High mitochondrial mass identifies a sub-population of stem-like cancer cells that are chemo-resistant

Abstract

Chemo-resistance is a clinical barrier to more effective anti-cancer therapy. In this context, cancer stem-like cells (CSCs) are thought to be chemo-resistant, resulting in tumor recurrence and distant metastasis. Our hypothesis is that chemo-resistance in CSCs is driven, in part, by enhanced mitochondrial function. Here, we used breast cell lines and metastatic breast cancer patient samples to begin to dissect the role of mitochondrial metabolism in conferring the CSC phenotype. More specifically, we employed fluorescent staining with MitoTracker (MT) to metabolically fractionate these cell lines into mito-high and mito-low sub-populations, by flow-cytometry. Interestingly, cells with high mitochondrial mass (mito-high) were specifically enriched in a number of known CSC markers, such as aldehyde dehydrogenase (ALDH) activity, and they were ESA+/CD24-/low and formed mammospheres with higher efficiency. Large cell size is another independent characteristic of the stem cell phenotype; here, we observed a >2-fold increase in mitochondrial mass in large cells (>12-μm), relative to the smaller cell population (4-8-μm). Moreover, the mito-high cell population showed a 2.4-fold enrichment in tumor-initiating cell activity, based on limiting dilution assays in murine xenografts. Importantly, primary human breast CSCs isolated from patients with metastatic breast cancer or a patient derived xenograft (PDX) also showed the co-enrichment of ALDH activity and mitochondrial mass. Most significantly, our investigations demonstrated that mito-high cells were resistant to paclitaxel, resulting in little or no DNA damage, as measured using the comet assay. In summary, increased mitochondrial mass in a sub-population of breast cancer cells confers a stem-like phenotype and chemo-resistance. As such, our current findings have important clinical implications for over-coming drug resistance, by therapeutically targeting the mito-high CSC population.

Keywords: MitoTracker; cancer stem cells; mitochondria; tumor metabolism.

Figure 1. Mitochondrial mass directly correlates with ALDH activity and the ESA+CD24-/low CSC population

Representative dot plots of ALDH activity in MDA MB 231 A. and MCF7 B. cells, showing ALDH+ and ALDH− cells in the absence of DEAB. Histograms represent typical staining intensity of MitoTracker in ALDH+ and ALDH− populations in both cell lines. C. Graph showing fold change in mean fluorescence intensity (MFI) of MitoTracker (Deep Red; 640 nM), within ALDH+ and ALDH− populations of MCF7 and MDA MB 231 cell lines (n = 4 independent experiments). D. Graph showing fold change in mean fluorescence intensity of MitoTracker (Deep Red; 640 nM) within the ESA+CD24− (cancer stem-like cell, CSC) population and ESA+/CD24− depleted (non-CSC) populations of MDA MB 231 cells (n = 4 independent experiments). Bar graphs are shown as the mean ± SEM, t-test, two-tailed test, *P ≤ 0.05, **P ≤ 0.01.

Figure 2. Mitochondrial mass directly correlates with the enriched breast CSC population, identified using large cell size

A. Typical dot plot showing side scatter (SSC) and forward scatter (FSC) of live breast cancer cells, gates represent cell size (RED- 4–8 μm; BLUE- 9–12 μm; BLACK > 12 μm; [29]). Histograms show MitoTracker mean fluorescence intensity within the 3 cell size groups of MDA MB 231 and MCF7 cells. Graphs showing fold change in the mean fluorescence intensity (MFI) of MDA MB 231 B. and MCF7s C. within the 9–12 μm and > 12 μm cell size compared to the smallest cells (4–8 μm), n = 3 independent experiments, 2 technical replicates. Bar graphs are shown as the mean ± SEM, t-test, two-tailed test, ***P < 0.001.

Figure 3. Mitochondrial mass correlates with ALDH activity in primary breast cancer cells isolated from metastatic breast cancer samples and a patient derived xenograft

Representative dot plots of ALDH activity in primary metastatic breast cancer samples BB3RC88 A. BB3RC84 B. and patient derived xenograft sample (BB3RC50) C. cells showing ALDH+ and ALDH− cells in the absence of DEAB. Histograms represent typical staining intensity of MitoTracker (MT) in ALDH+ and ALDH− populations. D. Graph showing fold change in mean fluorescence intensity (MFI) of MitoTracker (Deep Red; 640 nM) within ALDH+ and ALDH− populations of primary metastatic breast cancer samples (n = 4), BB3RC77, BB3RC79, BB3RC88, BB3RC84. E. Graph showing fold change in mean fluorescence intensity (MFI) of MitoTracker (Deep Red; 640 nM) within ALDH+ and ALDH− populations of a patient derived xenograft sample (BB3RC50). F. Table showing the original (invasive) breast cancer characteristics of the metastatic (MET) breast cancer samples. Estrogen receptor (ER) and progesterone receptor (PR) status, + = > 10% positive; HER2 status, + = 3+ or 2+ & amplified; BB3RC50* denotes the PDX sample. Bar graphs are shown as the mean ± SEM, t-test, two-tailed test, *P ≤ 0.05.

Figure 4. High mitochondrial mass is specifically associated with mammosphere formation and tumor-initiating activity

MDA MB 231 and MCF7 cells were subjected to flow-cytometry to isolate different populations of MitoTracker (MT) stained cells: mito-low = lowest 5% of MitoTracker stained cells; mito-high = highest 5% of MitoTracker stained cells. Mito-high and low populations were then seeded into mammosphere cultures A. or injected by limiting dilution into NOD scid gamma (NSG) mice B. (A) Graph showing fold change in mammosphere formation in mito-low and mito-high populations of MDA MB 231 and MCF7 cells, n = 4 independent experiments, ≥3 technical replicates. (B) Graph showing tumor size (mm³), 9 weeks after cell injections (sub-cutaneous), with 1, 5, 10 and 50 cells per injection of mito-low and mito-high MDA MB 231 cell populations. Each dot represents a tumor, hollow dots <100 μm = no tumor initiation. C. Extreme limiting dilution analysis (ELDA) was used to calculate the tumor initiation cell frequency within the mito-low and mito-high MDA MB 231 cell populations. Graphs are shown as the mean ± SEM, t-test, two-tailed test, **P ≤ 0.01, **** ≤ 0.0001.

Figure 5. CSCs with high mitochondrial mass preferentially survive paclitaxel treatment and show reduced DNA strand breaks

MDA MB 231 and MCF7 cells were subjected to flow-cytometry to isolate different populations of MitoTracker stained cells: mito-low = lowest 5% of MitoTracker stained cells; mito-high = highest 5% of MitoTracker stained cells. Mito-high and low populations were then seeded into mammosphere culture in the presence or absence of paclitaxel (A–B) or treated with paclitaxel or control (DMSO 1:10,000) for 24 h before DNA damage analysis via a comet assay (C–D). A–B. Graphs show fold change in mammosphere formation in mito-low and mito-high cells (MDA MB 231 and MCF7), after treatment with 0.1 μM and 0.5 μM paclitaxel compared to vehicle alone controls, n = 3 independent experiments, ≥3 technical replicates. C–D. Graph shows fold change in comet tail length in mito-low and mito-high cell populations (MDA MB 231 and MCF7), after 24 h of 0.5 μM paclitaxel treatment, the images show representative comet tails under all conditions. Bar graphs are shown as the mean ± SEM, (ANOVA) test with post-hoc dunetts multiple comparisons (A–B), t-test, two-tailed test (C–D), **P ≤ 0.01, **** ≤ 0.0001.

Figure 6. Therapeutic targeting of chemo-resistant CSCs: A new systematic approach

Here, we propose a new clinical strategy for over-coming drug resistance. We suggest that primary or metastatic clinical samples could be used to purify chemo-resistant “mito-high” CSCs by flow-cytometry, after live-staining with MitoTracker, to estimate mitochondrial mass. Then, these chemo-resistant CSCs would be subjected to phenotypic drug screening, with well-defined panels of i) conventional chemotherapies and/or ii) mitochondrially-targeted FDA-approved drugs (e.g., metformin and antibiotics). This would allow us to achieve the goals of personalized cancer treatment, by establishing the chemo-sensitivity profiles of mito-high CSCs in individual patients, leading to the prevention of tumor recurrence and metastasis, in multiple cancer types. Chemo-resistant “mito-high” CSCs could also be expanded by using 3D-spheroid cultures, possibly for biobanking and drug screening.

Figure 7. Understanding the relationship between mitochondrial mass, “stemness” and chemo-resistance in cancer cells

Our report suggests that high mitochondrial mass, quantified using live MitoTracker staining, is associated with a number of stem cell characteristics and biological functions including CSC markers ALDH+, ESA/CD24, large cell size, mammosphere formation in vitro, as well as tumor-initiating activity in vivo. Moreover, cells with high mitochondrial mass were preferentially resistant to paclitaxel during mammosphere formation and the comet assay (a marker DNA damage), when compared to low mitochondrial mass cells. Thus, we conclude that high mitochondrial mass can be used as a biomarker to isolate a sub-population of stem-like cancer cells that are chemo-resistant. This has important implications for better appreciating the role of tumor metabolic heterogeneity in driving chemo-resistance and treatment failure, via recurrence and metastasis.