Skip to main content
NCBI home page
Journal of Histochemistry and Cytochemistry logoLink to Journal of Histochemistry and Cytochemistry
. 2011 Dec;59(12):1087–1100. doi: 10.1369/0022155411423680

Human Breast Cancer Histoid

An In Vitro 3-Dimensional Co-culture Model That Mimics Breast Cancer Tissue

Pavinder Kaur 1,2,3,4,5, Brenda Ward 1,2,3,4,5, Baisakhi Saha 1,2,3,4,5, Lillian Young 1,2,3,4,5, Susan Groshen 1,2,3,4,5, Geza Techy 1,2,3,4,5, Yani Lu 1,2,3,4,5, Roscoe Atkinson 1,2,3,4,5, Clive R Taylor 1,2,3,4,5, Marylou Ingram 1,2,3,4,5, S Ashraf Imam 1,2,3,4,5,✉
PMCID: PMC3283087  PMID: 22034518

Abstract

Progress in our understanding of heterotypic cellular interaction in the tumor microenvironment, which is recognized to play major roles in cancer progression, has been hampered due to unavailability of an appropriate in vitro co-culture model. The aim of this study was to generate an in vitro 3-dimensional human breast cancer model, which consists of cancer cells and fibroblasts. Breast cancer cells (UACC-893) and fibroblasts at various densities were co-cultured in a rotating suspension culture system to establish co-culture parameters. Subsequently, UACC-893, BT.20, or MDA.MB.453 were co-cultured with fibroblasts for 9 days. Co-cultures resulted in the generation of breast cancer histoid (BCH) with cancer cells showing the invasion of fibroblast spheroids, which were visualized by immunohistochemical (IHC) staining of sections (4 µm thick) of BCH. A reproducible quantitative expression of C-erbB.2 was detected in UACC-893 cancer cells in BCH sections by IHC staining and the Automated Cellular Imaging System. BCH sections also consistently exhibited qualitative expression of pancytokeratins, p53, Ki-67, or E-cadherin in cancer cells and that of vimentin or GSTPi in fibroblasts, fibronectin in the basement membrane and collagen IV in the extracellular matrix. The expression of the protein analytes and cellular architecture of BCH were markedly similar to those of breast cancer tissue.

Keywords: 3-dimensional co-culture system, breast cancer histoid, tumor microenvironment

Breast cancer has been recognized to be heterogeneous, representing cancer as well as tumor microenvironment–associated stromal cells with diverse genetic and biological characteristics. The tumor microenvironment generated by the interaction between cancer and the tumor microenvironment–associated stromal cells is now recognized to have a major role in cancer development and progression (Orimo et al. 2001; Shekhar et al. 2001; McAlhany et al. 2003; Desmouliere et al. 2004; Nakagawa et al. 2004; Chung et al. 2005; Galie et al. 2005; Gallagher et al. 2005; Micke and Ostman 2005; Ide et al. 2006; Ohira et al. 2006; Patocs et al. 2007). The results of these studies suggest that the cancer cells may interact with the stromal cells to induce conditions that facilitate tumor growth. The stromal cells in turn are able to modify cancer cells by direct cell-to-cell, contacts through soluble factors or by modification of extracellular matrix (ECM) components, thereby altering the microenvironment to promote the malignant growth (Shekhar et al. 2001; Chung et al. 2005), reduced apoptosis (Desmouliere et al. 2004), induction of angiogenesis (McAlhany et al. 2003), and invasion and metastasis (Nakagawa et al. 2004; Gallagher et al. 2005). Moreover, 3-dimensional (3-D) static co-cultures of normal breast epithelial cells or breast cancer cells alone or with stromal cells, which were grown in collagen gel-matrix, have been utilized to study breast morphogenesis. These 3-D cultures or co-cultures yielded the formation of lumen, apicobasal polarity, and basement membrane (Gudjonsson et al. 2003; Sung et al. 2009; Bauer et al. 2010; Inman and Bissell 2010). The investigators also utilized these models in their pioneer work to study functional aspects of cellular interaction in 3-D culture or co-culture (Gudjonsson et al. 2002; Han et al. 2010; Beliveau et al. 2011; Sung et al. 2011; Yang et al. 2011). Taken together, results of these studies laid the foundation for the feasibility of studying the above complex phenomenon in a controllable environment of cell culture.

The role of stromal cell–derived molecular mediators in cancer progression has also been supported by clinical findings (Orimo et al. 2001; Patocs et al. 2007). In addition, recently reported experimental as well as clinical studies suggest that the development of angiogenesis and tumor growth are stimulated by tumor microenvironment–derived mediators such as the vascular endothelial growth factor (Hicklin and Ellis 2005; Burstein et al. 2008).

Progress in this field of study to unravel the mechanism of heterotypic interaction in cancer progression has been hampered by major drawbacks of most static co-culture systems. Firstly, the xenogenic source of precoated collagen gel–matrix support renders many of the systems inappropriate for the study of human cancer (Shekhar et al. 2001; Selvey et al. 2004; Che et al. 2006). Secondly, insufficient diffusion of nutrients through the gel matrix results in poor cell density. Thirdly, such co-cultures do not yield the cellular architecture of a simulated tissue-like microenvironment with endogeneously produced ECM and basement membrane components. The growing interest in 3-D tumor models and in methods for generating them has resulted in various types of 3-D constructs that are potentially improved models (Seidl et al. 2002; Hsiao et al. 2009). However, the technical aspects of growing multicellular tumor models remain challenging with respect to their 1) morphology that could mimic the tissue-like microenvironment and 2) cellular growth in endogeneously produced growth factors and ECM.

Here, we report the utilization of a 3-D rotary bioreactor to co-culture human breast cancer cells and fibroblasts, which resulted in the generation of breast cancer histoids (BCH). BCH, thus generated, exhibited reproducible expression of cell-specific analytes and cellular architecture that resembled breast cancer tissue, allowing heterotypic cellular interaction and growth as well as the expression of endogeneously produced proteins that are known to be associated with basement membrane and ECM in human breast cancer tissue.

Materials and Methods

Reagents

Rabbit polyclonal antibody to C-erbB.2 (an epidermal growth factor receptor), mouse monoclonal antibodies to pancytokeratins, epithelial cell–associated cytoskeleton proteins (clone AE-1/AE-3); vimentin, a fibroblast-associated cytoskeleton protein (clone V9); p53, an oncogene’s product (clone DO-7); Ki.67, a cell cycling protein (clone MIB-1); and collagen type IV, an ECM-associated protein (clone CIV 22), were purchased from Dako North America Incorporated (Carpinteria, CA). Monoclonal antibodies to E-cadherin, an epithelial-specific intercellular adhesion protein (clone 4A2C7), glutathione S-transferase Pi (GSTPi), a major detoxifier enzyme (clone LW29), and fibronectin, a major cellular basement membrane–associated protein (clone FBN11) were obtained from Zymed Laboratory (South San Francisco, CA), Novocastra (New Castle-upon-Tyne, UK), and LabVision (Fremont, CA), respectively. Normal goat or horse serum, biotinylated antibodies to goat anti-rabbit or horse anti-mouse immunoglobulins, and avidin-biotin-peroxidase (ABC) complex were bought from Vector Laboratories Incorporated (Burlington, CA). The blocking peptides specific for individual primary antibody were obtained from LabVision. The dilution of each antibody was empirically predetermined for optimum staining of target cells in a known positive control, which consisted of FFPE tissue biopsy specimens of breast cancer. All other reagents used were of the highest purity available from Sigma-Aldrich Chemical Company (St. Louis, MO).

Breast Cancer Cell Lines

The rationale for selecting the UACC-893 breast cancer cell line with overexpression of C-erbB.2 for the initial development of BCH was based on the fact that it represents a major subtype of breast cancer with poor prognosis (Meltzer et al. 1991). In addition to UACC-893, two other human breast cancer cell lines, namely BT.20 and MDA.MB.453, were also utilized for the generation of BCH. MDA.MB.453 is known to be positive for moderately membranous expression of C-erbB.2, whereas BT.20 is negative. These cell lines are known to be negative for the expression of estrogen and progesterone receptors (Meltzer et al. 1991). The cell lines employed in this study were obtained from the American Type Culture Collection (Rockville, MD). The cell lines were cultured in DMEM/F12 (1:1 mix; Mediatech Incorporated, Manassas, VA 20109), supplemented with penicillin (100 U/mL), streptomycin (100 µg/mL), and fetal bovine serum (FBS) (10%, v/v).

Preparation of Primary Culture of Human Foreskin Fibroblast

The foreskins were removed during routine circumcision of newborn male infants. The use of tissue specimens was approved by the Institutional Review Board (IRB) of Huntington Memorial Hospital (Pasadena, CA). Primary cultures were established from the foreskin specimens using standard tissue culture methods to obtain fibroblasts for histoid preparations.

3-D Co-culture System

The 3-D rotary bioreactor that was utilized in this study was developed in-house (Ingram et al. 2010). The system supports rotating suspension cultures in low-shear environment to maintain cells in free fall during culture. Each cell undergoes a small trajectory in the suspending liquid, allowing it to contact and communicate with other cells and to interact spontaneously with them in a manner that eventually results in a 3-D structure. The culture chamber was made from a gas exchange membrane (FEP Teflon fluorocarbon; American Fluoroseal Corporation, Gaithersburg, MD).

Co-culture of Breast Cancer Cells and FSF

Freshly harvested FSF were counted and washed. FSF at four different densities (0.5 × 107, 1.0 × 107, 2 × 107, or 4 × 107) were separately suspended in 10 mL of the complete medium. Each cell suspension was injected through 10-mL syringe attached without the needle to the opening port of an individual culture chamber (10-mL capacity). The four culture chambers were mounted on the bioreactor, transferred to a humidified cell culture incubator in 5% CO2 atmosphere, and rotated at 6 rmp for 6 hours. The resulting cultures were individually harvested from the culture chambers, washed, fixed in formalin, processed, and embedded in paraffin (FFPE) as blocks. FFPE blocks were sectioned at 4-µm thickness using a microtome and placed on histological glass slides as previously described (Saha et al. 2007). The representative sections from each of the four blocks were deparaffinized in xylene and rehydrated in decreasing concentrations of alcohol, stained with hematoxylin and eosin (H&E), and microscopically examined. FSF seeded at a density of 0.5 × 107 or 1.0 × 107 yielded individual FSF spheroids with less than 0.5 mm in diameter, whereas those with 4.0 × 107 yielded large clumps of cells. FSF seeded at a density of 2.0 × 107 resulted in the generation of individual FSF spheroids with approximately 1.0 mm in diameter without any detectable clumps of cells. As a result of this observation, the latter density of FSF was chosen for the subsequent co-culture experiments.

Stage 1

FSF (4.8 × 108) were resuspended in 192 mL of freshly prepared complete medium. The cell suspension (8.0 mL containing 2 × 107 FSF) was injected into each of the 24 culture chambers and cultured for 6 hours as described above.

Stage 2

The UACC-893 breast cancer cell line was initially utilized to establish parameters for obtaining co-cultures. UACC-893 at four different densities (3 × 106, 6 × 106, 12 × 106, or 24 × 106) were separately suspended in 12 mL of the complete medium. Six equal aliquots (2 mL each) from each of the above four different cell suspensions, each containing 0.5 × 106, 1.0 × 106, 2.0 × 106, or 4.0 × 106 breast cancer cells per 2 mL, were transferred into six separate culture chambers, which contained unfixed FSF spheroids that were prepared with 2 × 107 FSF as described above. The culture chambers were mounted on the bioreactor, and the co-cultures were maintained for 1, 2, 8, 9, 10, or 12 days, carefully replacing 5.0 mL of spent medium with the freshly prepared complete medium at 24-hour intervals. The resulting co-cultures, containing 0.5 × 106, 1.0 × 106, 2.0 × 106, or 4.0 × 106 breast cancer cells and a predetermined density of FSF (2 × 107), were individually harvested from the culture chambers at various time intervals as indicated above, washed, fixed in formalin, processed, and embedded in paraffin (FFPE) as blocks. A total of 24 blocks were obtained.

Immunohistochemical Evaluation of Co-cultures

The above FFPE blocks of co-cultures were sectioned at 4-µm thickness using a microtome and placed on histological glass slides as described above. The representative unstained sections from each of the 24 blocks (FFPE) were deparaffinized as stated above and subjected to localization of epithelial cell–associated cytoskeleton proteins (pancytokeratins) or a fibroblast-associated cytoskeleton protein (vimentin) by an immunohistochemical (IHC) staining method and microscopically examined as previously described (Saha et al. 2007). Briefly, the unmasking of epitopes of the antigens was performed by heating the histological slides in 0.01 M sodium citrate buffer, pH 6.0, in a microwave pressure cooker for 30 minutes. All the steps of the experiments stated below were carried out at ambient temperature unless otherwise stated. Following the incubation of histological slides with appropriate normal serum for 20 minutes to block nonspecific binding of the primary or secondary antibodies, sections were incubated overnight with 150 µL of mouse monoclonal antibody to pancytokeratins (0.1 µg/mL) or vimentin (0.08 µg/mL) in a humidified chamber. The optimum concentration of each primary antibody was empirically determined to yield the most intense immunostaining of known target cells in breast cancer tissue sections with an undetectable or negligible staining of nonspecific cells and stromal components. Biotinylated horse anti-mouse immunoglobulin antibody as secondary antibody, followed by the ABC conjugate, was applied to the sections. The concentrations of the secondary antibody and the ABC reagents used were as previously determined to be optimum on FFPE tissue biopsy specimens of breast cancer. Diaminobenzidine (DAB) was used as the chromogen and hematoxylin as the counterstain. For each experiment, tissue sections of breast cancer with known expression of pancytokeratins or vimentin served as controls. Moreover, for each experiment, application of preabsorbed primary antibody with the appropriate blocking peptide (10 mg protein/mL of working dilution of each antibody) served as a control to determine specificity of reactivity of each antibody to its target cells in BCH and tissue biopsy specimens of breast cancer.

The cancer cells seeded at a density of 1 × 106 yielded a co-culture preparation with most of the cancer cells showing invasion of FSF spheroids’ core and a minimum trace of free-floating cancer cells on the ninth day in co-culture, as shown in Figure 1D (for details, see Results section). The resulting co-culture with the above characteristics was referred to as BCH. Following the establishment of the procedure of generating BCH with the characteristics that are shown in Figure 1D, three separate batches of BCH preparations in triplicate with a predetermined density of UACC-893 cells (1 × 106) and FSF (2 × 107) per culture chamber and under the identical co-culture conditions for 9 days were generated at 1-month intervals. Moreover, two other breast cancer cell lines, BT.20 (1 × 106) or MDA.MB.453 (1 × 106), were separately co-cultured with FSF (2 × 107) under the identical culture conditions and duration (9 days) in the culture chamber. Each batch of the above BCH preparations was separately harvested from culture chambers, fixed in formalin and embedded in paraffin as blocks, sectioned at 4-µm thickness, and placed on histological glass slides as described above for their IHC characterization. Each of the above BCH blocks yielded approximately 75 sections (4-µm thick each), and each section contained individual BCH constructs, which ranged from 100 to 120.

Figure 1.

Figure 1.

Open in a new tab

Evaluation of UACC-893 breast cancer histoid (BCH). The sections (4-µm thick) of formalin-fixed and paraffin-embedded BCH, resulting from a co-culture of the human breast cancer cell line UACC-893 (1 × 106) and foreskin fibroblasts (FSF) (2 × 107) in the bioreactor for 1, 2, or 9 days, were immunohistochemically stained with mouse monoclonal antibodies to pancytokeratins or vimentin. A representative example of pancytokeratin-positive cancer cells (reddish brown staining, indicated by closed arrows) shows coating of the external layer of an individual pancytokeratin-negative FSF spheroid (indicated by open arrows) at the first 24 hours of co-culture (A). In a consecutive section, an individual FSF spheroid shows reactivity with anti-vimentin antibody (reddish brown staining, indicated by open arrows), whereas the surrounding cancer cells were nonreactive (absence of reddish brown staining, indicated by closed arrows) (B). The pancytokeratin-positive cancer cells (indicated by closed arrows) showed invasion of the core of an individual FSF spheroid (indicated by open arrows) at 48 hours in co-culture (C). Process of invasion of the FSF spheroid by the cancer cells was complete in 9 days in co-culture, as most of the pancytokeratin-positive cancer cells (closed arrows) were detected within the core of an individual FSF spheroid (open arrows) (D). The sections were counterstained with hematoxylin (blue nuclear staining). Original magnification: A–C (×200) and D (×120). Bar = 100 µm.

IHC Characterization of UACC-893 BCH

Intrabatch and interbatch reproducibility of UACC-893 BCH preparations in terms of quantitative expression of a breast cancer cell–associated protein (C-erbB.2), a well-known marker of breast cancer patients with poor prognosis, was determined as described below by IHC staining method and the Automated Cellular Imaging System III (ACIS III; Carl Zeiss Microimaging Incorporated, Aliso Viejo, CA) in breast cancer cells in histology sections from two of the three BCH preparations or blocks from each of the three separate batches of BCH preparations as described below. The number of UACC-893 BCH blocks utilized in this experiment was six. In addition to C-erbB.2, specificity and patterns of expression of other protein analytes, which are known to be associated with epithelial cell, fibroblast, basement membrane, or ECM in tissue biopsy specimens of breast cancer, were compared with those in the remaining one of the three BCH blocks from each of the three separate batches of BCH preparations by the IHC staining method. Three UACC-893 BCH blocks were utilized in this experiment. Moreover, histological sections from each of the two BCH, which were generated by co-cultures of BT.20 or MDA.MB.453 with FSF, were also immunohistochemically analyzed. One hundred fifty microliters of mouse monoclonal antibody to pancytokeratins (0.1 µg/mL), vimentin (0.08 µg/mL), p53 (0.16 µg/mL), Ki.67 (0.15 µg/mL), E-cadherin (2.5 µg/mL), GSTPi (0.4 µg/mL), fibronectin (0.2 µg/mL), collagen type IV (2.0 µg/mL), or rabbit polyclonal antibody to C-erbB.2 (0.5 µg/mL) was individually applied to the first nine histological sections, respectively, from the above-stated BCH blocks. The above antibodies were also individually applied in the same sequence on the remaining sections of each of the five BCH blocks, which consisted of three blocks of UACC-893 BCH, one of BT.20 BCH, and one of MDA.MB.453 BCH. In addition, sections (4-µm thick) from FFPE breast cancer tissue blocks with known reactivity with the above individual antibody were used as positive controls. The determination of optimum concentration of each primary antibody and the rest of the procedure including controls (test of specificity of reactivity of individual primary antibody) was as described above. The immunostained sections of BCH and tissue biopsy specimens of breast cancer were independently reviewed by three of the investigators (R.A., C.R.T., and S.A.I.) throughout this project.

Quantitative Image Analysis of C-erbB.2 Expression in UACC-893 BCH by ACIS III

The rationale for choosing to determine quantitative expression of C-erbB.2 protein as a basis for reproducibility of UACC-893 BCH was governed by its importance as a reliable marker to identify patients having invasive breast carcinoma with poor prognosis and who are likely to respond to treatment with Herceptin (Genentech Incorporated, South San Francisco, CA). Moreover, the procedure of determining quantitative expression of C-erbB.2 by ACIS III (Carl Zeiss Microimaging Incorporated) in breast cancer cells in sections of tissue biopsy specimens for the purpose of prognosis and making decisions for the treatment with Herceptin (Genentech Incorporated) is well established and approved by the Food and Drug Administration (FDA). Quantitative image analysis of C-erbB.2 expression in the cancer cells in histological sections of UACC-893 BCH was conducted by ACIS III (Carl Zeiss Microimaging Incorporated) as described below.

ACIS III (Carl Zeiss Microimaging Incorporated) consists of two major subassemblies. The first is the microscope with its associated electromechanical hardware, and the second is a computer with a frame grabber and image processing system. The microscope subsystem includes the components of a standard microscope (e.g., lamp, condenser, and turret) mounted in a special shock-resistant frame with a video camera. The camera is a 30–frames per second, 1024 × 768–pixel three-chip color camera (JAI CV-M9CL). Also included in the microscope assembly is an in-feed hopper; a stage and motors that provide X, Y, and Z (focus) translation; and an out-feed drawer. The system uses a carrier system, with four histological slides per carrier, and the input hopper is capable of holding 25 carriers. The computer subsystem consists of a dual 3.6-GHz Intel Xeon processor (Santa Clara, CA) running the Windows XP operating system (Microsoft, Redmond, WA). It has 2 GB of memory, 250 GB of RAID level 1 redundant data storage, and a 4.7-GB DVD+RW disc recorder.

For rare event detection, the ACIS III (Carl Zeiss Microimaging Incorporated) makes use of proprietary software, allowing for fast and highly sensitive color detection, along with the capability for the analysis of a variety of morphometric features. The application software available on the ACIS III (Carl Zeiss Microimaging Incorporated) for optical microscopy detection involves first scanning a microscope slide at low magnification (×10). The system next returns to objects that were originally identified for a second analysis at a higher magnification (×40 or ×60). In this case, a more comprehensive image analysis of color and morphometric characteristics (nuclear size and nuclear shape) is undertaken in an effort to exclude cellular debris, large clumps, and cells with morphological features typical of normal hematological mononuclear cells, as opposed to C-erbB.2–positive cancer cells. At the same time, objects that meet color- and morphometry-based criteria as likely tumor cells are collected and presented as montage images for review and classification by a pathologist or other laboratory professional. In the data file generated after specimen analysis, the X-Y coordinates of the object within a framelet are stored. The use of location data results in powerful sample navigation features. A “revisit” capability allows the user to double click on framelets of interest to return to the proper location on the specimen slide for further review under manual control of the microscope. In this mode, it is possible to navigate across the slide, adjust focus, and change microscope objectives. Comparison of the montage images that result from each repeated run is another feature that uses location data. Tumor cells or cell clusters found multiple times by the system can be identified by highlighting framelets with proximate locations as “suspected triplicates.”

The result of IHC staining of sections from two of the three BCH blocks from each of the three separate batches of UACC-893 BCH preparations (approximately 75 sections per block × 6 = 450 sections) revealed a microscopically comparable pattern and intensity of expression of C-erbB.2 protein (results not shown). As a result of this observation, quantitative image analysis of intensity of expression of C-erbB.2 protein in UACC-893 breast cancer cells in sections from BCH blocks was measured by ACIS III (Carl Zeiss Microimaging Incorporated) in every fifth including the first one from a total of 75 sections containing BCH constructs from each of the six BCH blocks (16 sections per BCH block × 6 = 96 sections). The digital images of 96 immunohistochemically stained BCH sections on slides with anti–C-erbB.2 antibody were captured automatically by the ACIS III (Carl Zeiss Microimaging Incorporated) instrument scanner at low magnification, and the whole slide images were viewed on a monitor by one of us (R.A.). R.A. is a diplomat of the American Board of Pathology and is an expert in assessing quantitative C-erbB.2 expression status in breast cancer tissue specimens. Six randomly selected BCH constructs per slide were collected and evaluated for membranous C-erbB.2 expression. The 40× circle tool on the ACIS III (Carl Zeiss Microimaging Incorporated) instrument was used for the selection of areas. After selection of the area, the instrument calculates a region score that was converted to a scoring algorithm consistent with the HercepTest scoring algorithm (Dako North America Incorporated). In accordance with the guidelines, region scores of 0 to 5.0 were given for IHC intensity of staining as described (Minot et al. 2009).

Statistical Analysis of IHC Staining Data

Analysis of variance (ANOVA) was used to determine any significant differences due to batch-to-batch and within-batch variability in histoids and to estimate the pooled variance in order to calculate the coefficient of variation (CV) using measurements on the natural logarithmic scale (Fray 1993). The F test was used to examine whether the measurements are different between experiments. High, low, and average scores were subjected to statistical analyses. The results were similar for each of the three levels of scoring. We report the results for the average score.

Results

Generation of BCH

The sections of formalin-fixed and paraffin-embedded (FFPE) blocks, resulting from co-cultures of a human breast cancer cell line (UACC-893) at various densities (0.5 × 106, 1 × 106, 2 × 106, or 4 × 106) and time intervals (1, 2, 8, 9, 10, or 12 days) with a predetermined density of human foreskin fibroblast (FSF) (2 × 107), were immunohistochemically stained with antibodies to pancytokeratins or vimentin. The immunohistochemically stained sections were microscopically examined to assess the extent of invasion of FSF spheroids by the cancer cells and cellular architecture. One hundred to 120 individual vimentin-positive FSF spheroids per section of co-culture were obtained. The breast cancer cells, following their introduction into the culture chamber, first coated the external layer of individual FSF spheroids on day 1 in co-culture. A representative example of an individual FSF spheroid, which is surrounded by the cancer cells, is shown in Figure 1A and 1B (reddish brown staining). The cancer cells were identified as being positive for cytoplasmic expression of epithelial cell–associated cytoskeleton proteins, pancytokeratins (reddish brown staining, Figure 1A), whereas the FSF spheroids for a fibroblast-associated cytoskeleton protein, vimentin (reddish brown staining, Figure 1B), were in BCH sections. After the initial period of 1 day in co-culture, the cancer cells began to progressively invade the FSF spheroids’ core. A representative example of an individual FSF spheroid with pancytokeratin-positive invading breast cancer cells on day 2 is shown in Figure 1C. The co-culture of 0.5 × 106 breast cancer cells and 2 × 107 FSF resulted in a co-culture with fewer cancer cells showing invasion of FSF spheroids’ core (result not shown), whereas those with 2 × 106 or 4 × 106 cancer cells yielded large clumps of free-floating cancer cells (result not shown). The co-culture of breast cancer cells seeded at 1 × 106 resulted in invasion of most of the FSF spheroids’ core by the cancer cells and a minimum trace of free-floating cancer cells on day 9. A representative example of an individual FSF spheroid with invading breast cancer cells (reddish brown staining) on the ninth day in co-culture is shown in Figure 1D. The process of invasion of FSF spheroids by the cancer cells was complete on the ninth day (Figure 1D), after which no further invasion was detected in co-culture under these conditions (result not shown).

Quantitative Image Analysis of C-erbB.2 Expression in UACC-893 BCH by ACIS III

The expression of C-erbB.2 protein was determined by the IHC staining method in sections (75 sections per BCH block) from two of the three BCH blocks from each of the three batches of BCH preparations, which were obtained by co-culture of UACC-893 and FSF (a total of 6 × 75 = 450 sections). A representative example of an individual BCH construct shows mostly membranous reactivity of anti–C-erbB.2 antibody with UACC-893 breast cancer cells (reddish brown staining, Figure 2A), whereas the fibroblasts (FSF) were nonreactive with the antibody, as indicated by the absence of reddish brown staining (Figure 2A). In breast cancer tissue, anti–C-erbB.2 antibody showed a similar pattern of membranous reactivity with the cancer cells (Figure 2B), whereas stromal cells were nonreactive (Figure 2B). The application of the preabsorbed antibody with specific blocking peptide led to abolition of the immunostaining of the cancer cells in sections of BCH (Figure 2C) and breast cancer tissue (Figure 2D), attesting to specificity of reactivity of the antibody. The morphological appearance and the patterns of membranous expression of C-erbB.2 in breast cancer cells in BCH (Figure 2A) were markedly similar to those of the breast cancer tissue (Figure 2B).

Figure 2.

Figure 2.

Open in a new tab

Immunohistochemical localization of C-erbB.2 protein expression in UACC-893 breast cancer histoid (BCH) and breast cancer tissue. The sections (4-µm thick) of formalin-fixed and paraffin-embedded BCH preparation (A, C), resulting from co-culture of breast cancer cell line UACC-893 (1 × 106) and fibroblast (FSF) (2 × 107) in the bioreactor for 9 days, or human breast cancer tissue (B, D) were immunohistochemically stained with rabbit polyclonal antibody to the C-erbB.2 protein. The antibody exhibited membranous staining of the cancer cells, as indicated by reddish brown staining (closed arrows) in BCH (A) or breast cancer tissue (B), whereas the fibroblasts, as indicated by the absence of reddish brown staining (open arrows), showed no reactivity with the antibody in BCH (A) or breast cancer tissue (B). The application of preabsorbed anti-C-erbB.2 antibody with specific blocking peptide led to abolition of immunostaining of the target cells, as indicated by the absence of reddish brown staining (closed arrow) in BCH (C) or breast cancer tissue (D). The sections were counterstained with hematoxylin (blue nuclear staining). Original magnification: ×310. Bar = 100 µm.

The intensity and patterns of membranous expression of C-erbB.2 protein in breast cancer cells by the IHC staining method were similar in 450 sections, each containing 100 to 120 individual BCH constructs, from six BCH blocks (results not shown). As a result of this observation, quantitative image analysis of intensity of C-erbB.2 expression in cancer cells by the ACIS III (Carl Zeiss Microimaging Incorporated) (Minot et al. 2009) was carried out in every fifth (including the first one) from a total of 75 prestained sections containing BCH from each of the six UACC-893 BCH blocks (16 sections per BCH block × 6 = 96 sections). The quantitative image analysis of intensity of C-erbB.2 expression in cancer cells yielded scores that ranged from 3.0 to 5.0 (standard error = 0.03). The CV was 4.2% for the average score. No significant difference in the score was observed in sections from intrabatches or interbatches of BCH preparations (p=0.61), demonstrating reproducible quantitative expression of C-erbB.2 in breast cancer cells in BCH (Figure 3).

Figure 3.

Figure 3.

Open in a new tab

Quantitative intensity of image analysis of immunochemical expression of the C-erbB.2 protein in the UACC-893 human breast cancer cell line in breast cancer histoid (BCH) by the Automated Cellular Imaging System III (ACIS III). Each block of BCH was obtained from a co-culture of breast cancer cell line (1 × 106) and FSF (2 × 107) for 9 days. Quantitative intensity of image of C-erbB.2 protein expression was analyzed by ACIS III on C-erbB.2–immunostained sections, representing every fifth section from a total of 75 sections from two of three formalin-fixed and paraffin-embedded BCH blocks from each of the three separate batches of BCH preparations (16 sections per block × 6 = 96 sections). The ACIS III quantitates images on a scale of 0 to 5. The bars represent standard error of immunohistochemical expression of C-erbB.2 within the six randomly selected BCH constructs per section. The quantitative expression of C-erbB.2 in the cancer cells in 16 sections from each of the six BCH preparations (blocks) showed scores that ranged from 3.1 to 5.0 (standard error = 0.03). The coefficient of variation (CV) was 4.21 for the average score. No statistically significant difference in the score was observed between the sections from intrabatches or interbatches of BCH preparations (p=0.61).

IHC Characterization of UACC-893 BCH

The sections from the remaining one of the three UACC-893 BCH blocks from each of the three separate batches of BCH preparations were immunohistochemically analyzed for the expression of selected protein analytes. Anti-pancytokeratin antibodies exhibited cytoplasmic reactivity with UACC-893 breast cancer cells in BCH, as indicated by the reddish brown staining (Figure 4A), whereas fibroblasts (FSF) were nonreactive with the antibody, as indicated by the absence of reddish brown staining (Figure 4A). In breast cancer tissue, anti-pancytokeratin antibodies showed a similar pattern of cytoplasmic reactivity with the cancer cells, whereas the stromal cells including fibroblasts were nonreactive (Figure 4B). The application of the preabsorbed antibody to pancytokeratins with specific blocking peptide led to abolition of the immunostaining of breast cancer cells in BCH (Figure 4C) and breast cancer tissue (Figure 4D), as indicated by the absence of reddish brown staining, attesting to specificity of reactivity of the antibody. Anti-vimentin antibody showed a similar pattern of cytoplasmic reactivity with fibroblasts in BCH (Figure 4E) and breast cancer tissue (Figure 4F), as indicated by reddish brown staining, whereas the cancer cells were nonreactive in both instances (Figure 4E and 4F), as indicated by the absence of reddish brown staining. The application of the preabsorbed antibody to vimentin with specific blocking peptide resulted in the abolition of the immunostaining of fibroblasts in BCH (Figure 4G) and breast cancer tissue (Figure 4H).

Figure 4.

Figure 4.

Open in a new tab

Immunohistochemical comparison of the expression of breast tissue–associated protein analytes between UACC-893 breast cancer histoid (BCH) and breast cancer tissue. The sections (4-µm thick) of formalin-fixed and paraffin-embedded BCH (A, C, E, G), resulting from a co-culture of a human breast cancer cell line (UACC-893) and fibroblasts (FSF) in the bioreactor for 9 days, or human breast cancer tissue (B, D, F, H) were immunohistochemically stained. The anti-pancytokeratin antibodies showed cytoplasmic reactivity with the cancer cells, as indicated by the reddish brown staining (closed arrow), in BCH (A) or breast cancer tissue (B), whereas the fibroblasts were nonreactive, as indicated by the absence of reddish brown staining (open arrow) (A, B). The application of preabsorbed anti-pancytokeratin antibodies led to the abolition of immunostaining of the target cells, as indicated by the absence of reddish brown staining (closed arrows), in BCH (C) or breast cancer tissue (D). The anti-vimentin antibody showed cytoplasmic reactivity with the fibroblasts, as indicated by reddish brown staining (closed arrows), in BCH (E) and breast cancer tissue (F), whereas the cancer cells were negative, indicated by the absence of reddish brown staining (open arrows) (E, F). The application of preabsorbed anti-vimentin antibody with specific blocking peptide led to the abolition of immunostaining of the target cells, as indicated by the absence of reddish brown staining (closed arrows), in BCH (G) or breast cancer tissue (H). The sections were counterstained with hematoxylin (blue nuclear staining). Original magnification: ×310. Bar = 100 µm.

Antibody to an oncogene’s product, p53, or a cell cycling protein, Ki-67, exhibited nuclear reactivities with breast cancer cells in BCH (Figure 5A and 5C) and breast cancer tissue (Figure 5B and 5D), as indicated by reddish brown staining. Antibody to an epithelial cell–specific intercellular adhesion protein, E-cadherin, showed membranous reactivity with the cancer cells in BCH (Figure 5E) and breast cancer tissue (Figure 5F). Antibody to a major detoxifier enzyme, GSTPi, showed cytoplasmic and nuclear reactivity with the fibroblasts in BCH (Figure 5G) and breast cancer tissue (Figure 5H), whereas the cancer cells were negative (Figure 5G and 5H), as indicated by the absence of reddish brown staining. Moreover, antibodies to a major cellular basement–associated protein, fibronectin, or an ECM-associated protein, collagen type IV, showed reactivity with the basement membrane component or ECM, respectively, in BCH (Figure 5I and 5K) and breast cancer tissue (Figure 5H and 5L), as indicated by reddish brown staining. The controls that consisted of the application of the preabsorbed antibody to the above analytes (p53, Ki-67, E-cadherin, GSTPi, fibronectin, or collagen IV) with specific blocking peptide led to the abolition of immunostaining of the target cells or stroma in BCH and breast cancer tissue sections, demonstrating the specificity of each antibody’s reactivity (results not shown).

Figure 5.

Figure 5.

Open in a new tab

Immunohistochemical comparison of the expression of breast cancer cells, fibroblast, basement membrane, or extracellular matrix–associated protein analytes between UACC-893 breast cancer histoid (BCH) and breast cancer tissue. The sections (4-µm thick) of formalin-fixed and paraffin-embedded BCH (A, C, E, G, I, K), resulting from a co-culture of a human breast cancer cell line (UACC-893) and fibroblasts (FSF) in the bioreactor for 9 days, or human breast cancer tissue (B, D, F, H, J, L) were immunohistochemically stained. The antibodies to p53 or Ki-67 exhibited nuclear reactivity with the cancer cells, as indicated by reddish brown staining (open arrows), in BCH (A, C) or breast cancer tissue (B, D), whereas the fibroblasts were nonreactive with the antibody, as indicated by the absence of reddish brown staining (open arrows), in BCH (A, C) or breast cancer tissue (B, D). Bar = 100 µm. Antibody to E-cadherin showed membranous reactivity with the cancer cells in BCH (E) and breast cancer tissue (F). Bar = 100 µm. Anti-GSTPi antibody showed cytoplasmic/nuclear reactivity with the fibroblasts, as indicated by brown staining (open arrows), in BCH (G) or breast cancer tissue (H), whereas the cancer cells were nonreactive, as indicated by the absence of brown staining (closed arrows), in BCH (G) or breast cancer tissue (H). The anti-fibronectin antibody exhibited reactivity with the basement membrane component, as indicated by reddish brown staining (closed arrows), in BCH (I) or breast cancer tissue (J). The anti–collagen IV antibody showed reactivity with the extracellular matrix, as indicated by reddish brown staining (closed arrows), in BCH (K) or breast cancer tissue (L). The sections were counterstained with hematoxylin (blue nuclear staining). Original magnification: A–D and G–L (×310), E (×150), and F (×210). Bar = 100 µm.

IHC Characterization of BT.20 and MDA.MB.453 BCH

The sections from each BCH block, generated by co-culture of either BT.20 and FSF or MDA.MB.453 and FSF, were immunohistochemically characterized. Anti-pancytokeratin antibodies exhibited cytoplasmic reactivity with BT.20 breast cancer cells in BCH, as indicated by the reddish brown staining (Figure 6A), whereas fibroblasts (FSF) were nonreactive with the antibody, as indicated by the absence of reddish brown staining (Figure 6A). The antibody showed a similar pattern of cytoplasmic reactivity with the MDA.MB.453 cells in BCH (result not shown). Anti–Ki-67 antibody exhibited nuclear reactivity with BT.20 cells (Figure 6B) and MDA.MB.453 cells (result not shown) in BCH. Anti–C-erbB.2 antibody showed a weak cytoplasmic reactivity with BT.20 cells (Figure 6C), whereas moderately membranous reactivity with MDA.MB.453 cells (result not shown) was in BCH. The patterns of reactivities of the rest of the antibodies with cells in the BCH were similar to those of UACC-893 BCH (results not shown).

Figure 6.

Figure 6.

Open in a new tab

Immunohistochemical expression of breast tissue–associated protein analytes in BT.20 breast cancer histoid (BCH). The sections (4-µm thick) of formalin-fixed and paraffin-embedded BCH, resulting from a co-culture of the BT.20 breast cancer cell line and fibroblasts (FSF) in the bioreactor for 9 days, were immunohistochemically stained. Antibodies to pancytokeratins or Ki-67 exhibited cytoplasmic or nuclear reactivity, respectively, as indicated by the reddish brown staining with BT.20 breast cancer cells in BCH (A, B, respectively) (closed arrow), whereas fibroblasts were nonreactive, as indicated by the absence of reddish brown staining with both antibodies (A, B) (open arrows). Antibody to C-erbB.2 exhibited a weak cytoplasmic reactivity with BT.20 breast cancer cells in BCH (C) (closed arrow), whereas the fibroblasts were nonreactive, as indicated by open arrows. The sections were counterstained with hematoxylin (blue nuclear staining). Original magnification: ×310. Bar = 100 µm.

Discussion

In order to study the in vitro cancer progression, several investigators have initially utilized 3-D culture of cancer cells on either matrigel or laminin-rich ECM with encouraging results (Kenny et al. 2007). Others have focused their investigations on the utilization of 3-D co-cultures of cancer and stromal cells to delineate the mechanism of heterotypic cellular interaction in cancer progression (Rhee et al. 2001; Sung et al. 2008). However, the progress in this field of study has been hampered mainly due to unavailability of an appropriate co-culture system that could facilitate studies of cancer progression in a tumor tissue-like microenvironment.

In this study, we focused firstly on the generation of BCH by co-culturing human breast cancer cell lines (UACC-893, BT.20, or MDA.MB.453) and fibroblasts (FSF) in our rotating 3-D suspension co-culture system; secondly on the quantitative determination of C-erbB.2 protein expression as a parameter of reproducibility of UACC-893 BCH by the FDA-approved ACIS III (Carl Zeiss Microimaging Incorporated); thirdly on the determination of expression of protein analytes that are known to be associated with breast cancer cells, fibroblasts, basement membrane, or ECM in breast cancer tissue; and fourthly on the comparison of BCH with breast cancer tissue in terms of cellular morphology and pattern of expression of the above analytes.

Several batches of BCH that consisted of co-cultures of breast cancer cell lines, representing phenotypically and biologically different subtypes of breast cancer (UACC-893, BT.20, or MDA.MB.453), and fibroblasts were immunohistochemically analyzed. The first parameter selected for analysis of UACC-893 BCH was the quantitative expression of C-erbB.2 protein, which is an important marker of poor prognosis of patients with invasive breast cancer (Slamon et al. 1987). UACC-893 cells in BCH consistently showed batch-to-batch reproducibility of quantitative expression of C-erbB.2, demonstrating the reliability of generating C-erbB.2–positive BCH. A inexhaustible and reliable supply of consistently reproducible C-erbB.2–positive BCH with a tissue-like architecture could be used as an appropriate and much needed IHC reference standard in lieu of breast cancer tissue, which is not readily available to control variations in test conditions at different laboratories (Press et al. 2005). The reliability of IHC assay is paramount to identify C-erbB.2–positive patients who are likely to respond to treatment with Herceptin (Genentech Incorpo-rated) (Singer et al. 2008).

In addition to quantitative determination of C-erbB.2 expression in BCH, other protein analytes, which are known to be associated with breast cancer cells, fibroblasts, basement membrane, or ECM in breast cancer tissue, were also qualitatively evaluated for their specificity and patterns of expression in BCH by the IHC staining method. The breast cancer cells in BCH preparations were homogeneously positive for nuclear expression of a cell cycling protein, Ki-67, demonstrating that these cells were not quiescent at the end of the 9-day co-culture experiments. The frequency of positivity of the cancer cells in BCH was similar to that observed in breast cancer tissue. The known specificity and patterns of expression of these protein analytes in tissue were maintained in BCH under the conditions employed in the 3-D co-culture system. BCH, thus generated, exhibited several distinct cellular features that have not been previously achieved. For example, BCH produced basement membrane–associated proteins, which were readily detectable by IHC staining of sections of BCH. Endogeneously produced collagen IV formed ECM in BCH in a manner that closely resembled that of the breast cancer tissue. Moreover, the stromal components that comprised fibroblasts and ECM were invaded by the cancer cells, resulting in the formation of cellular architecture that morphologically mimicked the breast cancer tissue (Table 1). Interbatch preparations of BCH were reproducible as determined by consistent quantitative expression of C-erbB.2 protein, a breast cancer cell–associated functional analyte. In contrast, the previously reported in vitro breast cancer models were not evaluated for their batch-to-batch reproducibility. The use of rotary bioreactors to carry out co-cultures of breast cancer cells and fibroblasts under defined and stringent culture conditions offered several distinct advantages over the traditional 3-D co-cultures on xenogenically derived gel matrix, which were used in the previously reported studies (Gudjonsson et al. 2002, 2003; Sung et al. 2009; Bauer et al. 2010; Han et al. 2010; Inman and Bissell 2010; Beliveau et al. 2011; Yang et al. 2011; Sung et al. 2011).

Table 1.

Comparison between Huntington Medical Research Institutes (HMRI) and Other 3-Dimensional Co-culture Systems

Characteristics HMRI Co-culture System Other 3-Dimensional Co-culture Systemsa
1. 3-dimensional microenvironment Yes Yes
2. Quality control: levels of expression of cell-specific analytes (e.g., C-erbB.2) as parameters to control variations in preparations of 3-dimensional co-culture Yes No
3. Endogeneously produced expression of human extracellular matrix (e.g., collagen IV) Yes No
4. Endogeneously produced human basement membrane–associated protein expression (e.g., fibronectin) Yes Not reported
5. Mimics tissue-like microenvironment architecture Yes No
Open in a new tab

The BCH with a marked resemblance to human breast cancer tissue and reproducible expression of both structural and functional protein analytes, which are known to be expressed in breast cancer tissue, could be a potentially improved model to study the roles of the tumor microenvironment as well as the significance of heterotypic cellular interaction in breast cancer progression in a controlled, defined, and breast tumor tissue–like microenvironment. The technology developed to generate BCH in this study can, with minor modifications, be applied to generate 3-D co-culture models of other solid tumors.

A study is underway to expand the scope of the BCH model by incorporating breast tumor microenvironment–associated stromal cells, such as cancer-associated fibroblasts (CAF), endothelial cells, or immune cells, in co-culture. Such a model, representing major cellular components of the breast tumor microenvironment, may serve as a basis for the study of molecular mediators of cancer progression, which promote malignant growth, angiogenesis, metastasis, and reduced apoptosis.

Acknowledgments

The authors are grateful to Dr. William Opel, president of Huntington Medical Research Institutes (HMRI), for encouragement and HMRI for financial support. They are also thankful to Ms. Jill Nuccio for data management and Mr. James Kingman for skillfully typing the article.

Footnotes

The authors declared no potential conflicts of interest with respect to the authorship and publication of this article.

The work was supported by Huntington Medical Research Institutes, Pasadena, CA.

References

  1. Bauer M, Su G, Beebe D, Friedl A. 2010. 3D microchannel co-culture: method and biological validation. Integr Biol. 2:371–378 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beliveau A, Mott JD, Lo A, et al. 2011. Raf-induced MMP9 disrupts tissue architecture of human breast cells in three-dimensional culture and is necessary for tumor growth in vivo. Genes Dev. 24:2800–2811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Burstein HJ, Chen Y-H, Parker LM, et al. 2008. VEGF as a marker for outcome among advanced breast cancer patients receiving anti-VEGF therapy with bevacizumab and vinorelbine chemotherapy. Clin Cancer Res. 14:7871–7877 [DOI] [PubMed] [Google Scholar]
  4. Che ZM, Jung TH, Choi JH, et al. 2006. Collagen-based co-culture for invasive study on cancer cells-fibroblasts interaction. Biochem Biophys Res Commun. 346:268–275 [DOI] [PubMed] [Google Scholar]
  5. Chung LW, Baseman A, Assikis V, et al. 2005. Molecular insights into prostate cancer progression: the missing link of tumor microenvironment. J Urol. 173:10–20 [DOI] [PubMed] [Google Scholar]
  6. Desmouliere A, Guyot C, Gabbiani G. 2004. The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol. 48:509–517 [DOI] [PubMed] [Google Scholar]
  7. Fray JC. 1993. Biological Data Analysis: A Practical Approach. Oxford: Oxford University Press [Google Scholar]
  8. Galie M, Sorrentino C, Montani M, et al. 2005. Mammary carcinoma provides highly tumorigenic and invasive reactive stromal cells. Carcinogenesis. 26:1868–1878 [DOI] [PubMed] [Google Scholar]
  9. Gallagher PG, Bao Y, Prorock A, et al. 2005. Gene expression profiling reveals cross-talk between melanoma and fibroblasts: implications for host-tumor interactions in metastasis. Cancer Res. 65:4134–4146 [DOI] [PubMed] [Google Scholar]
  10. Gudjonsson T, Ronnov-Jessen L, Villadsen R, et al. 2002. Normal and tumor-derived myoepithelial cell differed in their ability to interact with luminal breast epithelial cells for polarity and basement membrane deposition. J Cell Sci. 115:39–50 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gudjonsson T, Ronnov-Jessen L, Villadsen R, et al. 2003. To create the correct microenvironment: three-dimensional heterotypic collagen assays for human breast epithelial morphogenesis and neoplasia. Methods. 30:247–255 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Han J, Chang H, Giricz O, et al. 2010. Molecular predictors of 3D morphogenesis by breast cancer cell lines in 3D culture. PLoS Comput Biol. 6:1–12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hicklin DJ, Ellis LM. 2005. Role of the vascular and endothelial growth factor pathway in tumor growth and angiogenesis. J Clin Oncol. 23:1011–1027 [DOI] [PubMed] [Google Scholar]
  14. Hsiao AY, Torisawa YS, Tung YC, et al. 2009. Microfluidic system for formation of PC-3 prostate cancer co-culture spheroids. Biomaterials. 30:3020–3027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ide T, Kitajima Y, Miyoshi A, et al. 2006. Tumor-stromal cell interaction under hypoxia increases the invasiveness of pancreatic cancer cells through the hepatocyte growth factor/c-Met pathway. Int J Cancer. 119:2750–2759 [DOI] [PubMed] [Google Scholar]
  16. Ingram M, Techy GB, Ward BR, Imam SA, et al. 2010. Tissue engineered tumor model. Biotech Histochem. 85:213–229 [DOI] [PubMed] [Google Scholar]
  17. Inman JL, Bissell MJ. 2010. Apical polarity in three-dimensional culture systems: where to now? J Biol. 9:2–5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kenny PA, Lee GY, Myers CA, et al. 2007. The morphologies of breast cancer cell lines in three-dimensional assays correlate with their profiles of gene expression. Mol Oncol. 1:84–96 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. McAlhany SJ, Ressler SJ, Larsen M, et al. 2003. Promotion of angiogenesis by ps20 in the differential reactive stromal prostate cancer xenograft model. Cancer Res. 63:5859–5865 [PubMed] [Google Scholar]
  20. Meltzer P, Leibovitz A, Dalton W, et al. 1991. Establishment of two new cell lines derived from human breast carcinomas with HER-2/neu amplification. Br J Cancer. 63:727–735 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Micke P, Ostman A. 2005. Exploring the tumour environment: cancer-associated fibroblasts as targets in cancer therapy. Expert Opin Ther Targets. 9:1217–1233 [DOI] [PubMed] [Google Scholar]
  22. Minot DM, Kipp BR, Root RM, et al. 2009. Automated cellular imaging system III for assessing HER2 status in breast cancer specimens: development of a standardized scoring method that correlates with FISH. Am J Clin Pathol. 132:133–138 [DOI] [PubMed] [Google Scholar]
  23. Nakagawa H, Liyanarachchi S, Davuluri RV, et al. 2004. Role of cancer-associated stromal fibroblasts in metastatic colon cancer to the liver and their expression profiles. Oncogene. 23:7366–7377 [DOI] [PubMed] [Google Scholar]
  24. Ohira S, Sasaki M, Harada K, et al. 2006. Possible regulation of migration of intrahepatic cholangiocarcinoma cells by interaction of CXCR4 expressed in carcinoma cells with tumor necrosis factor-alpha and stromal-derived factor-1 released in stroma. Am J Pathol. 168:1155–1168 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Orimo A, Tomioka Y, Shimizu Y, et al. 2001. Cancer-associated myofibroblasts possess various factors to promote endometrial tumor progression. Clin Cancer Res. 7:3097–3105 [PubMed] [Google Scholar]
  26. Patocs A, Zhang L, Yaomin X, et al. 2007. Breast cancer stromal cells with TP53 mutations and nodel metastases. N Engl J Med. 357:2543–2551 [DOI] [PubMed] [Google Scholar]
  27. Press MF, Sauter G, Bernstein L, et al. 2005. Diagnostic evaluation of HER-2 as a molecular target: an assessment of accuracy and reproducibility of laboratory testing in large, prospective, randomized clinical trials. Clin Cancer Res. 11:6598–6607 [DOI] [PubMed] [Google Scholar]
  28. Rhee HW, Zhau HE, Pathak S, et al. 2001. Permanent phenotypic and genotypic changes to prostate cancer cells cultured in a three-dimensional rotating-wall vessel. In Vitro Cell Dev Biol Anim. 37:127–140 [DOI] [PubMed] [Google Scholar]
  29. Saha B, Chaiwun B, Tsao-Wei D, et al. 2007. Telomerase and markers of cellular proliferation are associated with the progression of cervical interepithelial neoplasia lesions. Int J Gyn Pathol. 26:214–222 [DOI] [PubMed] [Google Scholar]
  30. Seidl P, Huettinger R, Knuechel R, et al. 2002. Three-dimensional fibroblast-tumor cell interaction causes down-regulation of RACK1 mRNA expression in breast cancer cells in vitro. Int J Cancer. 102:129–136 [DOI] [PubMed] [Google Scholar]
  31. Selvey S, Haupt LM, Tompson EW, et al. 2004. Stimulation of MMP-11 (strometysin-3) expression in mouse fibroblasts by cytokines, collagen and co-cell culture with human breast cancer cell lines. BMC Cancer. 4:40–49 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shekhar MP, Werdell J, Santner SJ, et al. 2001. Breast stroma plays a dominant regulatory role in breast epithelial growth and differentiation: implications for tumor development and progression. Cancer Res. 61:1320–1326 [PubMed] [Google Scholar]
  33. Singer CF, Kostler WJ, Hudelist G. 2008. Predicting the efficacy of trastuzumab-based therapy in breast cancer: current standards and future strategies. Biochim Biophys Acta. 1786:105–113 [DOI] [PubMed] [Google Scholar]
  34. Slamon DJ, Clark GM, Wong SG, et al. 1987. Human breast cancer: correlation of relapses and survival with amplification of the Her-2/neu oncogene. Science. 235:177–182 [DOI] [PubMed] [Google Scholar]
  35. Sung SY, Hsieh C-L, Law A, et al. 2008. Coevolution of prostate cancer and bone stroma in three-dimensional co-culture: implication for cancer growth and metastasis. Cancer Res. 68:9996–10003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sung KE, Su G, Pehlke C, et al. 2009. Control of 3-dimensional collagen matrix polymerization for reproducible human mammary fibroblast cells culture in microfluidic devices. Biomaterials. 30:4833–4841 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sung KE, Yang N, Pehlke C, et al. 2011. Transition to invasion in breast cancer: a microfluidic in vitro model enables examination of spatial and temporal effects. Integr Biol. 3:439–450 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yang N, Mosher R, Seo S, et al. 2011. Syndecan-1 in breast cancer stroma fibroblasts regulates extracellular matrix fiber organization and carcinoma cell motility. Am J Pathol. 178:325–335 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Histochemistry and Cytochemistry are provided here courtesy of The Histochemical Society

RESOURCES