule inhibitors.
One factor limiting the high-throughput use of mouse models for preclinical trials is difficulty in performing longitudinal studies analyzing tumor-cell response. One of the most common alterations found in human cancer is inactivation of the RB pathway1. The mutations leading to this inactivation result in loss of RB-E2F transcriptional repressor complexes and activation of E2F-regulated transcription2. Here we present a new transgenic mouse for bioluminescence imaging (BLI) of the proliferative activity of glioma cells. In this mouse line, the gene encoding luciferase is controlled by the human E2F1 promoter, which exhibits tumor-specific activity in vivo3. We show that the activity of the RB pathway in endogenously induced brain tumor cells can be detected, followed over time and used to evaluate the efficacy of drug treatment in vivo. We use this technology to show that the proliferation of PDGF-induced gliomas is dependent on both PDGF receptor activation and mTOR signaling.
Results The Ef-luc transgenic construct Recently, use of in vivo BLI has become an established method4,
5,
6 that is more sensitive than other in vivo imaging methods presently available7,
8. To image the loss of RB pathway function in tumor cells in vivo we used the human E2F1 promoter controlling the expression of the firefly luciferase gene, the Ef-luc transgenic construct (see Supplementary Fig. 1a online). The E2F1 promoter is autoregulatory, containing four E2F binding sites. In vitro, the E2F1 promoter has been shown to have a strict cell cycle−regulated expression with elevated activity during late G1 to S phase2 resulting from E2F-mediated negative regulation where E2F-RB inhibitory complexes bind to the promoter during early G1. In vivo, however, the E2F1 promoter has been shown to mediate tumor-selective transgene expression unrelated to normal mitotic activity3, probably as a result of loss of RB pathway control in the tumor cells. Thus, bioluminescence from the Ef-luc transgenic mouse is expected to correlate not only with the number of tumor cells but also with the loss of RB pathway necessary for transformation activity and the proliferative capacity of tumor cells. We have developed this reporter line for use in mouse glioma models and report findings with this tumor type.
Bioluminescence imaging of PDGF-induced gliomagenesis To demonstrate that the Ef-luc mouse could be used for BLI of tumors, we crossbred the Ef-luc transgenic mouse line with the N−tv-a mouse strain allowing BLI in the well-characterized RCAS/TV-A mouse model system of brain tumors9 (Supplementary Fig. 1b online). The N−tv-a transgenic mouse expresses the viral receptor tv-a from the nestin promoter, allowing retroviral transduction to glial progenitor cells using replication-competent ASLV long terminal repeat with a splice acceptor (RCAS) vectors. The double transgenic Ef-luc N−tv-a mice were infected with the avian leukosis virus−based RCAS-PDGFB vector that induces oligodendrogliomas in N−tv-a mice10. First, we determined whether BLI could be used to detect brain tumors induced with the RCAS/TV-A model system. Mice injected with RCAS-PDGFB were imaged by BLI weekly. We found that all mice with the Ef-luc transgene showed a background light production, especially over the regions of skin not covered by fur (Supplementary Fig. 2 online). Mice found to emit additional light between the ears (Supplementary Fig. 2 online and Fig. 1a) were analyzed and the emitted light could be quantified using Living Image software. We confirmed the presence of gliomas in these mice by histologic analysis of the brains following imaging (Fig. 1b). For tumors of the same proliferation index, the amount of light production roughly correlated with the size of the tumor (Fig. 1).
Figure 1. Approximate correlation between BLI output and tumor size.
(a) Ef-luc BLI of PDGF-induced oligodendrogliomas. Luciferase activity in tumor-bearing Ef-luc N−tv-a transgenic mice. (b) Whole mount histologic analysis of the brains from the same mice as were imaged in a. Tumor size roughly correlates with the amount of emitted light.
Development of gliomas over time Next we performed a longitudinal study by screening Ef-luc N−tv-a mice that had been injected with RCAS-PDGFB every third day with BLI in order to follow tumor development. A representative graph of one such mouse that developed a brain tumor is shown (Fig. 2a). The time-dependent increase in light production represents the sum of the tumor cells' capacity to proliferate and the overall size of the tumor. In a subsequent study we imaged five mice daily that had been identified to harbor gliomas. These mice showed a variable increase in light production, indicating the baseline for glioma growth in this model (Fig. 2b).
Figure 2. Longitudinal studies of PDGFB-induced gliomagenesis in Ef-luc N−tv-a transgenic mice.
(a) One Ef-luc N−tv-a mouse imaged every third day for 39 d. (b) Five Ef-luc N−tv-a mice imaged daily for 5 d.
Bioluminescence imaging of therapeutic efficacy We then used the model in two longitudinal preclinical studies aimed at determining the requirements for maintenance of tumor-cell proliferation in this PDGF-driven glioma model. First, we used BLI to follow PDGFB-induced glioma-bearing mice that were treated with an inhibitor of the PDGF receptor, PTK787/ZK222584 (ref. 10). The result of this analysis for one representative mouse showing substantial decrease in light production over time is shown in Figure 3a. We compared five glioma-bearing mice treated with PTK787/ZK222584 to mice treated with buffer only. We imaged these two cohorts of mice daily and found that, whereas the buffer-treated mice showed a continued increase in photon emission from the brain area over 5 d (Fig. 3b), the mice treated with PTK787/ZK222584 showed a clear reduction in light emission (Fig. 3b). All of these mice were killed and their brains analyzed for evidence of proliferation using immunohistochemical staining for PCNA, Ki-67 and phosphohistone H3. Consistent with our previous data11, the PTK787/ZK222584-treated tumors showed minimal staining for these markers (less than 5% of tumor cells) (Fig. 3b) whereas the buffer-treated (and untreated) tumors showed a high level of PCNA staining (greater than 50% of tumor cells) (Fig. 3b). The Ki-67 and phosphohistone H3 immunostainings were consistent with the PCNA results (data not shown). In all cases there was minimal evidence of apoptosis, as determined by TUNEL and caspase 3 staining (data not shown), also consistent with our earlier findings11. These data indicate that the light production from these tumors in an Ef-luc N−tv-a transgenic background correlates with other indicators of cell cycle progression and proliferation.
Figure 3. Preclinical trials of PDGFB-induced glioma-bearing Ef-luc N−tv-a mice.
(a) Longitudinal imaging of one tumor-bearing mouse treated with PTK787/ZK222584 daily for 6 d. (b) Longitudinal study with five Ef-luc N−tv-a tumor-bearing mice in each cohort: untreated (left panel) or treated daily with PTK787/ZK222584 (middle panel) or CCI-779 (right panel). Upper panels show photon counts of emitted light and lower panels show immunohistochemical staining for PCNA as a measure of cell proliferation.
In the second preclinical study, we tested the effect of mTOR inhibition by using the rapamycin analog CCI-779 (ref. 12) in the PDGF-induced gliomas. We treated 5 tumor-bearing mice with CCI-779 and imaged them over 5 d. Inhibition of mTOR by treatment with CCI-779 reduced the light production from PDGF-induced gliomas to an extent similar to that seen with PDGF receptor inhibition (Fig. 3b). Furthermore, when we analyzed the brains from these mice with immunohistochemical markers for proliferation, the results showed essentially the same effect when inhibiting mTOR as seen for PDGF receptor inhibition (Fig. 3b and data not shown). Notably, there was minimal staining for TUNEL in these sections, indicating that blockade of mTOR does not induce apoptosis in this glioma model. Taken together, these data imply that the activities of both PDGF receptors and mTOR are essential for the proliferation of PDGF-induced gliomas in mice.
Discussion Experimental data generated in vitro does not always correlate with what is seen in vivo. Preclinical trials need to be performed in order to understand the rules that govern the maintenance of tumor cells in a living organism in vivo. Previous reports have described BLI of xenograft gliomas in which luciferase was driven by a constitutive promoter and the light output was a measure of cell number13,
14,
15,
16. In this study we show that the Ef-luc mouse can be used in such studies, as it generates light that is proportional to both cell number and proliferation index. In the therapeutic studies presented here, apoptotic cell loss appears not to occur. Therefore, the reduction in light production seen after treatment with either PTK787/ZK222584 or CCI-779 is predominantly an effect of inhibition of cell proliferation, a notion concordant with the results from the immunohistochemical analysis. By contrast, during the development of a tumor where both cell number and proliferation rates are changing over time, this would not be the case. Also, light is lost during penetration through tissue, making comparisons between mice difficult because of the variable depth of each tumor. But using each mouse as its own control normalizes these variations and reduces the number of mice required to generate compelling results in preclinical trials.
The Ef-luc model makes it possible to investigate the importance of downstream signaling pathways in glioma maintenance. Previous results have shown that the Akt activity in our PDGF-driven gliomas is quite low17 and therefore one might have guessed a priori that mTOR activity would not be important for the proliferation of these tumor cells. Notably, however, our present data show that this appears not to be the case. It is possible that the mTOR dependence of these tumors may be trivial, having to do with the nutrient status of the cells18. An alternative explanation for the substantial effect of mTOR inhibition would be that it interferes with the recruitment of oncogenic mRNAs into polysomes. Recent data have shown that this is one of the mechanisms of Ras- and Akt-driven gliomagenesis19. Although it is possible that the growth-inhibitory effect seen with CCI-779 results from effects on proteins other than mTOR, there are no other known targets of the rapamycin analogs, and it seems unlikely that an unidentified target of this family of molecules would happen to be essential for PDGF-induced glioma proliferation. Assuming the effect is a result of mTOR inhibition, these data imply that pathways not activated by oncogenic stimulation may still be essential for tumor maintenance and therefore valid therapeutic targets.
In conclusion, these studies make it clear that qualified guessing at the importance of signaling pathways in tumors in vivo cannot be substituted for experiments that directly test the hypothesis. The development of the Ef-luc mouse can be of great help in this process and will be a valuable tool in such in vivo explorations in the future.
Methods Generation of Ef-luc and Ef-luc N−tv-a transgenic mice. A 1714-bp SmaI-XbaI fragment of the pGL3-Basic vector encoding a modified firefly luciferase (Promega) was cloned into a plasmid behind a 273-bp PCR-generated fragment of the human E2F1 promoter. The PCR primers used were E2F1-1 (5'-GGAATTCCATCCGGACAAAGCCTGCGCGC-3') and E2F1-2 (5'-GGAATTCAGGCCTCGGCGAGGGCTCGAT-3'); we added an EcoRI site at each end of the E2F1 promoter. The gene encoding luciferase was followed by a polyadenylation sequence. We generated chimeric founder mice by pronuclear microinjection of the linearized Ef-luc construct into fertilized FVB oocytes. Genotyping of transgenic mice was done by PCR using primers E2F1-1 and EFLUC-1 (5'-TGCGGGAGTTTCACGCCACCA-3'), yielding a product of 335 bp. We screened EFLUC-positive mouse lines for correct expression of the transgene in vitro. We identified 10 chimera founders that lead to 8 founders with germline transmission; three of these showed tight cell-cycle regulation of expression in vitro. The luciferase activity was measured in primary cultured cells obtained from the brains of newborn mice as described previously10. Equal numbers of cells were cultured in the presence of 10% fetal bovine serum for 1 d or in 0.5% fetal bovine serum for 4 d before lysis. Luciferase activity was measured using the Luciferase Assay System (Promega) according to the manufacturer's protocol.
Generation of mouse brain tumors. Double-transgenic neonatal Ef-luc N−tv-a mice were injected intracranially with 1 l DF-1 cells producing RCAS-PDGFB retrovirus as described previously11. Mice were monitored carefully for symptoms of tumor development (hydrocephalus, lethargy). All injected mice were routinely screened with BLI, and image-positive mice were followed over time, treated and followed over time, or killed.
BLI of Ef-luc N−tv-a mice. Mice were anesthetized with 3% isofluorane before retro-orbital injection with 75 mg/kg body weight n-Luciferin (Xenogen). Three minutes after injection of the n-Luciferin, images were acquired for 2 min with the Xenogen IVIS system (Xenogen) using Living Image analysis and acquisition software (Xenogen). A photographic image was taken, onto which the pseudocolor image representing the spatial distribution of photon counts was projected. We defined a circular region between the ears and used it as a standard in all experiments. From this region the photon counts were compared between different mice.
Drug treatment of tumor bearing Ef-luc N−tv-a mice. Image-positive Ef-luc N−tv-a mice were treated daily with PTK787/ZK222584 at 100 mg/kg body weight, CCI-779 at 40 mg/kg body weight, or buffer only for the indicated number of days. All doses were administered through intraperitoneal injection. We took transgenic images 24−144 h after initiation of treatment.
The resuspension solution for PTK787/ZK222584 consists of 5% DMSO and 1% Tween-80 in distilled water stored at 4 °C. This solution was used when treating the buffer-treated mice. We prepared PTK787/ZK222584 at 10 mg/ml and stored it at -20 °C. PTK787/ZK222584 does not dissolve in solution and therefore we inverted the tubes before each use. CCI-779 was prepared at 2 mg/ml in a resuspension buffer containing 5% Tween-80 and 5% PEG400, and stored at -20 °C.
Histological analysis, immunohistochemistry and TUNEL analysis. We removed the brains of killed mice, fixed the brains in formalin and embedded them in paraffin. Immunohistochemical stainings were performed using antibodies for PCNA (Oncogene), caspase-3 (Cell Signaling), Ki-67 (Novacastra), and histone H3 (Upstate). TUNEL analysis (Roche) was performed according to the manufacturer's protocol.
Received 24 March 2004; Accepted 7 July 2004; Published online: 24 October 2004.
REFERENCES
Sherr, C.J. & McCormick, F. The RB and p53 pathways in cancer. Cancer Cell2, 103−112 (2002). | Article | PubMed | ISI | ChemPort |
Johnson, D.G., Ohtani, K. & Nevins, J.R. Autoregulatory control of E2F1 expression in response to positive and negative regulators of cell cycle progression. Genes Dev.8, 1514−1525 (1994). | PubMed | ISI | ChemPort |
Parr, M.J. et al. Tumor-selective transgene expression in vivo mediated by an E2F- responsive adenoviral vector. Nat. Med.3, 1145−1149 (1997). | Article | PubMed | ISI | ChemPort |
Contag, C.H. et al. Visualizing gene expression in living mammals using a bioluminescent reporter. Photochem. Photobiol.66, 523−531 (1997). | PubMed | ISI | ChemPort |
Gelovani Tjuvajev, J. & Blasberg, R.G. In vivo imaging of molecular-genetic targets for cancer therapy. Cancer Cell3, 327−332 (2003). | Article | PubMed |
Massoud, T.F. & Gambhir, S.S. Molecular imaging in living subjects: seeing fundamental biological processes in a new light. Genes Dev.17, 545−580 (2003). | Article | PubMed | ISI | ChemPort |
Choy, G. et al. Comparison of noninvasive fluorescent and bioluminescent small animal optical imaging. Biotechniques35, 1022−1030 (2003). | PubMed | ChemPort |
Dai, C. et al. PDGF autocrine stimulation dedifferentiates cultured astrocytes and induces oligodendrogliomas and oligoastrocytomas from neural progenitors and astrocytes in vivo. Genes Dev.15, 1913−1925 (2001). | Article | PubMed | ISI | ChemPort |
Shih, A. et al. Dose dependent effects of PDGF-B on glial tumorigenesis. Cancer Res. (in press).
Podsypanina, K. et al. An inhibitor of mTOR reduces neoplasia and normalizes p70/S6 kinase activity in Pten+/- mice. Proc. Natl. Acad. Sci. USA98, 10320−10325 (2001). | Article | PubMed | ChemPort |
Burgos, J.S. et al. Time course of bioluminescent signal in orthotopic and heterotopic brain tumors in nude mice. Biotechniques34, 1184−1188 (2003). | PubMed | ISI | ChemPort |
Rehemtulla, A. et al. Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging. Neoplasia2, 491−495 (2000). | Article | PubMed | ISI | ChemPort |
Rehemtulla, A. et al. Molecular imaging of gene expression and efficacy following adenoviral-mediated brain tumor gene therapy. Mol. Imaging1, 43−55 (2002). | Article | PubMed | ChemPort |
Shah, K., Tang, Y., Breakefield, X. & Weissleder, R. Real-time imaging of TRAIL-induced apoptosis of glioma tumors in vivo. Oncogene22, 6865−6872 (2003). | Article | PubMed | ISI | ChemPort |
Dai, C. et al. The characteristics of astrocytomas and oligodendrogliomas are caused by two distinct interchangable signaling formats. Neoplasia (in the press).
Garami, A. et al. Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2. Mol. Cell11, 1457−1466 (2003). | Article | PubMed | ISI | ChemPort |
Rajasekhar, V.K. et al. Oncogenic Ras and Akt signaling contribute to glioblastoma formation by differential recruitment of existing mRNAs to polysomes. Mol. Cell12, 889−901 (2003). | Article | PubMed | ISI | ChemPort |
Acknowledgments The authors would like to thank C. Glaster for preparation of this manuscript, C. Discafani (Wyeth Research) for the CCI-779 and J. Wood (Novartis Pharmaceuticals) for the PTK787/ZK222584. This work was supported by the Tow, Seroussi, Bressler and Kirby Foundations.
Competing interests statement:
The authors declare that they have no competing financial interests.
l DF-1 cells producing RCAS-PDGFB retrovirus as described previously
