Background
Breast cancer is the most frequent and the second deadliest neoplasm affecting females in the Western world [
1]. Despite the enormous advancement in breast cancer therapy in the last years, a better understanding of the mechanisms involved in breast carcinogenesis and progression may lead to novel therapies and increased patient survival.
For decades, processes involved in malignant transformation and progression have been ascribed only to genetic alterations occurring in cells, with no influence of the surrounding microenvironment where cells were located. Because of this, most of the earlier studies were focused exclusively on cancer cells. This concept changed over the last years, as epigenetic contributions from stromal cells in close proximity to cancer cells were found to play a crucial role in the growth, angiogenesis, invasion, and metastases of most carcinomas, including those originated in breast [
2‐
4]. In that aspect, the establishment and characterization of cell lines derived not only from neoplastic but also from stromal components of breast carcinomas are extremely important to study more accurately the biological consequences of these heterotypic cell interactions. Myofibroblasts, defined as fibroblasts with α-smooth muscle actin (SMA) expression [
5], constitute the most predominant carcinoma-associated stromal cells, and have been found to be involved in the production of different proteases responsible for invasion [
6‐
8], as well as in the stimulation of the proliferation of cancer cells [
9,
10].
In the present study, we succeeded at isolating epithelial (LM-234ep) and myofibroblast (LM-234mf) cell lines from two murine mammary adenocarcinomas derived from a common ancestor spontaneously arisen in BALB/c mice [
11]. Despite its malignant mammary carcinoma origin, the immortalized epithelial LM-234ep cell line was not tumorigenic in either syngeneic or immunodepressed mice. Moreover, the myofibroblast LM-234mf cell line was more invasive
in vitro than LM-234ep cell line. Paradoxically, we found that LM-234ep-derived factors were able not only to inhibit the proliferation of different cancer cells
in vitro but also their growth
in vivo. To our knowledge, this is the first report showing antiproliferative and antineoplastic activities induced by tumor-associated epithelial cells.
Methods
Establishment of LM-234 cell lines
M-234 is a mammary tumor spontaneously arisen in a BALB/c female mouse, histologically defined as a semidifferentiated carcinoma with low mitotic rate, and absence of estrogen and progesterone receptors [
11]. The tumor, obtained from the Animal Care Facility of School of Medical Sciences, University of Rosario, is maintained by serial subcutaneous passages in syngeneic mice. Two tumor variants with similar characteristics obtained from M-234, namely M-234p and M-234m, were used here. Non-necrotic fragments from both tumors were used for
in vitro culture. Briefly, palpable tumors were aseptically excised, and minced to about 1 mm
3. The tumor pieces were disaggregated by incubation at 37°C in phosphate buffered saline (PBS) containing 0.25% bovine serum albumin (BSA), 0.25% trypsin, 0.25% collagenase type II, and 0.1% hyaluronidase (all reagents from Sigma Chemical Co., St. Louis, MO). The resulting cell suspensions were centrifuged for 10 min at 1,000 rpm, and the pellets obtained resuspended in RPMI-1640 or DMEM culture media (Sigma) with 10% fetal bovine serum (FBS, Natocor, Córdoba, Argentina). The cells derived from M-234p grew faster in RPMI-1640 than in DMEM, whereas those derived from M-234m grew only in DMEM. The cells have been maintained in culture for over 60 passages. Cell lines derived from M-234p and M-234m are referred to as LM-234ep and LM-234mf, respectively.
Other cell lines
B16-F10, a C57BL/6 mouse melanoma cell line [
12], and HeLa, a human cervix adenocarcinoma cell line [
13], were both obtained from American Type Culture Collection (ATCC, Manassas, VA). They were cultured in DMEM supplemented with 10% FBS. BMA3.1A7 [
14], a mouse macrophage cell line derived from bone marrow, kindly provided by Dr. Kenneth Rock (Dana-Farber Cancer Institute, Boston, MA), was cultured in RPMI 1640 with 10% FBS. The osteoblast-like cell line 7F2 [
15], isolated from mouse bone marrow, was obtained from ATCC, and cultured in Alpha minimum essential medium with 2 mM L-glutamine, 1 mM sodium pyruvate without ribonucleosides and deoxyribonucleosides, supplemented with 10% FBS.
Immunocytochemistry and immunofluorescence analyses
LM-234ep and mf cells were grown overnight on sterile 13-mm Thermanox™ coverslips or Lab Tek™. 4-chamber glass slides, both from Nunc (Naperville, IL). Cells were rinsed in PBS, and then fixed in ice-cold methanol (-20°C) for 30 min. After rinsing in PBS, immunoreactive proteins were detected by one of three methods: 1) for detection of SMA, slides were incubated with a fluorescein-conjugated monoclonal antibody that cross reacts with human and mouse αSMA (20 μg/ml; Biomeda, Foster City, CA) for 1 h at room temperature, and then counterstained with DAPI (Sigma, 5 μg/ml Tris buffer) for 20 sec, and mounted with Vectashield (Vector Laboratories, Burlingame, USA); 2) for protein detection, slides were incubated for 1 h at 37°C with a 1:200 antibody dilution of mouse anti-c-Jun (Transduction Laboratories, San Diego, CA), rabbit anti-cyclin A or rabbit anti-cyclin D (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and then incubated for another hour with an anti-mouse or rabbit FITC-antibody; 3) for detection of cytokeratin, endogenous peroxidase activity was quenched with 3% hydrogen peroxide in methanol, and non-specific immunobinding blocked with Superblock (Scytek Laboratories, Logan, UT) for 10 min at room temperature. Slides were then incubated with an anti-Pan cytokeratin antibody (1:50; Sigma) for 2 h at room temperature. After rinsing with PBS, Vectastain Elite ABC kit for mouse IgG (Vector Laboratories) was used. Immunoreactive cytokeratins were recognized as a brown color after incubation with diaminobenzidine (DAB; Sigma)/H2O2 solution. Slides were lightly counterstained with Mayer hematoxylin before mounting. For each epitope analyzed, species/isotype-matched pre-immune IgGs served as negative primary antibody controls.
Karyotyping
Exponentially growing LM-234ep and LM-234mf cell cultures were exposed to 0.2 μg/ml colcemid (Sigma) for one hour at 37°C. Metaphase spreads were prepared and stained by conventional methods as described previously [
16]. A minimum of fifty metaphases for each cell line was selected and photographed to determine chromosome frequency distribution and morphology.
Zymography
Confluent monolayers of LM-234mf and LM-234ep cells were incubated for 48 h in FBS-free culture media containing 0.1% BSA, to obtain conditioned media (CM). All CM were cell number-normalized. The gelatinase activity was detected by zymography in 10% sodium dodecyl sulfate-polyacrylamide gels containing 1 mg/ml gelatin (Sigma), as described previously [
17]. Gelatinolytic bands were visualized as transparent areas against the dark-blue background. As positive controls for murine MMP-2 and MMP-9, CM derived from 7F2 and BMA3.1A7 cell lines were used, respectively.
Chemoinvasion assay
Cell invasion assay was carried out in Transwell
® cell culture chambers (Corning Costar, Cambridge, Massachusetts) essentially as previously described [
18]. Briefly, the upper surface of the insert membrane (8-μm pore size) was coated with 37.5 μg/100 μl of cold reconstituted basement membrane Matrigel™, and dried overnight at room temperature in a tissue culture hood. The lower compartment of the Transwell
® contained 5% FBS, which was used as chemoattractant. LM-234ep or LM-234mf cells were seeded at a density of 1 × 10
5per Transwell
® culture insert in 100 μl of DMEM containing 0.1% BSA. The chambers were incubated at 37°C for 18 h. Non-migrating cells on the upper surface of the filters were removed with a cotton swab, whereas the cells that traversed the filters were stained with Diff-Quick (Dade-Behring, Newark, DE), and counted in 10 random 20× fields using a light microscope. All the assays were carried out in triplicate.
Tumorigenic potential
Subconfluent LM-234ep and LM-234mf cell cultures were trypsinized and resuspended in their respective FBS-free culture medium. The viability of both cell types was above 90%, as determined by trypan blue exclusion. To test the tumorigenicity of the cell lines, 1 × 10
6-1 × 10
7 cells were subcutaneously or intramammary fat pad inoculated into 10–12 week-old female BALB/c (from the Rodent Facility of the University of Rosario School of Medical Sciences) or 8–12 week-old congenitally athymic female Swiss
nude mice (from the Mice Facility of Centro de Biología Molecular Severo Ochoa, Universidad Autónoma de Madrid). Co-inoculation of cells and Matrigel™ (BD Biosciences, Rockville, MD), a basement membrane preparation extracted from EHS mouse sarcoma, was performed in some cases in an attempt to enhance tumorigenicity of the cells [
19]. To study the effect of heterotypic interaction on tumor growth, experiments involving co-inoculation of LM-234mf and LM-234ep cells (1.5 × 10
6 cells each) in
nude mice were performed. Tumors volumes were calculated as:
π/6 ×
Length ×
Width ×
Height, based on periodic measurement on calipers.
Nude mice were maintained under aseptic conditions.
Co-culture studies
LM-234ep and LM-234mf cells were seeded at 1:1 ratio (104 cells for each cell line) onto Thermanox™ coverslips that were placed inside 24-well tissue culture plates. Cells were co-cultured in DMEM supplemented with 10% FBS, and coverslips were harvested 4 and 8 days after cell seeding for cytokeratin immunocytochemistry, performed as explained above. Digital photomicrographs were captured using a Zeiss Axioplan 2 microscope (Zeiss, Göttingen, Germany) equipped with a software-controlled (Axiovision, Zeiss) digital camera.
LM-234mf, B16-F10, and HeLa cells were seeded into 6-well plates at a density of 1 × 104 cells/well on day 0, and allowed to attach overnight. Thereafter, cells were incubated in the presence of 1 ml of either LM-234ep or LM-234mf CM, obtained after culturing confluent monolayers for 48 h in their culture media containing 10% FBS, or their respective complete culture medium (control). To avoid non specific effect of the LM-234ep or LM-234mf CM, the pH was controlled in all assays, the glucose, lactate and glutamine were measured using a Biochemistry analyzer YSI 2700 Select (YSI Inc., Yellow Springs, OH) and we have used the same cell numbers of LM-234ep in all the experiments to obtain the CM. Cell numbers were counted every 2 days up to day 7 with a Neubauer chamber.
Fluorescence-activated cell sorting (FACS) analysis
After being exposed to LM-234ep CM or culture medium, both supplemented with 10% FBS, LM-234mf, B16-F10, and HeLa cells were harvested by trypsin digestion, resuspended in cold PBS, and fixed in 75% ethanol at 4°C for at least 8 h. After a brief centrifugation, fixed cells were resuspended in PBS containing propidium iodide (5 μg/ml; Sigma)/RNase A (30 μg/ml, Sigma). After incubation for 30 min at room temperature, DNA staining was evaluated by fluorescence 2 intensity in the linear scale using FACScan.
Immunoblotting
Cells were lysed on ice in lysis buffer (50 mM Tris-HCl (pH 7.5), 300 mM NaCl, 0.5% Triton ×-100, 1 mM EDTA and protease inhibitor). Cellular debris was removed by centrifugation of lysates for 10 min at 14,000 g. The supernatant was collected, and equal amounts of proteins, as determined by the BCA protein micro assay (Bio-Rad Laboratories, Hercules, CA), were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electroblotted onto nitrocellulose membranes. After the transfer, membranes were blocked with a 10% (wt/vol) solution of skim milk powder in PBS, and then probed with one of the following primary antibodies: rabbit anti-p-ERK (1:200; Santa Cruz Biotechnology), mouse anti-c-Jun (1:1,000; Transduction Laboratories), rabbit anti-JunB (1:500; Santa Cruz Biotechnology), rabbit anti-cyclin E (1:200; Santa Cruz Biotechnology), rabbit anti-cyclin A (1:200; Santa Cruz Biotechnology), rabbit anti-cyclin D (1:200; Santa Cruz Biotechnology), mouse anti-Cdk2 (1:1,000; BD Transduction Laboratories), or rabbit anti-β-actin (1:200, Sigma). The different antigens were detected by chemiluminescence, using horseradish peroxidase-conjugated antibody (1:500; Dako, Carpinteria, CA).
For immunodetection of cytokeratins, membranes were incubated with a 1:500 dilution of monoclonal anti-pan cytokeratin (Sigma), which recognizes cytokeratins 1, 4, 5, 6, 8, 10, 13, 18 and 19. The cytokeratins detected by these antibodies range in molecular weight between 40 kDa and 68 kDa. For immunodetection of αSMA a monoclonal antibody (clone 1A4, Sigma) was used at 1:200.
Measurement of AP-1 activity
Cells were transfected with the AP-1-luciferase reporter gene plasmid carrying the collagenase promoter (p-73col-luc) [
20] or with the control plasmid without the AP-1 site and the pRL-tk-luc plasmid (Promega, Madison, WI) to normalize for transfection efficiency. After 24 h, the activity was determined using a Promega luciferase assay kit following the manufacturer's protocol. Data were normalized to
Renilla luciferase activity and are presented in relative luciferase units. Transfections were always run in triplicate and data presented as mean ± SD.
Treatment of tumor-bearing mice with LM-234ep conditioned medium
Nude mice were subcutaneously injected with 5 × 105 B16-F10 or 1 × 106 HeLa cells in both flanks. Tumors were allowed to grow until they became palpable, and then locally injected with 200 μl of FBS-free LM-234ep CM in the left flank, and with an equal volume of FBS-free culture medium contralaterally. Treatments were performed once every 2 days in mice injected with B16-F10 cells, and once a week in mice injected with HeLa cells. Treatments were continued until tumors reached a median volume of 500 mm3. Five mice per group were used.
Histological analysis and immunohistochemistry
Tumors were fixed in 10% buffered formaldehyde, paraffin-embedded, and processed by routine methods to get 5-μm consecutive histological sections that were used for H&E staining and immunohistochemical studies. To determine the proliferative activity within the tumors, sections were pretreated in Ag Citrus Plus Retrieval Solution (BioGenex, San Ramon, CA) in a microwave, and then incubated with anti-Ki67 monoclonal antibody (1:40; BioGenex, San Ramon, CA). Intratumoral microvascularity was assessed using a rat anti-mouse CD34 antibody (1:25; Cell Sciences, Canton, MA). Mouse on Mouse (MOM™) immunodetection peroxidase kit (Vector, Burlingame, CA) and Vectastain Elite ABC kit for rat IgG (Vector Laboratories) were used to visualize Ki-67 and CD34 immunoreactive sites, respectively. Slides were lightly counterstained with Mayer hematoxylin before mounting. In each case, species/isotype-matched pre-immune IgGs served as negative primary antibody controls.
Statistical analysis
Data were statistically analyzed using Student's t test using GraphPad InStat® version 3.0 (GraphPad Software, San Diego, CA). Differences were considered to be statistically significant at P < 0.05.
Discussion
For many decades, the neoplastic phenotype has been considered to be driven only by genetic alterations occurring in cells. However, many studies performed in the last years have demonstrated that genetic changes in individual cells may not be sufficient, and that the cellular microenvironment has a powerful influence in determining the final fate of cells in the process of transformation [
24‐
26]. The oncogenic potential of cells can be modulated by adjacent cells in opposite directions, sometimes contributing to tumor formation [
4,
27] and others impeding it [
28,
29]. Moreover, stromal cells present in the tumor microenvironment have a profound effect on tumor growth, invasion and metastatic progression [
30]. This is also true for breast cancer, where most of the studies have focused on the luminal epithelial cells found in the ducts and alveoli, despite the presence of stromal cells including fibroblasts, myofibroblasts, leukocytes, and myoepithelial cells [
4,
27]. Myofibroblasts, for instance, which are usually absent in normal tissue derived from breast, have been associated with tumor progression [
26,
31].
To attain a better understanding of the complex heterotypic interactions that occur in breast cancer, suitable model systems should involve cells representing both epithelial and stromal cells ideally derived from mammary tumors. Here, we isolated and established two cell lines from mammary carcinomas derived from a common tumor spontaneously arisen in a BALB/c female mouse [
11]. On the basis of their immunophenotypical features, the cytokeratin positive and SMA negative LM-234ep cell line was considered of epithelial origin, whereas LM-234mf, which is positive for SMA but lacks cytokeratin expression, was categorized as myofibroblast. Both cell lines have been growing in culture for over 60 passages, with abnormalities in the number and structure of their chromosomes. In the near-triploid LM-234ep cells, double-minutes are found, suggesting genomic amplification. As for the near-hexaploid LM-234mf cells, mostly chromosomes breaks are observed. Despite these features characteristic of cancerous cells, both cell lines did not generate progressing tumors mice when different numbers of cells were injected subcutaneously (with or without Matrigel) or orthotopically in the mammary fat pad of either syngeneic mice or
nude mice. Due to their epithelial phenotype, their immortalized nature, and the fact that they were derived from a mammary tumor, LM-234ep cells appear to derive from carcinoma cells, though they only revealed an incipient growth
in vivo that did not persist. Although some ER positive breast cancer cells do not form tumors unless the hosts are treated with β-estradiol, that would not be the case for LM-234ep cells that are ER negative (data not shown). As for LM-234mf myofibroblast cells, it is not clear why they escaped from senescence and revealed anomalous chromosome patterns usually present in cancerous cells, but did not induce tumor formation. These results agree with those obtained in an immortalized myofibroblast cell line obtained from a human liver angiosarcoma, which also shows severe chromosomal alterations and inability to induce tumors [
32]. Particularly in breast cancer, where epithelial-mesenchymal transition occurs in approximately 18% of tumors [
33], a transdifferentiation of epithelial tumor cells into myofibroblasts that lose their malignant phenotype has been reported [
34]. We hypothesize that this could have been the case for LM-234mf myofibroblasts, which are not tumorigenic but show aneuploidy, characteristic of cancer cells. In fact, myofibroblasts are considered a desmoplastic reaction that may facilitate mammary tumor growth and invasion [
34]. Myofibroblasts, for example, can invade certain tumors preceding endothelial invasion necessary for angiogenesis [
35,
36]. Here, we show that LM-234mf cells are highly invasive
in vitro and secrete MMP-2 in an active form, whereas epithelial LM-234ep cells are almost non invasive and only secrete MMP-2 as a zymogen.
It is now widely accepted that a concerted action of epithelial and stromal cells is necessary to support tumor growth and progression. Having two different cell types derived from murine mammary tumors of common origin, we aimed at elucidating the impact of their heterotypic interaction on tumor growth. As we mentioned above, only LM-234ep cells induced a small and brief tumescence in vivo that then regressed. Despite an enhanced growth obtained when LM-234mf and LM-234ep cells were co-inoculated in nude mice, this ended up in a complete regression. When both cell types were co-cultured, a predominance of LM-234ep over LM-234mf was evident. This cannot be explained in terms of different growth rates, as LM-234ep cells have a longer doubling time than LM-234mf cells. Taken together, our results suggests an inhibitory effect of LM-234ep on LM-234mf cells, and could explain why the initial enhancing effect revealed by LM-234mf on LM-234ep cell growth in vivo was finally hampered by the epithelial cell type. This inhibitory effect displayed by LM-234ep cells would be accomplished by diffusible factors, since LM-234ep CM revealed strong inhibitory activity in vitro on LM-234mf cell proliferation, which would be mediated by down-regulation of the ERK/AP-1 pathway. The antiproliferative effect was also observed on a mouse melanoma cell line (B16-F10) and on a human cervix adenocarcinoma (HeLa) and, in both cases, is the result of an arrest of the cells at the G0/G1 phase of the cell cycle. The reduced levels of cyclins E, A, and D, and of Cdk2 would account for the observed decrease in the percentage of cells in the S-G2/M phase. These results suggest that the inhibitory factor(s) derived from LM-234ep cells can directly or indirectly down-regulate the expression of those cell cycle genes. Moreover, the growth of tumors formed in nude mice by injection of HeLa and B16-F10 cells was inhibited by local treatment with LM-234ep CM. The immunohistochemical analysis of the nuclear Ki67 proliferation marker suggests an antiproliferative direct effect on the tumor cells.
Different studies have demonstrated that the normal myoepithelium can function as an inhibitor of growth, angiogenesis, invasion, and metastasis of breast cancer cells [
37‐
40]. However, to the best of our knowledge, there are no reports on growth inhibitory activity of soluble factors derived from tumor-associated epithelial cells. Our results suggest that within the highly heterogeneous mammary tumor cell populations there are epithelial cells acting as guardians that may counterbalance malignant progression through antiproliferative effects. Further studies are currently in progress to enlighten the nature of the factor(s) produced by tumor-derived LM-234ep that are responsible for the inhibitory effect demonstrated on tumors.
Conclusion
The results herein obtained indicate that LM-234ep and LM-234mf cell lines are suitable models to examine several aspects of tumor biology, in particular those related to the different pathways involved in the genetic events leading to tumor generation. We consider both cell lines as transformed but non tumorigenic. Also, as these cell lines are neither normal nor malignant, and can be situated somewhere in between those two stages, they could be used as a model to study different therapeutic strategies aiming to revert the transformed immortal cell type to a normal senescent one.
More interestingly, our data suggest the existence of epithelial cells with tumor-inhibitory activity within mammary tumors. The LM-234ep cell line characterized here would exemplify such cells that are transformed, as they displayed several chromosomal anomalies and did not senesce when cultured in vitro, and could represent a protective mechanism against tumor progression. The broad inhibitory spectrum activity demonstrated by the factor(s) secreted by LM-234ep cells on the in vitro proliferation and the in vivo growth of two non-related tumors justifies further studies that could derive in a novel therapeutic option for cancer.
Acknowledgements
This work was supported by Cancer Research Foundation (Fundación de Investigación del Cáncer, FUNDIC), Buenos Aires, Argentina. G.S. is an investigator of Consejo de Investigaciones, Universidad Nacional de Rosario. The authors thank Dr. Marta Izquierdo, for making possible many of the experiments developed by E.N.G. in her laboratory. E.N.G. currently holds a grant of the Comunidad de Madrid at the Department of Molecular Biology, Centro de Biología Molecular Severo Ochoa, Universidad Autónoma de Madrid, Spain. For animal care and handling we have strictly followed the legislation and guidelines in our country (Spanish Royal Decree 1201/2005), the ones from the British UKCCCR committee for the welfare of animals in experimental neoplasia (revised version of July 1997) and the European Union (2003/65/CE from the European Parliament and Council July 2003).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ENG carried out most of the studies, drafted the manuscript, and participated in the analysis and interpretation of data. SN optimized the gelatin zymographic studies and proliferation assays, and participated in the analysis and interpretation of data. HY and HM performed the immunohistochemical and proliferation assays. GS and DB conceived and coordinated the study, participated in the analysis and interpretation of data, and helped to draft the manuscript. All authors read and approved the final manuscript.