Role of Cancer Stem-like Cells in the Process of Invasion and Mesenchymal Transformation by a Reconstituted Triple-negative Breast Cancer Cell Population Resistant to p53-induced Apoptosis

Abstract

We investigated the role of cancer stem cells (CSCs) in a population of triple-negative breast cancer (TNBC) cells that are resistant to apoptosis. A human breast cancer cell population capable of inducing p53 expression with doxycycline (Dox) was created and used as an untreated control (UT). After the addition of Dox to UT for 5 days, the cell population reconstituted with cells showing resistance to apoptosis was named RE. Fluorescence-activated cell sorting (FACS) and immunostaining revealed that after the addition of Dox, the ratio of cells in the S and G2/M phases decreased in UT as apoptosis proceeded, but did not markedly change in apoptosis-resistant RE. CSC-like cells in RE exhibited a cell morphology with a larger ratio of the major/minor axis than UT. FACS showed that RE had a higher proportion of CSC-like cells and contained more CD44⁺CD24^- mesenchymal CSCs than ALDH1A3⁺ epithelial-like CSCs. In a Matrigel invasion assay, UT was more likely to form a three-dimensional cell population, whereas RE exhibited a planar population, higher migration ability, and the up-regulated expression of epithelial-mesenchymal transition-related genes. These results provide insights into the mechanisms by which TNBC cells acquire treatment resistance at the time of recurrence.

Keywords: apoptosis resistance; breast cancer; cancer stem-like cell; invasion.

Fig. 1.

Dox-induced p53 expression by a FACS analysis. UT Dox0d (A) and RE Dox0d (B) are indicated by a red line, and UT Dox2d (A) and RE Dox2d (B) by a green line.

Fig. 2.

Time-course changes in the number of cells and effects of the addition of Dox on growth curves. A: Changes in the number of cells without Dox. B: Examples of cell growth patterns with Dox analyzed by an inverted phase-contrast microscope. Bar = 200 μm. C and D: Changes in the number of viable cells with Dox in UT (C) and RE (D). The number of UT is expressed on a 1/10 scale of RE.

Fig. 3.

Cell cycle analysis by FACS. A, B, C, and D: Each cell was pulse labeled with BrdU and nuclear DNA was stained with 7-AAD. The red open square indicates the G0/G1 phase, blue the S phase, and orange the G2/M phase. A: UT Dox0d, B: UT Dox3d, C: RE Dox0d, and D: RE Dox3d. E and F: Each phase of the cell cycle, the G0/G1, S, and G2/M phases, was expressed as a ratio to the total number of cells.

Fig. 4.

Detection of CSC-like cells in UT and the reconstituted RE cell population by a FACS analysis. A and C: Positive cells are defined as a fluorescent signal intensity of 10³ or more, shown in purple areas in the histogram. A: UT Dox0d, C: RE Dox0d. B and D: CD24⁻ is defined as 10³ or less and CD44⁺ as 10³ or more, which are indicated by the open red square. E and F: Time-course analysis of the expression patterns of ALDH1A3 (E) and CD44⁺CD24⁻ (F) in UT and RE. The bar indicates means ± SE.

Fig. 5.

Analysis of the cell population constitution by immunocytochemical staining and morphological measurements. A: Double immunocytochemical staining. ALDH1A3 stains blue in the cytoplasm and Ki-67 stains brown in the nucleus. a and b: Examples of the expression patterns of ALDH1A3 and Ki-67 in Dox0d. a: UT Dox0d, b: RE Dox0d. solid arrow: ALDH1A3⁺Ki-67⁻, solid arrowhead: ALDH1A3⁺Ki-67⁺, open arrowhead: ALDH1A3⁻Ki-67⁺, and open arrow: ALDH1A3⁻Ki67⁻. c and d: Examples of L (≥ 50 μm) and S cells (< 50 μm), which are distinguished by the length of their major axis. Large arrows: L cells, small arrows: S cells. c: UT Dox1d, d: RE Dox1d. Bar = 50 μm. B and C: Time-course analysis of expression patterns of the CSC marker, ALDH1A3⁺Ki-67⁻ (B) and the cancer precursor cell marker, ALDH1A3⁺Ki-67⁺ (C) indicated by the expression rate with the addition of Dox. The bar indicates means ± SE. *p < 0.01. D: The rate of ALDH1A3⁺Ki-67⁻ cells without Dox sorted by cell size. The ratios of S cells and L cells divided by the total number of cells are shown. The bar indicates means ± SE. *p < 0.01 UT versus RE. E: The ratio of the major/minor axis of mesenchymal-like elongated cells in ALDH1A3⁺Ki-67⁻ cells. **p < 0.05 (the Student’s t-test).

Fig. 6.

Epithelial-like and mesenchymal-like CSCs by a FACS analysis. A: The representation of cell subpopulations in the diagram indicates CSCs with the epithelial-like ALDH1A3⁺ (E), mesenchymal-like CD44⁺CD24⁻ (M), and hybrid epithelial/mesenchymal ALDH1A3⁺/CD44⁺CD24⁻ (E/M) phenotypes. B: Ratio of cells from ALDH1A3⁺ cells (E) minus hybrid ALDH1A3⁺/CD44⁺CD24⁻ cells (E/M). C: Ratio of hybrid ALDH1A3⁺/CD44⁺CD24⁻ cells (E/M). The bar indicates means ± SE. *p < 0.01 UT versus RE. D: Ratio of cells from CD44⁺CD24⁻ cells (M) minus hybrid ALDH1A3⁺/CD44⁺CD24⁻ cells (E/M). The bar indicates means ± SE. *p < 0.01 UT versus RE.

Fig. 7.

Quantification of cell invasion and migration. A: Cells reaching the filter pores after a 6-hr culture in Matrigel were stained by Giemsa and photographed. Bar = 50 μm. B: Time-course analysis of the total number of invading cells per chamber. Each point indicates the average value and standard deviation of 3 chambers.

Fig. 8.

Depth and morphological characteristics of invading cells in Matrigel. A: Fluorescence images of cells infiltrating Matrigel by a confocal laser scanning microscope. Cells cultured in the Matrigel chamber were stained with DAPI and color-coded as follows. Regarding the three-dimensional construction of cells, Z-stack images are represented on a color code scale using Fiji/ImageJ software (version 1.52 g, Java 1.80_172, NIH). Bar = 100 μm. B: Depth of invasion in Matrigel. The distance from the surface of Matrigel to the invasion front of cell populations in Matrigel cultures for 1 hr and 6 hr was measured by a confocal laser scanning microscope. The values at each point are indicated by the mean and standard deviation (n = 5). *p < 0.01 (the Student’s t-test). C: Percentages of cell populations that form polygon and complex meshes during infiltration in Matrigel. The definition of the polygon and complex mesh was previously described (Aranda and Owen 2009 [2]). D and E: Expression levels of VM marker mRNA in infiltrating cells in Matrigel. D: MMP-9, E: VE-cadherin). p < 0.01 (the Student’s t-test).

Fig. 9.

Snail expression patterns in the process of Matrigel infiltration. Cells in the Matrigel culture for 6 hr were immunofluorescently stained with Snail. A: Fluorescence images of Snail expression patterns in Matrigel using a confocal laser scanning microscope. The nucleus is shown by DAPI (blue) and Snail expression is observed in the nucleus and cytoplasm (green). Bar = 100 μm. B: A quantitative analysis was performed on the Snail fluorescence channel using Fiji/ImageJ software. The average value of fluorescence intensity in 9 randomly selected fields of view where cells exist was measured, and the values of three wells were combined. The bar indicates means ± SE. *p < 0.01 UT versus RE. C: Semiquantitative analysis of Snail mRNA expression. RT-PCR was performed using RNA extracted from cells in the conventional monolayer culture and in Matrigel.