Background
Transgenic mouse models have been developed to recapitulate the complex effects of genes known to be involved in human breast cancer. These models can help to elucidate the mechanism of action of these genes during carcinogenesis, as well as their impact on normal mammary gland development. Imaging methods for mouse models of normal and cancerous mammary glands are in the developing stages and can help in the search for better ways to diagnosis human breast cancer earlier [
1].
In order to study early phenotypic effects of gene over-expression or lack of expression on mammary gland development and cancer traditional methods require that tissue is harvested from the animal and subjected to histological techniques to detect morphologically aberrant growth. These invasive procedures preclude further examination of the effects of these genetic changes on the process of carcinogenesis. Later, once the tumor becomes palpable, the size of the developing tumor can be measured and followed
in vivo over time to determine proliferative capacity. However, no information about initiation and progression can be gathered; only at the end of the experiment [
2] can information about the morphology or gene expression profile of the developing tumor be obtained.
The implementation of
in vivo imaging modalities to study normal mammary gland growth and disease progression has greatly improved the utility of these models, allowing the study of mammary differentiation or disease process, not simply the final effect [
3]. Biochemical and morphologic changes associated with early cancer change the optical properties of tissue, especially the absorption, scattering, and fluorescence, allowing the detection of these early carcinogenic effects with optical spectroscopy techniques [
4]. Imaging can allow researchers to, with minimal invasiveness, detect and follow abnormalities in ductal development during mammary differentiation in the same living animal. In cancer studies, imaging can detect undissected preneoplastic lesions and follow the behavior of these cancer cells, interactions with their stromal environment during the development of a tumor, angiogenesis, and metastatic disease. This can all be studied over time in the context of the cancer cells own physiological environment with an intact blood supply and interaction with surrounding tissues in 3-D and in real-time [
1,
5,
6]. Imaging regimens can also be adapted to evaluate efficacy and response of a cancer to prevention and therapeutic interventions [
7,
8] and to detect the presence of chemoresistance [
9]. Serial minimally invasive imaging of mice reduces the number of mice needed per experiment or in preclinical drug development since multiple time points can be observed in the same animal [
8]. The imaging modalities reflectance confocal microscopy (RCM), green fluorescent protein (GFP) imaging, and ultrasound imaging were utilized in this paper to image mammary glands and mammary tumors.
RCM provides real-time minimally invasive 3-D sectioning of
in vivo (living) or
ex vivo (newly biopsied) individual cells and tissues using variations in the optical properties of the natural backscattering of light from different cellular and subcellular structures without the use of labeling cells fluorescently or otherwise [
10,
11]. Optical techniques such as RCM have demonstrated high sensitivity for detecting cancer in their natural environment without using ionizing radiation [
12] and without time-consuming and potentially destructive fixation and staining, both of which may introduce artifacts and damage tissue [
13]. Tissue studied with RCM is treated with acetic acid, which induces DNA condensation providing increased reflectance to contrast nuclear versus cytoplasmic structure. We have shown that tissue treated with acetic acid can then be subjected to histological and immunohistological analyses without detrimental effects on the tissue [
14], facilitating further study into signaling pathways which may be active in the imaged structure [
15]. RCM has been performed on biopsy specimens to assess tumor margins [
16] and to identify precancerous lesions in human breast core needle biopsies [
14], the cervix [
17], and skin [
18].
Fluorescent protein labeling and epi-illumination spectroscopy microscopy are very powerful tools to follow primary tumor growth and metastasis with fluorophores
in vivo and in real time [
19]. Transgenic GFP optical imaging is one type of fluorescent protein label imaging and involves the detection of reporter transgene expression, namely a genetically encoded fluorescent protein, which is utilized to image cells within living tissue [
3,
20]. The specimen, often exposed surgically, is illuminated with blue light (488 nm excitation wavelength) which is absorbed by green fluorescent protein, a protein originally from the jellyfish
Aequorea Victoria [
6]. GFP then emits green light (509 nm peak shifted emission wavelength) which is collected by CCD cameras [
21]. GFP imaging can be used as a cell marker in both the living animal and in tissue culture and does not require a substrate for visualization [
22].
GFP transfected tissue culture cells and GFP transgenic mice have been used to monitor real time tumor growth and for mechanistic studies [
23,
24], evaluate the efficacy of therapy in a tumor xenograft model with metastasis [
25], monitor specificity of
in vivo gene therapy studies [
26], mark and sort potential mammary stem cells [
27], and examine mammary epithelial tumor cell behavior in metastasis [
28]. In addition to monitoring mammary gland development on the whole at the ductal morphology level as in our current study, GFP can be used to image single cells. This high resolution GFP imaging of cells
in vivo has been combined in a dual labeling approach with red fluorescent protein (RFP) to monitor tumor-stroma interactions and drug response of cancer and stromal cells [
29,
30].
Ultrasound imaging involves exposing tissues to high-frequency ultrasound waves (20–60 MHz in animals; 2–10 MHz in humans) by placing a transducer (which contains crystals that vibrate when exposed to small electrical currents and produce sound waves) on the skin and then detecting the ultrasound reflections from internal organs under investigation [
6,
31,
32]. This non-invasive technique produces a dynamic real-time image of the tissue from which structural and functional information can be obtained because sound waves travel though soft tissue based on the acoustic impedance of each tissue, which is a function of the tissue density [
31]. When two tissues with different densities are next to each other, a mismatch in the acoustic impedance causes sound waves to be reflected relative to the degree of mismatch; a greater acoustic impedance mismatch leads to a greater reflected pressure magnitude or intensity and is seen as a brighter image [
31].
Ultrasound is a rapid non-radiation method that has been used to detect cystic masses [
33] and superficial tumors [
34], differentiate between fibroadenomas and carcinomas in animal models [
35], noninvasively track liver metastases growth and evaluate potential therapy in liver metastasis models [
36], measure blood flow by Doppler [
37,
38], guide biopsy of a palpable breast mass [
38], and guide injections into target organs [
39].
In the present study, we use RCM, GFP, and ultrasound to visualize mammary gland and mammary tumor characteristics in vivo. We show that RCM can be used to study mammary development in an ex vivo whole organ culture setting with good resolution using a lower concentration of acetic acid. We show that GFP expression can be used to visualize mammary gland ducts, mammary tumor, and tumor vasculature, to follow lobuloaveolar development in an ex vivo whole organ culture experiment, and can be used to follow development of transplanted mammary glands. We show that ultrasound imaging can be used to visualize normal mammary gland, hyperplastic areas of preneoplasia, to follow tumor progression and liver metastases, and can be used to distinguish between mammary tumor and enlarged lymph node. In conclusion, we show that these modalities are, individually and in combination, useful in studying normal and carcinogenic biological processes in the mouse mammary gland longitudinally and with minimal invasiveness.
Methods
Mouse Models, Mammary Gland Whole Mounts, and Hematoxylin and Eosin Sections
Mammary glands from wild-type C57Bl/6 female mice, female mice from a model of Estrogen Receptor alpha (ERα) driven mammary cancer (tTA/TAg/ERα mice) [
40], and female mice from a WAP-TAg mammary cancer model [
41] were examined by different imaging modalities in this study. In general, after imaging was performed, one #4 mammary gland was fixed in formalin overnight, embedded in paraffin, slices (5 μm) mounted on glass slides, and stained with Hematoxylin and Eosin (H&E). For mammary gland whole mount staining, the other #4 mammary gland was fixed in Carnoy's fixative and stained in Carmine-alum as previously described [
42]. Visualization of carmine-alum and H&E stained mammary glands was performed on an Eclipse E800M microscope (Nikon Instruments Inc., Melville, NY, USA). Mammary gland preneoplastic lesions were measured
in situ upon necropsy to compare with measurements taken with the ultrasound software (n = 7). For the metastasis study, the primary mammary adenocarcinoma was removed at 10 months of age. Two weeks after the tumor was removed mammary glands and liver were imaged with ultrasound to screen for the development of new tumors and liver metastases. The mouse was euthanized at 12 months of age because of difficulty breathing and unresponsiveness, 1 month after the original ultrasound and before a second scheduled ultrasound could take place. The presence of liver, lung and omental metastases were confirmed on necropsy. All procedures involving animals were performed in accordance with current federal (National Institutes of Health Guide for the Care and Use of Laboratory Animals) and University guidelines and were reviewed and approved by the Georgetown University Institutional Animal Use and Care Committee.
Reflectance Confocal Microscopy
Mammary gland ductal and epithelial cell morphology from non-pregnant wild-type mice were directly imaged at different stages of development by reflectance confocal microscopy using the VivaCell 5000 Reflectance Confocal Microscope (VivaCell-TiBa, Rochester, NY, USA) with a 30× water immersion lens. Mice were euthanized prior to reflectance confocal imaging. Upon necropsy, the #3 mammary glands were injected with a dilute (5%) acetic acid in phosphate buffered saline (PBS) solution as a contrast agent to enhance visualization of the nuclei within cells by promoting condensation of nuclear material [
43,
44]. The mammary gland was then dissected and spread on the microscope stage above the objective on the Vivacell 5000 and images (originally 500 μm × 500 μm) were taken with the VS2000ui imaging software (version vs006.00.11, Lucid, Inc., Rochester, NY), as described previously [
15,
45]. After imaging, the mammary glands were fixed in formalin for H&E staining as described above.
GFP Imaging
Transgenic tTA/TAg/ERα mice were bred to mice carrying the enhanced Green Fluorescent Protein (GFP) transgene under the control of the chicken beta-actin promoter coupled with the cytomegalovirus (CMV) immediate early enhancer (FVB.Cg-Tg(ACTB-EGFP)B5Nagy/J strain from The Jackson Laboratories, Bar Harbor, Maine).
These mice express GFP in all cells of the mouse expressing actin, especially in the skin and mammary glands [
20]. Due to the nature of adipocytes having low fluorescence, even though they do express GFP, the mammary ductal tree was easily visualized in contrast to the less bright fat. The Nikon SMZ-1500 EPI-Fluorescence Digital Stereoscope System (Melville, NY) was used to visualize GFP-expressing mammary gland and mammary tumor tissues and pictures were taken with the Metamorph imaging software (Molecular Devices Corp, Sunnydale, CA). Using the Apo 1× objective with the 10× eyepiece, the Nikon SMZ-1500 EPI-Fluorescence Digital Stereoscope System has a field of view between 29.3 mm for the 0.75× zoom and 2.0 mm for the 11.25× zoom.
For the in vivo GFP experiment, a 4-week-old GFP mouse was anesthetized and incisions made in the abdominal skin such that the skin could be folded back to allow imaging of the ductal epithelial tree in the #4 mammary gland. After images were acquired as above, the skin was closed with surgical staples and the mouse was allowed to recover. At 8 weeks of age the mouse was euthanized and both #4 mammary glands harvested, placed on glass slides, and imaged by GFP fluorescence.
Mammary Gland Whole Organ Culture
Mammary gland whole organ culture (WOC) was carried out essentially as previously described [
46]. Briefly, 21–24 day old wild-type female mice were anesthetized and subcutaneously implanted in the interscapular area with a 21-day release 0.01 mg 17β-estradiol and 10 mg progesterone (E&P) pellet (Innovative Research of America, Sarasota, Florida) to prime the mammary glands for WOC. After 14 days of priming, the mice were sacrificed and the #4 mammary glands were harvested, placed on a square of cotton mesh and floated in WOC growth phase media, Waymouth's MB 752/1 media (Biosource Biofluids, Rockville, MD) supplemented with Antibiotic-Antimycotic (Gibco/Invitrogen, Grand Island, NY), insulin (5 μg/ml), prolactin (1 μg/ml), aldosterone (0.1 μg/ml), hydrocortisone (0.1 μg/ml) (IPAH). Glands were incubated at 50% oxygen, 5% CO2 in humidified air at 37°C (Heraeus Instruments, Newtown, CT).
A WOC experiment was performed to test viability of the mammary glands after treatment with acetic acid and visualization with RCM. Mammary glands of three-week-old wild-type female mice were primed with an E&P pellet as described above. After 14 days of priming, the mice were sacrificed and mammary glands #3 and #4 were harvested. One gland was immediately placed in IPAH WOC media. Before placing the other mammary glands in WOC, they were treated with PBS, 1, 3, or 5% acetic acid in PBS (3 mammary glands per treatment) and were imaged with RCM. After imaging for no more than 5–10 minutes, the mammary glands were placed in IPAH media as described above. After seven days in culture (post-WOC), the #4 mammary glands were whole mounted and #3 mammary glands were fixed and H&E stained as described above. RCM images from mammary glands treated with PBS or acetic acid pre-WOC were compared with the whole mounts and H&E slides post-WOC.
For the WOC GFP images, the mammary glands from mice expressing GFP alone were treated with IPAH for 10 days. Each day, the mammary glands were visualized with GFP imaging to detect and follow changes and growth in the ductal tree.
Ultrasound
Mice were anesthetized by inhalation of isoflurane with 1–3% oxygen and ventral hair was removed using a mild depilatory cream. Mice were placed on a thermostatically controlled heating pad to help maintain mouse body temperature. A water based ultrasonic gel was applied between the imaging probe transducer and the mouse skin and the liver, mammary tumors, and all ten mammary glands from a total of 16 wild-type, tTA/TAg/ERα, and WAP-TAg mice were imaged with the Visualsonics Vivo 660 High-Resolution Imaging System for small animal ultrasound (Toronto, Ontario, Canada). Transducers were 55 MHz for mammary gland and 40 MHz for liver imaging. Orientation of the mammary gland on ultrasound was accomplished by visualizing the lymph node, which is less echogenic and appears as a black hole surrounded by the echogenic mammary gland tissue. Mammary gland preneoplastic lesions were measured with the ultrasound software using acquired images in the plane showing the largest cross-sectional area of each lesion. These ultrasound measurements were compared with measurements taken at necropsy (n = 7).
Mammary Gland Transplantation
Mammary gland transplantation is a useful technique to study the specific effects of hormone or even genetic influences on mammary gland growth, differentiation, and carcinogenesis. Unfortunately, it is not always possible to distinguish whether the transplanted mammary gland grew or if it was actually the host mammary gland. Sometimes the host mammary gland does grow even if the fat pad of the host is cleared of epithelial cells, which should ensure that the host mammary gland cannot grow. In order to definitively determine the mammary gland of origin and therefore a successful versus unsuccessful mammary gland transplantation, mammary glands from GFP expressing mice were transplanted into non-GFP expressing hosts. If the mammary gland that grew expressed GFP, then the mammary gland was from the transplanted mammary gland (GFP positive). If the mammary gland that grew did not express GFP, then it was residual host mammary gland (GFP negative). Both #4 mammary glands from 1 to 2-day-old newborn female and male pups were removed under a dissecting stereomicroscope (Zeiss Stemi SV 11, Germany) and placed in a culture dish containing DMEM:Ham's F-12 (1:1) with 10% fetal calf serum, 2 mM glutamine and 1% penicillin-streptomycin to keep the glands moist and maintain tissue viability before transplantation. The mammary glands were then transplanted into 3- to 4-week-old female nude mice without clearing the mammary fat pad: an incision was made and a pocket was created between the #3 and #4 mammary glands in the nude mouse and both mammary glands from the newborn pups were introduced in the pocket [
47]. Post-transplantation, a cohort of the host mice were housed with males to become pregnant. Mice that were not housed with males were euthanized 8 weeks after transplantation. For the studies of pregnant transplanted mammary gland, mice that became pregnant were euthanized during late pregnancy (17–19 days pregnancy). For both cohorts, the transplanted glands were removed at necropsy, imaged for GFP expression as described above, and then whole mounted for morphological studies.
Conclusion
The studies presented here compare the well-established techniques of mammary gland whole mounting and hematoxylin and eosin histology with RCM, GFP, and ultrasound to study mammary gland and mammary tumor development. RCM, GFP, and ultrasound are quick techniques that do not require tissue processing for immediate imaging of mammary gland structures. RCM has the potential to advance screening and diagnosis, especially for the early detection of a variety of precancerous lesions [
13,
49‐
51]. We have shown here and previously that RCM is very useful in optical serial imaging of normal mammary gland ductal structures and tumors in harvested tissues from genetically engineered mice [
15]. It has a resolution comparable to the ductal structure resolution of a mammary whole mount and the cellular resolution of mammary histology. This technique can be used for 3-D reconstruction of MG morphology and can be used in living tissue. In this study RCM was also used in combination with whole mammary gland organ culture. Additional RCM applications include using it to excise specific mammary structures for transplant studies. GFP imaging is also useful for
in vivo studies, as well as for whole organ culture and transplantation, where it can be used to follow development and/or disease progression. GFP has been shown to be invaluable in mammary gland transplantation studies where it can successfully answer such questions as whether a growth factor acts in an autocrine or paracrine fashion during mammary gland development by allowing for labeling and following the development of specific mammary cells transplanted into a fat pad [
52]. Ultrasound demonstrates the least cellular resolution and requires an experienced operator to obtain consistent images, but is very useful for
in vivo and non-invasive imaging of development of non-palpable preneoplastic lesions into mammary adenocarcinomas. 3-D ultrasound imaging software can be used to obtain direct measurements of lesion volume, if needed. In summary, all three techniques are valuable adjuvants to the study of mammary development and cancer progression.
The imaging modalities used in this paper, RCM, GFP, and ultrasound imaging, are just a few of the many techniques being developed to study the mammary gland and mammary cancer. These versatile techniques can be combined with each other (i.e. fluorescence and RCM), as well as with other techniques, such as those that involve the detection of specific probes to image targeted cells while simultaneously acquiring confocal contrast images to localize the targeted cells within the histological context of the tissue being imaged [
53,
54]. Combining these techniques allows the researcher to obtain actual
in vivo molecular expression information from the image enabling the study of the molecular basis of initiation and progression of mammary cancer. All of these minimally invasive techniques allow longitudinal imaging to provide complete and precise information about mammary gland development, as well as, tumor initiation and progression in any transgenic cancer mouse model. The further development of mouse imaging techniques may well lead to the advancement of new technologies that can be translated to more sensitively and specifically detect precancerous abnormalities, diagnose curable pre-cancerous lesions, and to increase patient survival and quality of life in breast cancer patients.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MTT performed the RCM whole organ culture experiment, comparison of GFP and whole mount, GFP lymphnode imaging, supervised and performed the ultrasound to follow preneoplasia and tumor progression experiments, and wrote and coordinated the manuscript. ARP performed the RCM of mice during development and during the whole organ culture experiment. IC performed the GFP mammary gland transplant experiment and the comparison of ultrasound and in situ measurements. LPJ performed the GFP whole organ culture experiment. MDJ initially developed the GFP imaging procedures, performed the live GFP imaging, and supervised imaging of whole mammary gland organ cultures, mammary gland transplants, and the mouse mammary tumor. PAF designed and supervised the experiments in collaboration with the other investigators and edited the manuscript. All authors have read and approved the final version of the manuscript.