An ex vivo co-culture model system to evaluate stromal–epithelial interactions in breast cancer

Authors

  • T.S. Salameh,

    1. Department of Biological Sciences, Purdue University, West Lafayette, IN
    Search for more papers by this author
  • T.T. Le,

    1. Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN
    Search for more papers by this author
  • M.B. Nichols,

    1. Department of Biological Sciences, Purdue University, West Lafayette, IN
    Search for more papers by this author
  • E. Bauer,

    1. Department of Biological Sciences, Purdue University, West Lafayette, IN
    Search for more papers by this author
  • J. Cheng,

    1. Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN
    2. Department of Chemistry, Purdue University, West Lafayette, IN
    3. Purdue Center for Cancer Research, Purdue University, West Lafayette, IN
    4. Purdue Oncological Sciences Center, Purdue University, West Lafayette, IN
    Search for more papers by this author
  • I.G. Camarillo

    Corresponding author
    1. Department of Biological Sciences, Purdue University, West Lafayette, IN
    2. Purdue Center for Cancer Research, Purdue University, West Lafayette, IN
    3. Purdue Oncological Sciences Center, Purdue University, West Lafayette, IN
    • Department of Biological Sciences, Lilly Hall of Life Sciences, 915 West State Street, Purdue University, West Lafayette, IN 49706, USA
    Search for more papers by this author
    • Tel.: +765-494-0172


Abstract

Breast cancer is the most commonly diagnosed cancer among women worldwide. High breast cancer incidence and mortality rates, especially in obese patients, emphasize the need for a better biological understanding of this disease. Previous studies provide substantial evidence for a vital role of the local extracellular environment in multiple steps of tumor progression, including proliferation and invasion. Current evidence supports the role of adipocytes as an endocrine organ, which produces steroid hormones, pro-inflammatory cytokines and adipokines, such as leptin. To further define the role of the mammary microenvironment on tumorigenesis, we have developed an adipose-tumor epithelial cell co-culture system designed to reproduce the in vivo mammary environment. We validate this model through use of coherent anti-Stokes Raman scattering (CARS) microscopy, a label-free vibrational imaging technique. CARS analysis demonstrates the sustained viability of the adipocytes, and that mammary cancer cell morphology parallels that of tumors in vivo. Also, characterized was the influence of mammary adipose tissue on tumor cell growth and migration. Adipose tissue co-cultured with mammary tumor epithelial cells, in the absence of any serum or supplemental growth factors, resulted in substantial increases in growth and migration of tumor cells. In conclusion, this novel co-culture system provides an ideal model to study epithelial–stromal interactions in the mammary gland. Understanding the relationship between adipose tissue, the most abundant and least studied component of the breast stroma and tumor epithelial cells is critical to clarifying the influence of obesity on the development, progression and prognosis of breast cancer.

Stromal–epithelial interactions play an integral role in regulating tumor development in the mammary gland.1, 2 To address the complexity of these interactions, the establishment of a 3D tissue culture model is essential for understanding the inter-relationship between the epithelial cells and the breast stroma. Previous evidence supports the vital role of the local environment in subsequent steps of tumorigenesis, including proliferation and local invasion.3 To date, the majority of work done in breast cancer, regarding cell–cell signaling mechanisms, has focused on how fibroblasts,4, 5 myofibroblasts, macrophages6, 7 and other inflammatory cells each affect tumorigenesis. Adipocytes, while they are the most abundant cell type surrounding breast epithelial and tumor tissue, are one of the least studied. Until recently, adipocytes were mainly considered an energy storage depot; however, now it is well recognized that adipose tissue acts as an endocrine organ which produces hormones, pro-inflammatory cytokines, adipokines and other molecules that affect normal development, as well as tumor growth and metastasis.8–10 Along these lines, a recent report demonstrates that adipocytes enhance the metastatic properties of breast cancer in vitro and in vivo through the release of inflammatory proteins.11 Although these studies highlight the importance of individual components of the stroma on tumorigenesis, the majority of these previous experiments employ cell lines which can behave differently from stromal cells existing in a normal environment.

A limited number of studies have used mammary-derived adipose tissue explants to gain insights into stromal–mammary epithelial interactions. A primary advantage of using explants is that the cells remain in their natural extracellular matrix environment and therefore can maintain in vivo functionality. Furthermore, cells may be studied and manipulated in their microenvironment context or subsequently isolated and evaluated individually. For example, Beck et al. have shown mouse mammary fat pads cultured in serum-free medium, supplemented with growth factors, release factors that stimulate growth of CL-S1 cells, a mammary epithelial cell line derived from a pre-neoplastic outgrowth, in vitro.12 In their study, conditioned media was prepared from the mammary fat pad and was supplemented with various growth hormones. Their results indicated that cell proliferation was stimulated by the mammary fat pad. A similar study by Hovey et al. demonstrated that factors from the mammary fat pad modulated the ability of COMMA-1D cells, a mouse mammary epithelial cell line, to respond to IGF-1 and insulin in vitro.13 In that study, epithelial-free fat pad explants were placed on siliconized lens paper and co-cultured with COMMA-1D cells, in hormone-amended media. Their findings indicate that murine fat pads are the source of a diffusible mitogenic activity that markedly enhances the growth factor-stimulated proliferation response of mammary epithelial cells.

The use of mammary explant models has revealed important fundamental aspects of the importance of adipose tissue in epithelial growth; however, these systems can have limitations regarding quality of their in vivo reflection of epithelial–stromal interactions. As classically employed, explant studies use porous lens paper, with poorly defined pore sizes, leading to inconsistencies in diffusion of paracrine/autocrine factors, necessitating addition of growth factors to cell-culture media. To circumvent these inconsistencies, we utilized a co-culture system that is physiologically parallel to the in vivo mammary gland/tumor microenvironment. This system is based on the premise of two major coexisting compartments of mammary gland tissue, the stroma and epithelium, which are separated by a basement membrane that allows the passage of large and small molecules. In the co-culture system described herein, a membrane insert of defined size is used to act as a basement membrane, allowing for a more consistent molecular communication between the stroma and the tumor cells. Different from previous explant work focusing on mammary development; in these experiments our co-culture model was used to analyze the effect of the mammary stroma on mammary tumor epithelial cell proliferation and migration. In addition, the technique of coherent anti-stokes Raman scattering (CARS) microscopy was utilized for characterization of this system.14, 15 CARS microscopy is a technique for high-sensitivity vibrational imaging of live cells, in the absence of cell labeling dyes, for 3D visualization of live cell–cell interactions occurring between intact mammary glands or epithelial-free mammary fat pads and rat mammary tumor epithelial cells.

The goal of our studies is to develop an ex vivo co-culture model to facilitate our understanding of the relationship between the intact mammary adipose stroma and the tumor epithelial cells. Unraveling this crosstalk is especially important considering the current pandemic of obesity in adults and children,16 and its association with the pathogenesis of many diseases including breast cancer.17 The mechanisms by which increased adiposity contributes to breast cancer incidence, progression and morbidity are not understood. Thus, this co-culture system, coupled with CARS imaging techniques, provide important tools for evaluating and defining the molecular details of this relationship.

Material and Methods

Mammary tumor cell culture

CRL1743 cells, a rat mammary N-methylnitrosourea (MNU)-induced tumor-derived epithelial cell line, were used for these experiments (ATCC, Manassas, VA). Cells were cultured in complete medium, which consisted of Eagles minimal essential media (EMEM) (ATCC, Manassas, VA) supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/mL) and streptomycin (0.1 mg/mL). All cultures were maintained at 37°C under 5% CO2. No additional hormonal or growth factors were added to the medium. Cells for proliferation experiments were harvested using 1× Trypsin containing EDTA (Invitrogen, Carlsbad, CA) and plated in complete medium into 12-well culture plates at a density of 4 × 104 cells/well. After 24 hr, cells were placed in serum-free media for synchronization. Cell proliferation was determined at 24, 48 and 72 hr under the following experimental conditions: cells in serum-free media, cells in complete media, cells with epithelial-free fat pad in serum-free media, cells with epithelial-free fat pad in complete media, cells with intact mammary gland in serum-free media and cells with intact mammary gland in complete media. At the end of each experimental period, cells were harvested and counted using a hemocytometer. Cells from three wells of the 12-well plate were combined for an n = 1. An n = 3 was achieved in each experiment and each experiment was completed at least three times.

Preparation of epithelial-free rat mammary fat pads

Virgin female Sprague-Dawley rats (Harlan, Indianapolis, IN), between 21 and 24 days of age, were cauterized at the 4th pair of mammary gland on the right side to remove the developing epithelial component as described previously.18 The left side of the 4th mammary glands was left intact as an intact control. The 4th pair was chosen because of its accessibility in comparison to the other glands; however, all glands can potentially be used.19 Animals were housed in the Purdue University animal facility. All rats had free access to food and water and they were maintained at 25°C on a 12-hr light/12-hr dark cycle. Animal procedures were approved by the Purdue Animal Care and Use Committee.

Co-culture of mammary tissue with CRL1743 tumor cells

At least 60 days postsurgery rats were sacrificed by CO2 inhalation and cervical dislocation. Lymph nodes were removed from the epithelial-free fat pad and the intact mammary gland. Intraperitoneal fat was also isolated from each animal. Isolated tissues were washed in serum-free media. Tissues were cut into explants of approximately 1 × 1 × 1 mm3 pieces. Explants were placed onto cell-culture inserts (BD Biosciences; 3.0 micron pore size, high-density membrane) so that 80% of the surface area was covered by tissue. The tissue pieces placed across inserts of replicate wells were randomized. Prior to addition of tissue, inserts were equilibrated with serum-free media. Media was changed to either serum-free or complete media (FBS, 10%) dependent on the experimental condition. Experiments with mammary tissues were conducted for a period of 24, 48 or 72 hr. Experiments with intraperitoneal fat were conducted for a period of 24 and 72 hr. At the conclusion of each time point, cells were trypsinized and counted with a hemocytometer. A schematic diagram of the co-culture system is shown in Figure 1. This diagram has been adapted from a figure prepared by Walden et al.20

Figure 1.

A schematic diagram of the mammary tissue and tumor cell co-culture system. Virgin female Sprague–Dawley rats are cauterized at the 4th pair of mammary glands at 21–24 days of age to generate an epithelial-free mammary fat pad. One gland is left intact to serve as an internal control. After approximately 60 days postsurgery, the intact mammary gland and the epithelial-free fat pad are isolated and cut into 1 mm3 sections. These sections are randomly distributed onto cell culture inserts for the examination of CRL1743 rat mammary tumor epithelial cell proliferation and migration.

Migration assay

Cell migration was assessed using an established wound healing assay as described previously.21 Briefly, CRL1743 cells were plated in 12-well plates with complete growth media. When cells were 80–90% confluent, they were cultured in serum-free medium for 24 hr to synchronize cell growth cycle. The confluent monolayer of cells was then wounded by scraping a narrow object (pipette tip, 10 μL) down the center of the well. Cells were then placed in one of the six experimental groups; cells in serum-free media, cells in complete media, cells with epithelial-free fat pad in serum-free media, cells with epithelial-free fat pad in complete media, cells with intact mammary gland in serum-free media and cells with intact mammary gland in complete media. Images were taken of the wound at the time of infliction using a VistaVision inverted microscope (VWR, Westchester, PA) and at 18 hr postinfliction. Each wound was examined by phase contrast microscopy for the amount of wound closure by measuring the physical separation remaining between the original wound widths using Adobe Photoshop CS3. Three separate measurements were made per well, and each experiment was performed at n = 5–6.

Coherent anti-stokes Raman scattering microscopy

A multimodal microscopy technique has been described previously15 for evaluating rat mammary tissue and tumors. Briefly, the wave number difference ωp − ωs was tuned to 2,840 cm−1 for CARS imaging, which matches the Raman shift of symmetric CH2 stretch vibration. The same CARS laser sources were used for two-photon imaging. CARS and two-photon signals were collected through a 600/65 nm (Ealing Catalog; Rocklin, CA; Cat. No. 42-7336) and a 520/40 nm (Chroma; Rockingham, VT; Cat. No. HQ520/40) bandpass filters, respectively. The combined laser power at the sample was kept constantly at 40 mW.

Rat mammary tumor tissue

The mammary tumor tissue used for CARS imaging was collected from MNU-induced tumors in Sprague–Dawley rats. This animal study has been described previously.15 Mammary tumor tissue was cut into 2 × 2 × 2 mm sections and maintained in 10% FBS EMEM for a period of no >24 hr before imaging.

Adipocyte purification

Rat mammary tissues were incubated with 1 mg/mL collagenase (Cat. No. C2139, Sigma-Aldrich, St. Louis, MO) in DMEM supplemented with 10% FBS for 1 hr at 5% CO2 and 37°C with occasional stirring. Treated tissues were centrifuged at 125g for 5 min. Adipocytes floating in the supernatant were collected and concentrated through 25-μM sterile filters.

Isolectin-B4 staining of macrophages

Mammary tissues were incubated with 10 μg/mL of isolectin IB4 conjugated with FITC (Cat. No. L2895, Sigma-Aldrich, St. Louis, MO) diluted in minimal DMEM media for 4 hr. Stained tissues were washed thoroughly with minimal DMEM media to remove unstained excess dyes. Then, tissues were imaged with two-photon fluorescence microscopy using excitation wavelength at 890 nm and 510/40 nm detection filter.

DiOC18 staining of mammary adipose tissue

Mammary adipose tissue was incubated with 100 μM of 3,3′-Dioctadecyloxacarbocyanine Perchlorate (DiOC18), a green fluorescent dye used to stain the plasma membrane (Molecular Probes, Eugene, OR), diluted in DMEM media, for 1 hr in a 37°C incubator with 5% CO2. Stained tissues were washed thoroughly with 1× PBS to remove unstained excess dyes. Tissues were then imaged with CARS and two-photon excitation fluorescence (TPEF) as described previously.15

Statistical analysis

Data represented here are displayed as means ± standard deviation. Differences in data between the groups were compared to Student's paired two-tailed t-test. A p-value of <0.05 was considered statistically significant.

Results

Intact mammary gland and epithelial-free fat pad induce tumor cell growth in the absence of serum

Intact mammary glands and epithelial-free fat pads were isolated from rats at least 60 days postclearing surgery and used for co-culture. Tissues were co-cultured with the rat mammary cancer cell line CRL1743 in 10% serum supplemented media for 24–72 hr (Fig. 2a). No additional hormonal or growth factors were used in these studies. As a control, CRL1743 cells were cultured in the absence of tissue but in the same serum-augmented media. When CRL1743 cells were cultured in the presence of either intact mammary gland or epithelial-free fat pad, a significant increase in cell number occurred over the 3-day time period. At 24 hr, CRL1743 cell growth with serum-supplemented media increased (4.67 × 105 ± 1.24 × 104) at a level comparable to the increase observed in the presence of the intact mammary gland (5.13 × 105 ± 1.64 × 104) and in the presence of the epithelial-free fat pad (4.55 × 105 ± 2.4 × 104). At 48 hr, the intact mammary gland (1.63 × 106 ± 5.68 × 104) and the epithelial-free fat pad (1.67 × 106 ± 9.96 × 104) induced significantly higher CRL1743 cell growth compared to CRL1743 cells (1.33 × 106 ± 6.4 × 104) cultured in the absence of tissue. This suggests an autocrine/paracrine feedback to enhance the growth rate of cancer cells at 48 hr, regardless of the presence or absence of epithelial cells in mammary tissue. At 72 hr, CRL1743 cells in the absence of tissue increased to 2.92 × 106 ± 1.12 × 105, which was significantly higher compared to cells co-cultured with either the intact mammary gland (2.29 × 106 ± 1.35 × 105) or the epithelial-free fat pad (2.53 × 106 ± 1.15 × 105). There was no significant difference in tumor cell growth between the intact mammary gland and the epithelial-free fat pad at any time point evaluated. The importance of the presence of mammary fat tissue is further shown in Figure 2b. CRL1743 cells in serum-free media, in the absence of tissue, resulted in a decline in cell number over the 3-day time period as anticipated. Conversely, a significant increase (p < 0.05) in cell number was observed over the 3-day time period when CRL1743 cells were co-cultured with either intact mammary gland or epithelial-free fat pad in serum-free media. When cultured in the absence of serum and tissue, CRL1743 cell numbers decreased from 1.17 × 105 ± 2.96 × 104 at 24 hr to 5.54 × 104 ± 2.61 × 104 at 72 hr. However, when cells were co-cultured in the presence of intact mammary gland or epithelial-free fat pad, a significant increase in mammary tumor cell number occurred at every time point when compared to the serum-free control. For example, at 48 hr in co-culture, the intact mammary gland and epithelial-free fat pad stimulated increased tumor cell numbers to 6.43 × 105 ± 7.84 × 104 and 6.08 × 105 ± 6.14 × 104, respectively, which is significantly higher than the serum-free control tumor cells (7.75 × 104 ± 2.28 × 104). No significant differences in tumor cell growth were found regardless of whether they were cultured in the presence of the intact mammary gland or the epithelial-free fat pad at any time point measured. It is noteworthy that the substantial increase in cell growth occurs equally when cells were cultured with the intact mammary gland as well as the epithelial-free fat pad, despite the absence of serum or addition of any external growth factors. Furthermore, growth patterns appeared optimal under both conditions, as there is close to a doubling of tumor cell number approximately every 24 hr. These data demonstrate that the epithelial-free fat pad alone is sufficient to sustain cancer cell growth.

Figure 2.

Proliferation and migration of mammary cancer cells co-cultured with the intact mammary gland or epithelial-free fat pad. Synchronized CRL1743 cells were co-cultured for 24–72 hr in (a) media containing 10% serum or (b) serum-free media, in the presence of either an intact mammary gland or epithelial-free fat pad. As a control, CRL1743 cells were incubated in the presence or absence of serum without the addition of tissue. At the end of each time period, CRL1743 cells were trypsinized and counted. Data represent an n = 9–14 from three to four different rats. Intraperitoneal fat (c) was co-cultured with CRL1743 cells, in the absence of serum, for 24 or 72 hr. Data represent an n = 6–12. CRL1743 cell migration (d) was quantified by wound healing assays. At t = 0 hr a 10-μL tip was used to scrap away a section of cells and photographed. CRL1743 cells were co-cultured, with or without serum, in the presence or absence of the epithelial-free fat pad or intact mammary gland, and photographed again at t = 18 hr. Data represent an n = 5–6 from three animals. Values are expressed as means ± SD. Asterisk denotes a significant difference (p < 0.05) when compared to the control at each particular time point. #A significant difference (p < 0.05) between the experimental groups at a particular time point.

Intraperitoneal fat elicits a similar tumor cell growth response as mammary adipose tissue in co-culture

To determine if the effect of mammary fat tissue on tumor cell growth was fat tissue specific, parallel co-culture experiments were conducted using intraperitoneal fat collected from the same Sprague–Dawley rats that had undergone the mammary fat pad clearing protocol (Fig. 2c). Intraperitoneal fat was handled identically to mammary tissue. Briefly, at 24 hr intraperitoneal fat stimulated CRL1743 cell growth (4.68 × 105 ± 3.12 × 104), which was comparable to that seen in the presence of the epithelial-free mammary fat pad (3.69 × 105 ± 2.75 × 104). Interestingly, at 72 hr the rate of tumor cell growth stimulated by the intraperitoneal fat (1.03 × 106 ± 8.59 × 104) was significantly lower (p < 0.05) compared to the epithelial-free fat pad (1.33 × 106 ± 4.62 × 104). These results imply each fat tissue depot can support tumor cell growth; however, the underlying factors responsible for this difference remain to be elucidated and warrant further investigation.

Mammary adipose tissue stimulates tumor cell migration in co-culture

To further study the sustainability of mammary adipose tissue on tumor cell behaviors, the influence of mammary tissue on CRL1743 migration was tested. Wound-healing assays were conducted on CRL1743 cells co-cultured in the presence of either intact mammary gland or epithelial-free fat pad with and without 10% serum (Fig. 2d). Tumor cell images were taken at t = 0 hr and then again at t = 18 hr. This latter time point was chosen based on the 24 hr CRL1743 cell doubling time to ensure that our observations are the result of tumor cell migration and not proliferation. Each combination of mammary tissue/fat pad explants, supplemented or without serum, was able to induce CRL1743 migration (p < 0.05) as compared to CRL1743 controls incubated without serum and tissue. Each experimental group containing mammary tissue was not significantly different from CRL1743 cells cultured with 10% serum. Also, no significant difference in migration induction existed between cells cultured with the intact mammary gland and the epithelial-free fat pad. The CRL1743 cell line is not highly migratory in nature, and thus the inability of adipose tissue to induce migration above the level of CRL1743 cells incubated with 10% FBS is not surprising. These results do demonstrate the mammary fat pad with and without the presence of mammary epithelial cells is sufficient to sustain mammary tumor cell migration even in the absence of serum.

Adipocyte stability throughout the time course of co-culture experiments is verified using CARS

Coherent Anti-stokes Raman Scattering (CARS) microscopy imaging techniques were used to determine the stability of the adipocytes in the intact mammary gland and the epithelial-free fat pad. 3D imaging (Fig. 3) reveals viable intact lipid containing adipocytes with distinct well-defined cell borders. For the time course observed, there were no differences in adipocyte morphology when comparing across or within mammary tissue types. Morphological characteristics indicative of apoptosis are not visible in the adipocytes until at least 14 days in co-culture (data not shown). Adipocyte visualization by CARS was verified with DiOC18 staining (Fig. 4). These findings establish our ability to rapidly and easily image mammary adipocytes with CARS and indicate adipocytes remain viable well beyond the time course employed in our tumor cell proliferation and migration studies.

Figure 3.

CARS ex vivo imaging of rat mammary adipocytes and tumors. The intact mammary gland and epithelial-free fat pad were imaged for a period of 96 hr to show stability of adipocytes during the time course of the co-culture studies. Adipocytes in both the intact (a, b) and the epithelial-free mammary gland (c, d) show no morphological apoptotic features between one another, at the start or throughout the study time course. CRL1743 cells in co-culture for 96 hr with the epithelial-free fat pad (e) and mammary tumor tissue from an MNU-treated Sprague–Dawley rat (f). Lipid droplets can be seen around the periphery of the epithelial cells in both panels. CRL1743 cells were imaged at 96 hr in co-culture to obtain the correct density needed for CARS imaging. Data shown here are representative of results obtained from three rats. Scale bar is 25 μm.

Figure 4.

Verification of CARS adipocyte detection by DiOC18 staining. Images of mammary adipocytes taken with (a) CARS (red) and (b) TPEF (green). Overlay of CARS and TPEF images (c). Images taken with a 60× water immersion objective. Scale bar is 25 μm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Co-culture model parallels in vivo rat mammary tumor environment

To demonstrate that this model is similar in morphology to the in vivo environment, a comparison of the in vitro model with a rat mammary tumor was performed. Figure 3 compares CARS images of CRL1743 cells in co-culture (Fig. 3c) and a freshly dissected rat mammary tumor (Fig. 3f). These images demonstrate lipid droplets form around the periphery of CRL1743 rat epithelial tumor cells in vitro in the same manner as they form around rat tumor epithelial cells in vivo. CARS analysis not only identified viable adipocytes, but was also able to detect macrophages, indicating a physiologically active tissue (Figs. 5a5i). Macrophages were observed at day 10 (Fig. 5c) when the epithelial-free fat pad was cultured without CRL1743 tumor epithelial cells, in growth media containing 10% FBS and no additional growth factors. When epithelial-free fat pads were co-cultured with CRL1743 cells, macrophages were evident at day 4 (Fig. 5e). When adipocytes were purified from the epithelial-free fat pad and cultured in the absence of mammary epithelial cancer cells, no macrophages were detected (Figs. 5g5i). Tissue-associated cells were verified as macrophages by staining with Isolectin B4-FITC, which binds to cell membrane glycoconjugates bearing terminal alpha-D-galactose, which macrophages possess (Fig. 6). We found that CRL1743 mammary cancer cells promote lipid secretion from mammary tissues in co-culture (data not shown). We also observed lipids from mammary tissues are taken up by macrophages, leading to an earlier appearance of lipid-rich macrophages in co-cultures. Along these lines, mammary tissue macrophages were not observed on day 0 with CARS, but were detectable with fluorescence labeling and two-photon imaging (Fig. 6). However, as macrophages accumulate cytoplasmic lipid droplets, an indication of activation, they become visible with CARS imaging.

Figure 5.

CARS imaging of epithelial-free fat pad in co-culture with CRL1743 cells reveals the presence of macrophages. (ac) When the epithelial-free fat pad was cultured in the absence of tumor epithelial cells, macrophage detection occurs after 10 days in culture. (df) When the epithelial-free fat pad was cultured in the presence of tumor epithelial cells, macrophage detection occurs after 4 days in culture. (gi) Adipocytes were purified from the epithelial-free fat pad and cultured in the absence of CRL1743 tumor epithelial cells. No macrophages were observed. Scale bar is 25 μm.

Figure 6.

Macrophage detection in explanted mammary tissues stained with Isolectin B4-FITC. Isolectin B4-FITC conjugate was used for staining on day 0 after tissue extraction. Isolectin B4-FITC also stains endothelial cells of blood vessels and hence the visualization of microvessels. Images were taken with two-photon fluorescence. Arrows indicate blood vessels and macrophages. Scale bar is 25 μm.

Discussion

It has become increasingly apparent that the microenvironment plays a vital role in tumorigenesis.22, 23 The mammary gland microenvironment is composed of many components, of which adipocytes are the most abundant and the least studied. Work done by Beck et al. and Hovey et al. has shown the importance of the mammary fat pad on normal mammary epithelial growth. These early studies did not investigate interactions between the mammary fat pad and the tumor epithelial cells; however, they did emphasize the significance of using a system where adipocytes remain in their native microenvironment to study the effects of mammary stroma. Supporting the idea that adipocytes function optimally when cultured in the presence of a structural microenvironment, Stacey et al. established a preadipocyte/adipocyte co-culture technique using hydrogels of polyethylene glycol as scaffold material.24 When scaffold properties, culture dimensionality and culture media treatments were investigated, an up regulation of adipogenesis was observed in co-culture, compared to traditional differentiation media techniques. This report reveals adipose engineering requires multicomponent environments containing bioactive molecules. Recognizing the critical influence of the adipocyte milieu, and toward studying the important interrelationship between adipocytes and tumor epithelial cells, we have developed a new ex vivo co-culture model system that mimics the natural in vivo mammary tumor microenvironment.

In characterizing our co-culture system, we gained valuable insights regarding effects of adipose stroma on mammary tumor cells and vice versa. We demonstrate mammary tumor cells remain viable and show substantial and comparable growth, in the presence of either the intact mammary gland or the epithelial-free mammary fat pad. This observation indicates increased tumor cell growth is primarily the result of the mammary fat pad and not of normal epithelial cells within the intact mammary gland. Herein, we also provide the first evidence suggesting adipose tissue from different regions can directly and disparately affect mammary tumorigenesis. In our study, we demonstrate that adipose tissue from different depots each promoted mammary tumor cell growth, but that intraperitoneal adipose supports a tumor cell growth pattern that is distinct from mammary fat. This coincides with the previous reports where adipose derived from different regions, that is subcutaneous vs. intraperitoneal, can maintain unique features such as hormone secretory profiles.25, 26. Furthermore, a depot-specific impact on tumor growth is in accord with the association between breast cancer risk and adiposity at particular sites, such as central or abdominal obesity.27, 28 This observation calls attention to the possibility that discrete adipose–tumor cell interactions can occur based on adipose type and elucidating differences in molecular interplay may uncover the role of adipose on divergent stages of tumor progression.

It is well established that mammary adipose is a rich source of adipokines, lipids and growth factors, and that these factors can influence tumor growth and aggressiveness.29, 30 In developing this co-culture system, our goal was to create a more physiologically relevant culture model to unravel this multicomponent crosstalk. In these initial experiments, we focus on evaluating the effect of the fat pad on tumor invasiveness and characterizing the impact of tumor cells on mammary fat. Indeed, we demonstrate that the mammary fat pad, in the absence of serum, promotes mammary epithelial tumor cell migration. Subsequently, we show that our co-culture model is reflective of the morphology of the in vivo rat mammary tumor environment. Through use of CARS microscopy, we demonstrate adipose tissue was not only viable but maintained morphological integrity throughout the time course of our tumor cell analysis experiments. Comparison between the in vitro co-culture model and the in vivo mammary tumor stroma detailed similar lipid droplet formation around the periphery of epithelial cells in both environments. Furthermore, CARS analysis showed stability of mammary adipocytes and viability and activity of mammary macrophages in co-culture. Interestingly, mammary macrophages were detected several days earlier when fat pads were co-cultured with mammary tumor cells, as compared to culture in the absence of tumor cells. As CARS microscopy is exquisitely sensitive for imaging lipids,14, 31 this detection can be explained by increased lipid accumulation within macrophages. Given that lipid accumulation is a marker for macrophage activation, these observations demonstrate that CARS is a valuable technique for quantifying macrophage function in tissue. In a previous report, we used an in vivo model to convey that CARS, combined with other nonlinear optical methods, is a rapid effective means for evaluating the impact of obesity on components of the mammary microenvironment, including adipocyte, blood vessels and extracellular matrix proteins. In advancing our study, herein we describe the utility of CARS microscopy for evaluating the direct influence of tumor cells on the mammary microenvironment. Ultimately, a more comprehensive description and quantification of the phenotypic changes in the mammary stroma combined with improved insights into the intermolecular crosstalk will be critical for better understanding tumor–stroma interactions.

In summary, we have developed a unique co-culture system to study the influence of the intact mammary gland or the epithelial-free mammary fat pad on tumor epithelial cells, without the need for any serum or exogenous growth factor media supplements. In conjunction with CARS imaging, we illustrate the reconstructed mammary tumor environment in this model reflects the in vivo mammary tumor microenvironment. We also show mammary adipose tissue alone is sufficient to promote tumor cell migration and the growth and viability of tumor cells for extended periods of time. In brief, this system provides a tremendous advantage compared to models of mammary epithelial–stromal interactions requiring confounding media supplements. It can be exploited not only to study the effects of mammary adipose on tumor cell behavior, signaling and metabolism but also to delineate the physiological significance of local stromal-derived growth factors, fatty acids and extracellular matrix proteins on tumorigenesis. Furthermore, in contrast to co-culture methods employing isolated adipocytes11–13 this model can contribute novel biochemical, physiological and 3D morphological insights regarding the role of tumor cells in modulating stromal microenvironment components. Importantly, we put forth the premise that this co-culture protocol is applicable to other tumor cell and adipose tissue types. Thus, an invaluable method is presented capable of generating physiologically meaningful information without the extensive costs of in vivo work. In addition, this technique can be useful for testing the effect of individual agents on epithelial–stromal interactions, and provides an excellent intermediate, between in vitro and animal studies, for screening early-stage cancer drugs candidates. Ultimately, the coupling of co-culture ex vivo models with advanced imaging will facilitate our understanding of the role obesity plays in the development, progression and prognosis of cancer and can expedite identification of new targets and strategies for breast cancer treatment.

Acknowledgements

The authors are grateful for Dr. Charles Rehrer and Dr. Michael Karadimos for their help in reviewing the manuscript.

Ancillary