Neoplastic transformation of breast epithelial cells by genotoxic stress

MCF 10A cells (ATCC, Rockville, MD) were irradiated at 0.67 cGy/min to the desired dose (0.1 Gy) using a γ-cell 40 ¹³⁷Cesium irradiator (Atomic Energy of Canada). A 40 mg/ml stock solution of cigarette smoke condensate (Murthy Pharmaceuticals, USA) was prepared in DMSO and diluted with culture medium to a final concentration of 10 mg/ml. Exponentially growing cells were first irradiated and then exposed to Csc (10 μg/ml) for 72 hr [29].

Determination of DNA content by flow cytometry was performed as previously described [30]. Briefly, 2 × 10⁶ MCF 10A cells were incubated for 72 hr following individual and combined treatments. Fixed cells were incubated with a staining solution containing 0.56% NP-40, 3.7% formaldehyde, and 0.01 mg/ml Hoechst 33258 in phosphate-citrate buffered (pH 7.2) solution. DNA content was analyzed by flow cytometry instrumentation (BD Biosciences, San Jose, CA).

For immunoblot experiments, 20 μg of total cellular protein was subjected to SDS-PAGE. Membranes were probed with the primary antibodies indicated in the figure legend.

Matrigel (100 μl; 7-8 mg/ml) in serum-free medium was added to each well of a Transwell Corning Costar plate (Costar, Acton, MA, USA) and dried overnight in a tissue culture hood. The following day, 2.5 × 10⁴ cells in serum-free medium were pipetted onto the Matrigel and complete medium was added to the bottom trough. Following incubation, the transmembrane filter was stained with crystal violet and the number of cells counted.

For wound healing, a small area was cleared along a diameter of the 10 cm dishes of confluent monolayers of MCF 10A and MCF 10A treated cells with a sterile pipette tip. Cell migration was measured and photographed from the wound/scratch edge every 8 hr.

Following treatments, cells were incubated in 0.01 mg/ml of Hoechst 33242 dye for 45 minutes and then washed and incubated for a further 45 minutes and photographed using the Nikon 80i fluorescent microscope. Fluorescence intensity per cell (blue channel) was analyzed using image J software (n = 4 in each case).

RNA was extracted using the Qiagen mRNA extraction kit. The RNA samples were analyzed with Affymetrix GeneChip Human 133 2.0 Arrays. The quality of the microarray experiments was assessed with affyPLM and Affy, found at Bioconductor http://www.bioconductor.org. All computation was performed under R environment http://www.r-project.org. The Affy software was used to estimate the gene expression signals and evaluated using the Robust Multi-array Average [31]. Data normalization was performed using a Bayes method at Bioconductor, which includes log normal modeling. EBarrays was used to estimate the posterior probabilities of the differential expression of genes between the control and treated sample [32, 33]. This data has been deposited with Gene Expression Omnibus (GEO-accession number GSE21066).

Single cell suspensions (total 5000 cells in 2 ml) in Dulbecco's modified Eagle's medium/F-12 containing 5 mg/mL insulin, 0.5 mg/mL hydrocortisone, 2% B27 (Invitrogen Ltd., Paisley, Scotland), and 20 ng/mL epidermal growth factor were seeded into ultra-low attachment plates (Corning, Lowell, MA) and incubated for 7-10 days in presence of 20% O₂ and 5% CO₂ at 37°C. Subsequently, the number of mammospheres formed were counted using a microscope.

The effect of low dose ionizing radiation and cigarette smoke condensate was tested using MCF 10A normal immortalized breast cells. Following treatments, we observed differences between the morphological appearances of the differently treated cultures. Single treatment regimes generated consistent changes in cellular morphologies relative to the untreated controls (Figure 1A). Thus, irradiation or Csc treatment resulted in populations of cells with a rounded or slightly elongated morphology (Figure 1A top right and lower left panels). In contrast, combined treatment (Rad + Csc) of MCF 10A cells generated a phenotype that was somewhat fibroblast-like (Figure 1A lower right panel), which grew with fewer cell to cell contacts and was less cuboidal than the parental cells. Flow cytometry analyses were performed to evaluate whether the latter phenotypic changes could be associated with a change in cell growth rate or cell cycle frequency and hence a change in proliferation rate. As shown in Figure 1B, irradiation (0.1 Gray) increased the percentage of cells found in G2/M from the 3.7% found in the untreated cells to 39.2% while Csc treatment increased the percentage of cells in G0/G1 from 49% found in untreated cells to 75.4%. Thus, irradiation did not cause G1 arrest for MCF 10A cells while Csc treatment caused G1 arrest. However, treatment of MCF 10A cells with a combination of radiation plus Csc resulted in 42.9% of the cells being in S phase, which is roughly double that of the untreated cells and provides an indication that the combined treatment increased the percentage of proliferating cells (Figure 1B). Subsequently, immunoblot analyses were performed to investigate whether the cell cycle changes mediated by genotoxic stress could be correlated to the expression of cell cycle regulatory proteins. As indicated in Figure 1C, p53 protein levels were slightly increased following low dose irradiation, which was matched to an increase in the levels of its downstream target p21. In contrast to Rad, Csc or a combination of radiation plus Csc treatment resulted in a down regulation of p53 and p21 levels (Figure 1C). This indicates that Rad + Csc can alter the levels of regulatory molecules to dysregulate phases of the cell cycle.

Bild vergrößern

Increased cell motility and invasion are characteristics of a neoplastic transformation. To test whether the genotoxic stresses used in our experimental setting could increase the migration of MCF 10A cells, we used a scratch/wound assay. As shown in Figure 2A, MCF 10A cells treated with Rad as well as Rad + Csc were able to migrate into the cleared area within 24 hr of incubation. On the other hand, untreated and Csc treated MCF 10A cells showed little migration into the cleared region during this time period. We next investigated the effect of radiation and Csc on cell invasion in a modified Boyden chamber assay. As depicted in Figure 2B, MCF 10 A cells treated with Csc lack the ability to invade whereas Rad or the combination Rad + Csc cells were extremely invasive (four fold increase) as compared to the untreated and Csc treated MCF 10A cells. Taken together these data indicate that the Rad as well as Rad + Csc treated cells acquired genetic/biochemical alterations that increased the cells ability to migrate and invade, i.e., a possible increase in dissemination or metastatic potential.

Bild vergrößern

A distinct characteristic of a metastatic phenotype is generally increased drug resistant properties. Some of the pathways that promote drug resistance are elevated expression of ATP-binding cassette (ABC) glycoprotein transporters at the cell surface [34]. The elevated expression of these transporters can exclude vital dyes such as Hoechst 33342 analogues to chemotherapeutic drugs, thus promoting multi-drug resistance. As MCF 10A cells treated with Rad + Csc induced a transformed phenotype, we evaluated the ability of these cells to extrude the Hoechst 33342 dye, a surrogate marker for drug uptake. As seen in Figure 2C &2D, cells treated with both Rad + Csc retained less dye as compared to the untreated MCF 10A cells. This indicates that the combination treatment in MCF 10A cells can induce a drug resistant phenotype.

In an effort to determine which genes might be contributing to the increased cell invasion and migration capabilities of Rad or Rad + Csc treated MCF 10A cells, we performed expression analyses using the Affymetrix platform. To identify differentially expressed gene sets following treatment, we compared the individual (Rad or Csc) treatment data sets with the combined treatment (Rad + Csc) data set using a Venn diagram (Figure 3A). The Venn diagram indicates that 53 genes differentially expressed in the combination treated cells compared to untreated cells. Among these, 18 genes (12 up regulated and 6 down regulated) have been reported as participating in cellular pathways of cancer (Figure 3B). Included in this group were genes involved in tissue remodeling, metabolism and cell adhesion molecules. One cell adhesion molecule identified was CD44, which has been directly correlated to human breast cancer grade [35]. Moreover, a high CD44 low CD24 expression profile in breast cancer cell lines has been associated with a putative breast cancer initiating cell phenotype [36]. In order to validate the microarray data, we estimated the CD44 levels in treated versus untreated cell lines. The expression level of CD44 was up regulated 2 fold in the combined treatment as compared to the untreated and Rad treated (Figure 3C). Recently, cells exhibiting a high CD44 low CD24 expression profile have been shown to form mammospheres in culture conditions, which were resistant to chemotherapeutic regimes [37]. Therefore, we investigated the effect of genotoxic stress on MCF 10A cells to form mammospheres in vitro. Figure 4A shows representative mammosphere images under different genotoxic stress. The combined treated cells formed the most number of mammospheres as compared to individual treatments (Figure 4B). Thus, the Affymetrix data indicates that the combined treatment can initiate gene expression patterns that are associated with known cancers. That along with enhanced mammosphere formation strengthens our suggestion that further experiments will indicate that the altered gene profiles are consistent representations of known cancer cell phenotypes.

Bild vergrößern

Bild vergrößern