Introduction
Development of breast cancer mouse models that provide consistent primary mammary tumors and metastasis to clinically relevant tissues such as lymph nodes, lungs, and bone remain a challenge in the preclinical evaluation of therapies for human breast cancer. Current xenograft models of breast carcinoma involve murine or human breast cancer cell lines implanted into the mammary fat pad of mice or injected systemically by intravenous or intracardiac routes. Tumor cells injected into the mammary tissue yield reproducible tumors, but can require weeks to several months for primary tumor development and produce varied spontaneous metastasis depending on the cell line and mouse strain used in the study [
1].
One common human breast cancer cell line used in xenograft animals models is MDA-MB-231. These cells originated from a human metastatic ductal breast carcinoma sample [
2], are estrogen independent, and exhibit preferential growth in the mammary fat pad of immune compromised mice [
3]. MDA-MB-231 cells develop primary tumors that produce spontaneous metastasis to lymph nodes and micrometastases to the lungs [
4]. Detection of metastasis has relied primarily upon histological or PCR analysis of selected tissues at experimental end-points. Spontaneous metastasis to bone or soft organs from primary mammary tumors has not been reported.
Reproducible bone metastasis in breast cancer xenograft models has been achieved with intracardiac injection of MDA-MB-231 cells [
5,
6]. Passaging tumor cells harvested from the bone lesions several times
in vivo has created MDA-MB-231 sublines with exclusive propensity for bone metastasis [
7‐
10]. The bone metastases are typically identified in animals by radiographic or histological procedures. Recently, researchers have begun to apply luciferase-based imaging methods to detect widespread metastasis in mouse breast cancer models [
10‐
13]. In studies using luciferase-expressing MDA-MB-231 tumor cell sublines specifically selected for skeletal metastasis,
in vivo imaging was able to monitor experimental bone metastasis in mice to a level comparable to that of X-ray analysis [
10,
13]. Our goal was to develop a bioluminescent human breast cancer cell line that would offer a similar level of detection for both primary and metastatic tumors and would more fully mimic clinical breast cancer by metastasizing to multiple tissues, including lymph nodes, lungs, bone, and visceral organs.
This report describes bioluminescent xenograft mouse models based on more widely metastatic derivatives of MDA-MB-231 cells. These two luciferase-expressing cell lines, D3H1 and D3H2LN, were isolated for stable firefly luciferase expression in vitro and were passaged in mice to enhance their tumorigenic and metastatic properties. We evaluated the effect of long-term in vivo growth on the stability of cellular bioluminescence. In vivo and ex vivo imaging was used to monitor and compare the primary tumor growth rates and metastatic potential from mammary tumors and after intracardiac injection of the two MDA-MB-231 sublines over time.
Materials and methods
Tumor cell line
Human breast cancer cell line MDA-MB-231 was obtained from the American Type Culture Collection (Rockville, MA, USA). Cells were cultured in Minimum Essential Medium with Earl's Balanced Salts Solution MEM/EBSS medium supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 1% L-glutamine, and 1% sodium pyruvate (all from Hyclone, Logan, UT, USA) at 37°C in a humidified atmosphere containing 5% carbon dioxide.
Co-transfection and selection of cell line
MDA-MB-231 cells were co-transfected with plasmids expressing the firefly luciferase gene (luc; pGL3-Red, Chris Contag, Stanford University, Stanford, CA, USA) and zeocin resistance gene (pSV40/Zeo; Invitrogen, Carlsbad, CA, USA) using lipofectamine (Invitrogen). Transfected cells were then selected for antibiotic resistance (zeocin; Invitrogen) and surviving colonies were screened for bioluminescence in complete media supplemented with 150 μg/ml D-luciferin (Biosynth International, Inc, Naperville, IL, USA) by in vitro imaging using the IVIS™ camera system (Xenogen, Alameda, CA, USA; see below). Bioluminescent, antibiotic resistant, single cell clones were amplified in culture and characterized for stable luminescence in vitro and tumorigenic potential in vivo. One clonal cell line, MDA-MB-231-luc-D3, was initially selected. This D3 clone was passaged through mice as a primary tumor in the mammary fat pad to produce the D3H1 cell line. A spontaneous lymph node metastasis from a D3H1 mammary fat pad tumor was also harvested and designated D3H2LN. These three bioluminescent derivatives of MDA-MB-231 cells were used for further in vitro or in vivo studies.
Bioluminescent imaging
Bioluminescent imaging was performed with a highly sensitive, cooled CCD camera mounted in a light-tight specimen box (IVIS™; Xenogen), using protocols similar to those described previously [
14‐
17]. Imaging and quantification of signals were controlled by the acquisition and analysis software Living Image
® (Xenogen). For
in vitro imaging, bioluminescent cells were serially diluted from 4000 to 8 cells in complete media into black, clear bottomed, 96-well plates (Costar, Acton, MA, USA). D-luciferin (Biosynth International, Inc.) at 150 μg/ml in media was added to each well 5–10 min before imaging. Imaging time was 1 min/plate. For
in vivo imaging, animals were given the substrate D-luciferin by intraperitoneal injection at 150 mg/kg in DPBS Dulbecco's Phosphate Buffered Saline (Invitrogen, Carlsbad, CA, USA), and anesthetized (1–3% isoflurane). Mice were then placed onto the warmed stage inside the light-tight camera box with continuous exposure to 1–2% isoflurane. Imaging times ranged from 1 s to 3 min, depending on the tumor model and time point. Generally, two to three mice were imaged at a time. The low levels of light emitted from the bioluminescent tumors or cells were detected by the IVIS™ camera system, integrated, digitized, and displayed. Regions of interest from displayed images were identified around the tumor sites and were quantified as total photon counts or photons/s using Living Image
® software (Xenogen). Background bioluminescence
in vivo was in the region of 1 × 10
4 photon counts or 1–2 × 10
5 photons/s. For
ex vivo imaging, 150 mg/kg D-luciferin was injected into the mice just before necropsy. Tissues of interest were excised, placed into 24-well tissue culture plates with 300 μg/ml D-luciferin in DPBS, and imaged for 1–2 min. Tissues were subsequently fixed in 10% formalin (Sigma, St. Louis, MO, USA) and prepared for standard histopathology evaluation.
Mouse strain and animal care
Strict animal care procedures set forth by the Institutional Animal Care and Use Committee based on guidelines from the US National Institutes of Health Guide for the Care and Use of Laboratory Animals were followed for all experiments [
18]. Mice used in these studies were either female athymic nude-nu mice aged 8–10 weeks (Harlan, Indianapolis, IN, USA) or female nude-beige (NIH-bg-nu-xidBR) mice aged 8–10 weeks (Charles River Laboratories, Wilmington, MA, USA).
Mammary fat pad spontaneous metastasis model
Female nude mice or female nude-beige mice aged 8–10 weeks were anesthetized by exposure to 1–3% isoflurane and injected with 50 μl of 2 × 106 MDA-MB-231 cells suspended in 50% DPBS/50% matrigel into the abdominal mammary fat pad. At 10–15 min after luciferin injection, mice were placed in the IVIS™ Imaging System and imaged from the ventral view. Tumor growth was monitored weekly by bioluminescent imaging and external caliper measurements (tumor size = [length × width × height] × 0.52) for 5–9 weeks. In some experiments, the lower portion of each animal was shielded before reimaging in order to minimize the bioluminescence from the primary tumor so that the signals from the metastatic regions could be observed in vivo. The front limbs were secured with tape to better expose the axillary/brachial lymph node areas. The imaging time ranged from 1 s to 1 min, depending on the size of the primary tumors, but was consistently 3 min for detection of metastases.
Female nude mice (age 8–10 weeks) were anesthetized by intramuscular injection of 120 mg/kg ketamine hydrochloride with 6 mg/kg xylazine on the day of injections, and by exposure to 1–3% isoflurane on subsequent imaging days. On day 0, anesthetized animals were injected with 1 × 10
5 MDA-MB-231 cells suspended in 100 μl sterile DPBS into the left ventricle of the heart by nonsurgical means [
19]. Anesthetized mice were placed in the IVIS™ Imaging System and imaged from both dorsal and ventral views approximately 10–15 min after intraperitoneal injection of D-luciferin. A successful intracardiac injection was indicated on day 0 by images showing systemic bioluminescence distributed throughout the animal. Only mice with evidence of a satisfactory injection continued in the experiment. Assessment of subsequent metastasis was monitored
in vivo once a week by imaging for up to 5 weeks.
Histopathology
To confirm the presence of neoplastic cells, selected tissues were excised from the mice at necropsy and were preserved in 10% formalin solution (Sigma) immediately after ex vivo imaging. Tissues were prepared for histopathology (paraffin preparation, sectioning, and hematoxylin and eosin staining) and analyzed by subsequent microscopic evaluation by IDEXX Veterinary Services (West Sacramento, CA, USA).
Statistical analyses
The mean bioluminescence (photons/s), tumor volume, and corresponding standard errors of the mean were determined for each experiment. Regression plots were used to describe the relationship between bioluminescence and cell number and tumor volume; R2 values are reported to assess the quality of the regression model.
Discussion
Bioluminescent imaging has been used to detect primary cancer growth and metastasis in a growing number of animal models [
14‐
20]. In xenograft tumor models, human cell lines expressing luciferase have permitted studies that yield real-time, noninvasive monitoring of tumor sites in the same cohort of animals over time. This study is the first application of
in vivo bioluminescent imaging to monitor breast cancer tumor growth in animals and to detect spontaneous metastasis of tumor cells from the mammary fat pad to lymph nodes and lungs. The luciferase-expressing subclones of MDA-MB-231 characterized in this report also produced multiple metastases at high frequencies to clinically relevant tissues such as bone and brain following intracardiac injection of cells. With the ability to produce lymph node and lung metastasis from primary tumors and widespread metastasis after intracardiac injection, the D3H1 and D3H2LN subclones of MDA-MB-231 cells fully mimic the range of breast cancer development in humans.
Previous studies have noted lymph node and lung metastasis from parental MDA-MB-231 mammary tumors in mice but have largely relied on postmortem histology or PCR to document the presence of limited metastasis, especially to the lungs, from MDA-MB-231 primary tumors [
4]. Our imaging experiments yielded similar findings but they did not require time consuming tissue processing and analysis. The low bioluminescence of the micrometastasis to the lung from mammary tumors was detected immediately in excised tissues by
ex vivo analysis. The more pervasive lymph node metastasis was detected in both live animals over time and by
ex vivo examination. Subsequent microscopic evaluation of the metastatic tissues initially identified as positive or negative by bioluminescent imaging was used to confirm tumor spread and allowed precise localization and sizing of lesions at the cellular level.
MDA-MB-231 luciferase-expressing cells were injected in the mammary fat pad of mice and passaged one or two times in vivo to generate a more robust xenograft model of human breast cancer with enhanced primary tumor growth and subsequent metastasis. In evaluating the photon emission of these cell lines, we found no decrease in bioluminescence between the D3H1 cells, D3H2LN cells, and the original unpassaged D3 subclone from the parental MDA-MB-231 cells, despite 12 weeks of tumor growth for D3H1 cells and 24 weeks of accumulated in vivo growth time for the D3H2LN line. Additionally, the D3H1 and D3H2LN cell lines maintained consistent levels of bioluminescence in vitro for up to 2 months or approximately eight to 10 passages in culture (data not shown). These data demonstrate remarkable long-term stability of luciferase expression after multiple rounds of in vivo growth and continuous periods of in vitro culturing.
The intracardiac injection model is used to generate widespread arterial dissemination of tumor cells by bypassing the lungs in order to seed cells to various organs. This method has been employed with luciferase-positive MDA-MB-231 derivative cell lines that were selected for enhanced
in vivo metastasis to bone [
10,
13]. The metastatic spread of these cells after intracardiac injection was primarily to bone sites, with limited visceral metastasis to the adrenals or pancreas. Our D3H1 and D3H2LN sublines expand the utility of such models by producing consistent metastatic lesions to brain and various visceral organs as well as skeletal sites following intracardiac injection. In previous reports researchers compared bioluminescent imaging with radiometric measurements of bone metastasis, and found
in vivo imaging to be the faster and more sensitive detection method [
10]. Similarly, our intracardiac experiment compared imaging with histological examination and showed that, in many cases, tissues that had been found to be positive by
in vivo or
ex vivo imaging required multiple rounds of sectioning and histology to identify the micrometastasis.
In both nude and nude-beige mice, the D3H2LN tumors produced a faster rate of tumor growth in the mammary fat pad as well as enhanced spontaneous metastasis to the lymph nodes and lungs compared with the D3H1 cells (Fig.
2 and Table
1). The D3H2LN cells also exhibited higher frequencies of metastasis to bone, brain, lungs, and soft tissues after intracardiac injection (Table
2). Taken together, these findings suggest a more invasive phenotype in the D3H2LN cells and demonstrate an enhanced utility in animal metastasis studies. We did find that the rapid growth of D3H2LN cells in some mammary fat pad experiments created necrotic tumors that showed a decrease in bioluminescent signal
in vivo at later time points, despite an incremental increase in tumor volume (Fig.
2a). A similar disparity between volume measurements and bioluminescence with larger tumors was reported in a study using subcutaneous tumors of human prostate cells expressing luciferase [
11]. Here, the authors noted that tumor volume became static after treatment of animals, while a drop in bioluminescence of the same tumors indicated a tumoricidal effect on cells that was subsequently confirmed by histology.
Conclusion
The bioluminescent stability of the D3H1 and D3H2LN cell lines, coupled with their enhanced tumorigenicity and widespread metastatic potential, provides a sensitive in vivo model system for preclinical assessment of breast cancer growth, dissemination, and response to anticancer therapies.
Competing interests
All authors are current or former employees of Xenogen Corporation. Xenogen manufactures the IVIS Imaging System™ and provides commercially available bioluminescent tumor cell lines as part of their Bioware™.
Authors' contributions
DEJ conceived the study, directed the experiments, and wrote the manuscript. YSH and YO carried out the in vivo experiments, organized the data, and contributed to drafts of the manuscript. JD carried out in vitro experiments and participated in drafting the final manuscript. TP supervised the project and participated in drafting the manuscript.