We currently have limited local strategies for treating vertebral metastases.10,12,24,28-30 However, because there has been a decline in age-standardized cancer mortality from improved adjuvant therapies, skeletal metastases are increasingly prevalent.32 Thus, the ability to test interventions on early-stage metastatic disease is important to enhance survival. When considering the development and evaluation of innovative local therapies, animal models provide a necessary step after in vitro studies to introduce therapies into clinical trials.
Several (ie, early, asymptomatic) animal models of osteolytic metastases have been reported.14,21,26,33 Orthotopic injection of malignant cells in athymic rodent models has been described, however, direct vertebral injection would be technically difficult given vertebral size in this model.6,9 Intracardiac injection bone metastasis models are also well-established with the advantage of more closely mimicking the process of metastatic spread.4,7,19,21-23,26 Several human breast cancer cells lines have been evaluated in rodent models, however few specifically detail vertebral metastases.2,8,14,21,26 Potential differences in the temporal development of appendicular versus vertebral metastases have not been extensively quantified. An advantage of our rodent model is the use of human breast cancer cells, which is attractive in that it facilitates assessment of the intrinsic sensitivity of the cancer cell type of clinical interest to experimental treatment. Of the human primary malignancies with metastatic propensity in bone, patients with breast cancer can live for years even with established metastasis.1,13,17 Our model represents an attractive preclinical animal model given its relevance to the potential clinical population that may be at risk for skeletally related events. The evaluation of bisphosphonates in the systemic treatment of breast cancer bone metastases has been an excellent example of how this model can test therapies before clinical application.7,17,19,26
Because of the reliance on circulatory seeding of tumor cells in vertebrae, a limitation of this model is the time course of metastatic spread and early identification of vertebral lesions that can be targeted for local therapies. Unlike the scenario of systemic therapy, where the effects on overall metastatic bone burden can be readily quantifiable using end points including animal survival, evaluating a local vertebral therapy requires early identification of a treatable vertebral lesion so treatment can be initiated and outcomes evaluated before acceptable end points in animal care and survival. The potential feasibility of this model to test local vertebral therapies is not fully established. A timeline must be established for the detection of vertebral metastasis in the context of metastatic burden and animal survival.
The purposes of this study were to: (1) determine if there are differences in the temporal development of appendicular versus axial (vertebral) metastases as quantified by longitudinal conventional in vivo radiography (ie, fine detail radiography and two-dimensional (2-D) fluoroscopy); (2) compare detection rates between bioluminescent reporter imaging and conventional radiographs in identifying vertebral metastasis; and (3) compare tumor burden as quantified by bioluminescent imaging with subsequent ex vivo microcomputed tomography (microCT) analysis of vertebral osteolysis.
MATERIALS AND METHODS
Thirty-five athymic rats were given an intracardiac injection of human MT-1 cells. Detection rates for appendicular and axial metastases were documented longitudinally using in vivo conventional fine detail radiography and 2-D fluoroscopy at serial times after tumor cell injection. Animals were euthanized 28 days after cell injection or at acceptable tumor end points for animal care. In a subset of 15 animals, bioluminescence imaging was performed at 21 days and compared with conventional radiography for detection of vertebral metastases. Postmortem necroscopy, histology, and ex vivo microCT imaging of vertebral metastases were performed. Tumor burden as quantified by bioluminescence imaging was compared with ex vivo microCT evaluation of osteolysis.
Human breast cancer MT-1 cells were cultured in RPMI media at 37oC with 5% CO2. At 70% to 80% subconfluency, the cells were given fresh media 24 hours before inoculation into athymic rnu/rnu rats. We determined cell viability by trypan blue exclusion, and we used cell suspensions with greater than 95% viability without cell clumping. Intracardiac injections of 2 × 106 MT-1 cells in 0.2 mL free RPMI media were given into the left heart ventricle of 20 anesthetized rnu/rnu female rats. Each animal (4-6 weeks) was anesthetized using nose-cone inhalation of a 2% halothane/air mixture. We prepared the anterior chest with 100% ethanol and injected the 0.2 mL mixture of 2 × 106 MT-1 cells into the left ventricle using a 1-mL syringe with a 26-gauge needle, similar to that described by other investigators.2,8,25 The spontaneous, pulsatile entrance of bright red oxygenated blood into the syringe determined appropriate positioning in the left ventricle of the heart, and the mixture was injected slowly over 30 seconds. Animals were allowed ad lib cage activity after recovery from anesthesia. The study protocol was approved in accordance with institutional and national animal care standards.
The rodents were anesthetized and had imaging by fine detail radiography (Faxitron MX 20®, Faxitron X-ray Corporation, Wheeling, IL) (35kV, 2 seconds, ×2 magnification) at serial times after cell inoculation (14, 21, and 28 days). We also performed mini-C-arm 2-D fluoroscopy (Fluoroscan I, Fluoroscan Imaging Systems, Inc, Northbrook, IL) to detect vertebral metastases. We recorded weight at the time of initial injection and during serial radiographic imaging times.
Animals were euthanized by CO2 inhalation 28 days after tumor cell injection, or at accepted institutional tumor end points for animal care (range, 22-28 days). We then harvested the vertebral column, which was imaged using an ex vivo microCT imaging system (EVS, London, Ontario, Canada) (axial slice thickness, 52 microns). Sagittal and coronal reformatted views of the vertebral column were created using MicroView™ software (Microview International, Aurora, Ontario, Canada). The extent of tumor involvement in the spinal vertebrae, as evidenced by osteolytic bone destruction, was quantified using a Nikon slide scanner and Image Pro software (Media Cybernetics, Silver Spring, MD). The maximal areas of osteolytic involvement (mm2) in the midsagittal and midcoronal images in an involved vertebra were quantified relative to the vertebral body area (mm2).
Samples were fixed in 10% formalin for 7 days and decalcified in 10% formic acid for another 7 days. We then embedded the samples in paraffin and prepared two adjacent 10-μm midsagittal sections. Histologic sections were analyzed by an experienced pathologist (RK) using hematoxylin and eosin staining. A human horseradish peroxidase-labeled primary antibody to the endothelial growth factor receptor (EGF-r; Invitrogen Inc, Paisley, Scotland) was used to label human MT-1 cells in rat vertebral bone marrow.
Cultured human MT-1 cells were cotransfected using dioctadecylamidoglycylspermine-trifluoroacetate salt (Transfectam; E1231, Promega) with plasmid DNA containing the luciferase gene of the firefly (Photinus pyralis, pGL3 control vector E1741, Promega) and a plasmid containing the neomycin resistance gene (PCI-Neo, E184, Promega). Luciferin (Beetle, potassium salt anhydrous, E1603) and neomycin antibiotic (G-418 sulphate) also were obtained from Promega. Stable transfects of luciferase-expressing MT-1 cells (MT-1Luc) were confirmed by growing cells in the presence of G-418 antibiotic (100 μg/mL) for 10 days, at which time colonies expressing the highest bioluminescent signal were isolated and cultured. We measured bioluminescence images using a commercial system (Xenogen IVIS System, Xenogen Corporation, Alameda, CA) after adding the luciferin substrate (25 μmol/L) to cells in vitro. Ten microliters of luciferin stock (0.5 mmol/L in PBS) was added to MT-1Luc cells containing 190 μL of growth media (without phenol red). The plates were agitated gently and placed in the bioluminescent IVIS imaging system. The bioluminescent signal was integrated over 5 minutes using the Living Image™ software and plotted against time.
Fifteen rnu/rnu rats were inoculated with 2 × 106 MTLuc cells as previously described. Anesthetized animals were imaged using the bioluminescent imaging system 21 days after cell inoculation. The anesthetized animals were given an intraperitoneal (30 mg/kg) injection of luciferin. Each animal, held on a custom radiolucent stereotactic frame, was then imaged starting 5 minutes after injection with a 5-minute acquisition. Bioluminescent signal was captured as the absolute total flux (photons/steradian/cm2) emitted with a 5-minute integration time and plotted against time. The animals were imaged in the lateral decubitus and dorsoventral positions on the radiolucent frame, and suspected vertebral metastases were identified based on bioluminescent signal. The epicenter of maximal signal then was localized to coordinates based on horizontal and vertical grid-marks demarcated in standard increments on the radiolucent frame. The animals then were imaged on the frame by mini-C-arm fluoroscopy during the same surgical setting. The use of a 21-gauge needle guided identification of the vertebral level of interest. Using fluoroscopy, the needle could be localized in conjunction with the coordinates rendered by the frame. Animals were euthanized after imaging, and ex vivo microCT scan imaging and histologic analysis were performed on the targeted vertebrae. We performed statistical analyses using SPSS 6.1.3 (SPSS Inc, Chicago, IL) using Spearman's rank correlation coefficient for comparative analysis between image modality and quantification of tumor burden.
Appendicular metastases were detectable earlier in this model when compared with axial (vertebral) metastases using longitudinal in vivo conventional radiography. In the 20 rnu/rnu rats inoculated by intracardiac injection of MT-1 cells, including animals in our learning curve, 14 had metastatic disease develop as confirmed by postmortem necroscopy and histologic analysis. The 14 animals had a mean survival of 25 days. All animals with metastatic tumors became cachectic by approximately 18 days. Weight loss was an early sensitive indicator identifying animals with subsequent metastases. Osteolytic metastases in the appendicular skeleton (ie, femur, humerus) were identified as early as 14 days by fine detail radiography (12/14 animals at 14 days). However, vertebral metastases could not be detected either by fine detail radiography or fluoroscopy at this time or by 21 days after cell injection (0/14 animals at 21 days versus 14/14 at 21 days for detection of appendicular lesions; p < 0.05). We observed that in vivo conventional fine detail radiography and 2-D fluoroscopy were poor early detectors of vertebral metastases. Despite the lack of visual vertebral metastases on conventional radiographs by 21 days, ex vivo microCT scan analysis of the thoracic and lumbar spine after sacrifice of the animals showed multiple osteolytic vertebral lesions, which were confirmed histologically. The lesions, approximating as much as ⅓ of the vertebral body size in the thoracic and lumbar spine, were apparent in the majority of contiguous vertebrae (Fig 1; Table 1).
In animals subject to in vivo bioluminescence imaging, the technique was more sensitive (p < 0.01) for earlier detection of vertebral lesions, showing multiple vertebral metastases in all imaged animals by 21 days (n = 15/15). By comparison, vertebral metastasis was identified in only one of 15 animals in this experimental group when imaged at 21 days by conventional radiography. Vertebral metastases could not be identified reliably by fine detail radiography or fluoroscopy until after 25 to 28 days, which was late during the course of metastatic spread when animals were closely approaching accepted tumor end points for animal care.
Bioluminescent signal intensity varied among animals, with areas of highest vertebral signal correlating to subsequent ex vivo microCT (r = 0.71) quantification of vertebral tumor burden. The use of the stereotactic radiolucent frame facilitated the colocalization of bioluminescent signal to vertebral levels (Fig 2). Areas of appendicular skeletal involvement (eg, femur, humerus) also had increased signal intensity. In four animals, the highest signal intensity occurred in the chest cavity indicating involvement of the lung which subsequently was confirmed by gross evaluation postmortem. Histologic staining with hematoxylin and eosin demarcated areas of bone marrow infiltration as evidenced by cellular necrosis. Immunohistochemical staining for human EGF-r confirmed the presence and provided reliable labeling of human breast cancer cells in areas of bone marrow infiltration in harvested rat vertebral specimens (Fig 3).
When considering the development and evaluation of local therapies to treat vertebral metastases, detection of early-stage metastatic disease is desirable so treatment intervention and outcome can be evaluated in animal models before acceptable tumor end points in animal care. We evaluated several image modalities for detection of vertebral metastases. Fine detail radiography and 2-D mini-C-arm fluoroscopy could not reliably detect vertebral metastases until very late in the course of metastatic spread. Earlier detection of vertebral metastases was facilitated by a bioluminescent tumor model created using MT-1 cells transfected with PGL3 and PCI-neo plasmids conferring luciferase activity and neomycin resistance. The vertebral regions of highest bioluminescent signal could be colocalized to individual vertebrae by subsequent fluoroscopy using a custom radiolucent stereotactic frame.
We note several limitations to our study. Varying factors influence temporal development of metastasis after intracardiac injection of tumor cells. Although we used cell suspensions with greater than 95% viability, not all inoculated animals had metastasis develop as confirmed clinically, radiographically, at the time of necroscopy, or during subsequent histologic evaluation. The development of metastasis depends on the intracardiac injection technique and the number and type of cells inoculated.2 Jenkins et al reported on the use of bioluminescence imaging to confirm systemic distribution of injected cells after intracardiac injection to improve identification of animals that have metastases develop.15 We did not evaluate different cell concentrations, and the potential impact of a different concentration on the temporal development of metastasis requires ongoing evaluation.
An advantage of bioluminescence imaging is that it can be repeated serially with time in the same animal to monitor growth kinetics for each metastatic site, thus providing end-point analysis of experimental treatment of vertebral lesions. We performed bioluminescence imaging to detect vertebral lesions at a single time-point following tumor cell injection. The potential utility of the approach at earlier times warrant additional study given the high sensitivity of the imaging for detecting vertebral metastasis at 21 days in our model. We also did not evaluate other potential biomarkers for tumor imaging.
Choy et al observed bioluminescence imaging in a mouse tumor model was more sensitive than green fluorescent protein GFP imaging for early detection of viable tumor cells, noting bioluminescence imaging was able to detect labeled cells as early as 1 day after cell injection compared with 7 days later for fluorescence imaging.5 Although an advantage of fluorescence imaging versus bioluminescence is that it does not require injection of a substrate at the time of image acquisition. Fluorescence imaging may be more suitable to applications where sensitivity and early detection are not as critical.5
Bioluminescence imaging was not compared with other potential in vivo biologic imaging modalities such as bone scanning with Tc99m-MDP or positron emission tomography (PET). From a clinical perspective, bone scan imaging has a reported high sensitivity (85-90%) with a low specificity (80%) for bony metastasis, and FDP-PET scanning may be more specific for detection of metastatic disease although the sensitivity and specificity of the test may vary according to tumor type.3,11,18,27 A theoretical advantage of bioluminescent imaging is the detection of viable labeled tumor cells which potentially could facilitate gauging direct pretreatment and posttreatment effects directed toward tumor ablation. In the context of technetium bone scanning, background activity either by baseline bone metabolic activity or periphyseal activity in skeletally immature rodents used in this study and the theoretic ability to distinguish viable tumor cell growth versus host bony repair response are additional theoretical considerations for detecting early vertebral metastases. Finally, the use of in vivo microCT scanning, which was not an available resource for our study and thus a study limitation, could facilitate longitudinal evaluation of vertebral bony changes associated with tumor burden in each animal. Additional exploration comparing bioluminescent imaging with other available imaging modalities may be of value.
Consistent with other studies, our data support bioluminescence imaging for earlier detection of skeletal and vertebral metastases when compared with conventional radiography.23,31 In a murine intracardiac inoculation bone metastases model, Wetterwald et al described the sensitivity of bioluminescence in detecting intramedullary tumor growth preceding the appearance of radiographic osteolysis by approximately 2 weeks.31 In the study by Minn et al, skeletal radiographs and bioluminescent images performed 16 days and 45 days after intracardiac injection of MDA MB-231 cells in a xenograft mouse model were compared.20 They found bioluminescence imaging was more sensitive for earlier detection of bony metastases at 16 days, and osteolysis was not apparent on radiographs.20 They also observed, in a longitudinal followup at 45 days, that the anatomic sites of bioluminescent signal corresponding to the skeleton further increased in signal intensity, consistent with increasing tumor burden that was reflected in the development of radiographically apparent osteolysis not seen during the earlier evaluation.20
The use of multimodality imaging is a strategy that can guide end-point assessment in vertebral metastasis research. A consideration of bioluminescent imaging for detection of viable metastatic sites is the differentiation between bony versus nonbony metastases. The colocalization of bioluminescent signal during the same surgical setting by subsequent conventional radiographic imaging of animals on a radiolucent frame facilitated identification of vertebral metastasis and guided subsequent evaluation of tumor burden as determined by microCT imaging and histologic analysis. The utility of such an approach to guide early identification of bony metastasis and facilitate subsequent evaluation of tumor burden by conventional parameters, such as standard external calipers measurements of tumor volume, histology, and subsequent radiographic osteolysis, is gaining increasing enthusiasm and seems to complement assessment of tumor burden as determined by bioluminescent reporter imaging.16,20 The timeline of metastatic spread in this animal model is an important issue for early detection of vertebral metastasis when considering its application to gauge potential treatment effects after local therapeutic intervention.
We thank the Canadian Breast Cancer Research Alliance (formerly Canadian Breast Cancer Research Initiative-IDEAS grant) for supporting this research. The human breast cancer MT-1 cells were kindly provided by Dr. O. Engebraaten, Norwegian Radium Hospital, Oslo, Norway. We also thank Dr. Burton Yang, Dr. Joel Finkelstein, Ms. Bing Yang, and Dr. M. Henkelman for assistance in this study.
1. Ali SM, Harvey HA, Lipton A. Metastatic breast cancer: overview of treatment. Clin Orthop Relat Res
. 2003;415 (suppl):S132-S137.
2. Arguello F, Baggs RB, Frantz CN. A murine model of experimental metastasis to bone and bone marrow. Cancer Res
3. Bares R. Skeletal scintigraphy in breast cancer management. Q J Nucl Med
4. Blomme EA, Dougherty KM, Pienta KJ, Capen CC, Rosol TJ, McCauley LK. Skeletal metastasis of prostate adenocarcinoma in rats: morphometric analysis and role of parathyroid hormone-related protein. Prostate
5. Choy G, O'Connor S, Diehn FE, Costouros N, Alexander HR, Choyke P, Libutti SK. Comparison of noninvasive fluorescent and bioluminescent small animal optical imaging. Biotechniques
. 2003;35:1022-1026, 1028-1030.
6. Cincotta L, Szeto D, Lampros E, Hasan T, Cincotta AH. Benzophenothiazine and benzoporphyrin derivative combination phototherapy effectively eradicates large murine sarcomas. Photochem Photobiol
7. El-Abdaimi K, Ste-Marie LG, Papavasiliou V, Dion N, Cardinal PE, Huang D, Kremer R. Pamidronate prevents the development of skeletal metastasis in nude mice transplanted with human breast cancer cells by reducing tumor burden within bone. Int J Oncol
8. Engebraaten O, Fodstad O. Site-specific experimental metastasis patterns of two human breast cancer cell lines in nude rats. Int J Cancer
9. Fingar VH, Kik PK, Haydon PS, Cerrito PB, Tseng M, Abang E, Wieman TJ. Analysis of acute vascular damage after photodynamic therapy using benzoporphyrin derivative (BPD). Br J Cancer
10. Finkelstein J, Zaveri G, Wai E, Vidmar M, Kreder H, Chow E. A population-based study of surgery for spinal metastases: survival rates and complications. J Bone Joint Surg Br
11. Fogelman I, Cook G, Israel O, Van der Wall H. Positron emission tomography and bone metastases. Semin Nucl Med
12. Ghogawala Z, Mansfield FL, Borges LF. Spinal radiation before surgical decompression adversely affects outcomes of surgery for symptomatic metastatic spinal cord compression. Spine
13. Guarneri V, Conte PF. The curability of breast cancer and the treatment of advanced disease. Eur J Nucl Med Mol Imaging
. 2004;31 (Suppl 1):S149-S161.
14. Hiraga T, Williams PJ, Mundy GR, Yoneda T. The bisphosphonate ibandronate promotes apoptosis in MDA-MB-231 human breast cancer cells in bone metastases. Cancer Res
15. Jenkins DE, Hornig YS, Oei Y, Dusich J, Purchio T. Bioluminescent human breast cancer cell lines that permit rapid and sensitive in vivo detection of mammary tumors and multiple metastases in immune deficient mice. Breast Cancer Res
16. Jenkins DE, Oei Y, Hornig YS, Yu SF, Dusich J, Purchio T, Contag PR. Bioluminescent imaging (BLI) to improve and refine traditional murine models of tumor growth and metastasis. Clin Exp Metastasis
17. Jimeno A, Amador ML, Gonzalez-Cortijo L, Tornamira MV, Ropero S, Valentin V, Hornedo J, Cortes-Funes H, Colomer R. Initially metastatic breast carcinoma has a distinct disease pattern but an equivalent outcome compared with recurrent metastatic breast carcinoma. Cancer
18. Koga S, Tsuda S, Nishikido M, Ogawa Y, Hayashi K, Hayashi T, Kanetake H. The diagnostic value of bone scan in patients with renal cell carcinoma. J Urol
19. Michigami T, Hiraga T, Williams PJ, Niewolna M, Nishimura R, Mundy GR, Yoneda T. The effect of the bisphosphonate ibandronate on breast cancer metastasis to visceral organs. Breast Cancer Res Treat
20. Minn AJ, Kang Y, Serganova I, Gupta GP, Giri DD, Doubrovin M, Ponomarev V, Gerald WL, Blasberg R, Massague J. Distinct organspecific metastatic potential of individual breast cancer cells and primary tumors. J Clin Invest
21. Mundy G. Preclinical models of bone metastases. Semin Oncol
22. Nakata A, Tsujimura T, Sugihara A, Okamura H, Iwasaki T, Shinkai K, Iwata N, Kakishita E, Akedo H, Terada N. Inhibition by interleukin 18 of osteolytic bone metastasis by human breast cancer cells. Anticancer Res
23. Peyruchaud O, Winding B, Pecheur I, Serre CM, Delmas P, Clezardin P. Early detection of bone metastases in a murine model using fluorescent human breast cancer cells: application to the use of the bisphosphonate zoledronic acid in the treatment of osteolytic lesions. J Bone Miner Res
24. Ratliff J, Nguyen T, Heiss J. Root and spinal cord compression from methylmethacrylate vertebroplasty. Spine
25. Ree AH, Tvermyr M, Engebraaten O, Rooman M, Rosok O, Hovig E, Meza-Zepeda LA, Bruland OS, Fodstad O. Expression of a novel factor in human breast cancer cells with metastatic potential. Cancer Res
26. Sasaki A, Boyce BF, Story B, Wright KR, Chapman M, Boyce R, Mundy GR, Yoneda T. Bisphosphonate risedronate reduces metastatic human breast cancer burden in bone in nude mice. Cancer Res
27. Soderlund V. Radiological diagnosis of skeletal metastases. Eur Radiol
28. Wai EK, Finkelstein JA, Tangente RP, Holden L, Chow E, Ford M, Yee A. Quality of life in surgical treatment of metastatic spine disease. Spine
29. Wedin R, Bauer HC, Rutqvist LE. Surgical treatment for skeletal breast cancer metastases: a population-based study of 641 patients. Cancer
30. Weigel B, Maghsudi M, Neumann C, Kretschmer R, Muller FJ, Nerlich M. Surgical management of symptomatic spinal metastases: postoperative outcome and quality of life. Spine
31. Wetterwald A, van der Pluijm G, Que I, Sijmons B, Buijs J, Karperien M, Lowik CW, Gautschi E, Thalmann GN, Cecchini MG. Optical imaging of cancer metastasis to bone marrow: a mouse model of minimal residual disease. Am J Pathol
32. Wong DA, Fornasier VL, MacNab I. Spinal metastases: the obvious, the occult, and the impostors. Spine
33. Yoneda T, Michigami T, Yi B, Williams PJ, Niewolna M, Hiraga T. Actions of bisphosphonate on bone metastasis in animal models of breast carcinoma. Cancer