Accumulating evidence supports cancer to initiate and develop from a small population of stem-like cells termed as cancer stem cells (CSC). The exact phenotype of CSC and their counterparts in normal mammary gland is not well characterized. In this study our aim was to evaluate the phenotype and function of stem/progenitor cells in normal mammary epithelial cell populations and their malignant counterparts.
Freshly isolated cells from both normal and malignant human breasts were sorted using 13 widely used stem/progenitor cell markers individually or in combination by multi-parametric (up to 9 colors) cell sorting. The sorted populations were functionally evaluated by their ability to form colonies and mammospheres, in vitro.
We have compared, for the first time, the stem/progenitor markers of normal and malignant breasts side-by-side. Amongst all markers tested, we found CD44high/CD24low cell surface marker combination to be the most efficient at selecting normal epithelial progenitors. Further fractionation of CD44high/CD24low positive cells showed that this phenotype selects for luminal progenitors within Ep-CAMhigh/CD49f + cells, and enriches for basal progenitors within Ep-CAM-/low/CD49f + cells. On the other hand, primary breast cancer samples, which were mainly luminal Ep-CAMhigh, had CD44high/CD24low cells among both CD49fneg and CD49f + cancer cell fractions. However, functionally, CSC were predominantly CD49f + proposing the use of CD44high/CD24low in combination with Ep-CAM/CD49f cell surface markers to further enrich for CSC.
Our study clearly demonstrates that both normal and malignant breast cells with the CD44high/CD24low phenotype have the highest stem/progenitor cell ability when used in combination with Ep-CAM/CD49f reference markers. We believe that this extensive characterization study will help in understanding breast cancer carcinogenesis, heterogeneity and drug resistance.
Normal breastStem cellsBreast cancerFlow cytometryCD10ALDHCD44high/CD24lowMUC-1Mammary gland
Breast cancer is the most common cancer in women and, despite various treatment regimens, many patients die from the disease. A subpopulation of tumor cells, called cancer stem cells (CSC), is believed to contribute to the failure of breast cancer therapy due to their reported resistance to chemotherapy [1, 2] and radiotherapy . Due to their self-renewal abilities, even a minute population of CSC can form tumor when isolated and injected into an appropriate mouse model, while the remaining cells fail to do so . Breast CSC have been characterized as CD44high/CD24low , or aldehyde dehydrogenase enzyme (ALDH)high. However, how CSC relate to the different stem/progenitor cell populations of normal human mammary gland and whether CSC arise from normal mammary stem/progenitor or even differentiated cells, remains unanswered .
The normal mammary gland epithelium is composed of two types of epithelial cells: 1) basal contractile cells (mostly myoepithelial) that are in direct contact with the basement membrane and, 2) secretory luminal cells that face the lumen of ducts/lobules. The phenotype of the normal human mammary gland stem/progenitor cells has been described in various reports as being aldehyde dehydrogenase (ALDH)high , CD10+ [7, 8], CD44highCD24low  or Ep-CAM+/MUC-1neg . Additionally, two other markers have become a standard combination for studying human normal mammary gland cells: EP-CAM (epithelial specific antigen, ESA) and CD49f (α-6-integrin) . Previous reports indicate that Ep-CAMhigh labels luminal epithelial cells where the Ep-CAMhigh/CD49f + fraction contains the luminal progenitors, while Ep-CAMhigh/CD49fneg cells represent the differentiated luminal cells. On the other hand, Ep-CAM-/low/CD49f + phenotype characterize mainly the basal fraction of the human epithelial cells [8, 11]. Whether the aforementioned stem/progenitor cell markers, described in different reports, identify overlapping cell populations and whether they are related to CSC remains unknown.
In this study, we demonstrated that human mammary epithelial cells with CD44high/CD24low phenotype have the highest progenitor ability compared to all other stem/progenitor subpopulations. Furthermore, we have demonstrated that, in both normal and malignant breasts, there are multiple CD44high/CD24low subpopulations. In the majority of breast cancer cases, CSC with CD44high/CD24low phenotype existed in the Ep-CAMhigh/CD49f + fraction of cancer cells. Stem/progenitor markers should be used in combination with Ep-CAM/CD49f reference markers to indicate the pure/specific epithelial stem/progenitor cells. We believe this study may provide a better understanding of breast cancer carcinogenesis as well as facilitate the more accurate identification of CSC. Subsequently, these findings might help in monitoring and/or targeting of this population in the future.
Methods and materials
Patient selection and consenting
This study was conducted in accordance with the Helsinki Declaration and approved by the Research Advisory Council (RAC# 2080-045) of King Faisal Specialist Hospital and Research Centre (KFSH&RC). Normal human mammary gland tissues were obtained from 16 patients admitted to KFSH&RC who underwent reduction mammoplasty with no previous history of breast cancer. Breast cancer samples were obtained from 16 patients diagnosed with invasive ductal carcinoma of the breast (cases were ER + (n = 5), ER/Her2 (n = 3), Her2 (n = 3) and basal tumors (n = 5) ER = ER+/PR+/Her2neg, ER/Her2 = ER+/PR + orneg/Her2+, Her2 = ERneg/PRneg/Her2+, and basal = ERneg/PRneg/Her2neg). All patients signed an informed consent approved by KFSH&RC.
Tissue processing and cell preparation
Processing of breast cancer tissues was performed after routine pathological examination. Gross tissue specimens were macro-dissected by pathologist and frozen sections were examined by hematoxylin staining to ensure that they contained carcinoma cells. Breast tissues (plastic surgery or breast cancer) were kept at 4°C in complete medium, (DMEM medium with 10% fetal bovine serum (FBS)) and processed within 1-2 hours. Tissues were minced, transferred to collagenase digestion medium (Stem Cell Technologies (SCT), Vancouver, Canada) and agitated with an Adams Nutator Mixer (Becton Dickinson, Franklin Lakes, NJ) at 37°C.
Normal Tissue pieces were digested overnight and cell suspensions were centrifuged at 800 g for 8 minutes at 4°C. Fat was aspirated and cell pellets were re-suspended in phosphate buffered saline (PBS). Cells were differentially centrifuged at 120 g for 2 minutes at 4°C to enrich for epithelial cells. The epithelial-enriched pellet was further digested with accutase (SCT) at 37°C (30-60 minutes) until organoids disaggregated into single cells. Epithelial cells were washed once with medium or PBS, filtered through a 70 μm mesh (BD Falcon, Bedford, MA, USA), re-suspended in freezing medium (composed of 90% FBS and 10% DMSO (sigma, St. Louis, Mo, USA)) and stored frozen under liquid nitrogen for later use.
Tumor Tissues were digested, as stated above, for 6 hours followed by centrifugation at 800 g for 8 minutes at 4°C. Cell pellets were further digested with accutase for 10-15 minutes and stored frozen in liquid nitrogen.
Breast cancer cell lines were cultured in DMEM/F12 with the exception of SK-BR-3, which was cultured in McCoy's 5A medium. Both DMEM/F12 and McCoy's 5A media were supplemented with 10% fetal bovine serum and Antibiotics and Antimycotics (all from Invitrogen).
Cells were thawed, washed and, if deemed necessary, 20 to 60 μl of DNase (SCT) was added to disaggregate clumps. Cells were depleted of CD31+ endothelial cells, whenever necessary, using MACS system (Miltenyi Biotec, Germany), as per manufacturer’s instructions. Cells were stained with panels of 8 different antibodies labeled with up to 8 different fluorescent dyes in addition to 4',6-diamidino-2-phenylindole (DAPI, Invitrogen) viability dye. The panels and antibody combination used are listed in Tables 1 and 2 respectively. An ALDH kit (Stem Cell Technologies) was used to stain ALDH population as per manufacture instructions. ALDH staining was performed first followed by the addition of the other antibodies as recommended by the ALDH kit.
The panels of antibodies used to analyze the breast cells
* Replaced with ALDH, CD31-FITC or others. ♦ Replaced in some experiments with c-kit (CD117) ♣Replaced with EPCR (CD201) in some experiments, || replaced with SSEA-4 in some experiments.
Source, clone and dilution of Primary antibodies
Added volume (per 1-6 x 10^6 cells) or dilution
CD49f-PE, FITC, PE-Cy5
EPCR (CD201) PE
ABCG2-Percp-Cy5.5 or APC
CD24 FITC, or PE
CXCR-4 (CD184) PE-Cy7
Ep-CAM-APC, FITC, or PE
* Indirectly labeled with secondary anti-rabbit pacific blue antibody.
Data were acquired using an LSR II Flow Cytometer while a FACSAria was utilized for cell sorting (both from BD, Biosciences, New Jersey, USA) using BD FACSDiva operating software according to guidelines set for analysis and sorting of stem cells by flow cytometry . Positive staining was considered based on the negativity of an isotype control. A minimum of 10,000 events were recorded for all samples. Most phenotypic data were validated using the same antibodies with alternate labels.
We adopted a gating strategy to analyze single viable epithelial mammary cells. To this end, gates were established to exclude debris (using FSC and SSC) and hematopoietic cells (by gating on CD45neg cells) and include only single viable cells (using the relation between FSC-A and FSC-W followed by gating on DAPIneg cells). Gated cells were then examined with Ep-CAM and CD49f antibodies followed by a final gate to select stem/progenitor cells within each Ep-CAM/CD49f epithelial cell fraction. An example of the sequential gating is present in Additional file 1: Figure S1. ALDHhigh positivity was selected based on staining parallel cells with DEAB inhibitor of ALDH activity (Additional file 2: Figure S2). Positivity for antibodies like CD44 and CD24 (as shown by the quadrant) was identified using isotype control (Additional file 2: Figure S2). CD44high expression level was selected arbitrarily to include cells having fluorescence intensity (FI) units greater than 3000 FI (i.e. 2 minor ticks above 103 FI). Similarly, CD24low expression level was set to include cells having FI lower than 3000. Using these criteria we compared the abundance of CD44high/CD24low cells in Ep-CAMhigh or Ep-CAMlow populations. Due to lower expression of CD44 in luminal cells, compared with basal cells, we used the relative CD44high/CD24low of Ep-CAMhigh/CD49f + cells during sorting and in functional assays (to obtain enough cells) and compared them with the bulk of Ep-CAMhigh/CD49f + cells (CD44high expression levels were identified as cells with only 1 tick above 103 FI i.e. 2000 FI). This increased the CD44high/CD24low fraction of luminal Ep-CAMhigh/CD49f + cells from an average of 7% to 23%.
Primary breast cancer single viable cells were selected as above. In addition, lineage negative were selected using the CD10neg, CD31neg, and CD45neg phenotypes and after excluding Ep-CAMneg/CD49fneg mesenchymal cell fraction.
As a quality control, expression patterns were always cross referenced with known phenotypic data of breast cells subsets (e.g. basal and mesenchymal cells are CD24neg, mesenchymal cells are Ep-CAMneg/CD49fneg/MUC-1neg) This assured that our gating strategy was accurate.
Routine immunohistochemistry of formalin-fixed, paraffin-embedded breast cancer samples were evaluated for Her2, estrogen receptor, and progesterone receptor status as reported previously .
Sorted stem/progenitor cell subpopulations were attached to glass slides and stained as described previously . Briefly, cells were cytospin to glass slides by centrifugation at 800 rpm for 3 min. Slides were then air-dried overnight, acetone fixed for 15 min and stored at -80 C until stained. For staining, the cell membrane was permeabilized with 0.5% triton-X (Sigma) followed by overnight incubation with primary antibodies. After washing, Envision + polymer (ready to use; Dako) was used as a secondary antibody. Color was developed with 3,3′-diaminobenzidine (DAB) and instant hematoxylin (Shandon) was used for counterstaining.
Quantification of estrogen receptor (ER) positive cells, in normal sorted epithelial breast cells, was achieved by counting the number of ER + cells in several high magnification (x500) fields. The percentage was obtained by dividing the number of estrogen receptor positive cells by the total number of cells examined.
Mammosphere formation assay
Cells were seeded in ultra-low attachment plates (Corning, Tewksbury, MA, USA) at a density of 1000 viable cells/well in 96-well plate in 120 μL/well of medium composed of DMEM/F12 medium supplemented with epidermal growth factor (20 ng/mL), hydrocortisone, (500 ng/mL), and Insulin (5 μg/mL), (all from Sigma) as well as B27 (1:50) and antibiotic/antimycotics (1:100) from Invitrogen (Grand Island, NY, USA) .
In addition to counting the number of mammospheres formed, we also measured the size of mammospheres by summing the volume (calculated as =4/3*π*r3 where r = radius of each mammosphere) of all formed mammospheres per cell group. This enabled us to assess the progenitor ability of the cells and to minimize the effect of mammosphere aggregation. Results are displayed as total sphere volume in microns3 or, where possible, mammosphere formation was normalized to the total mammosphere size of one of the Ep-CAM/CD49f main cell populations (i.e. Ep-CAMlow/CD49f+, Ep-CAMhigh/CD49f + cells).
Colony formation assay
Sorted cells were cultured at a density of 300 to 500 cells/cm2 on irradiated NIH 3T3 mouse cells seeded 1 day prior at a density of 50,000 cells/cm2. Cells were cultured in serum containing (2%) Ep-Wang medium  for 24 hours followed by medium exchange to serum free medium (Ep-sfm) . Similar experiments utilizing Epi-Cult B medium (SCT) revealed comparable data. At the end of the experiment, colonies were counted under phase contrast microscope or directly after fixation with 4% PFA and staining with Giemsa (Fisher Scientific).
Mouse xenotransplantation studies
All animal work, including anesthesia and euthanasia, was done in accordance to protocols approved by the Animal Care and Use Committee (ACUC) of KFSH&RC. Sorted (30, 000) MDA-MB-468 breast cancer cells were suspended in 50 uL FBS, mixed (1:1) with matrigel (BD Biosciences) and injected subcutaneously (MDA-MB-468).
Significance in expression or mammosphere formation was determined by T-test using Excel software. P < 0.05 was used to indicate significance. Correlation coefficient was also calculated using Excel software. Error bars are presented as standard error of the mean (SEM).
In this study we used the Ep-CAM/CD49f antibody combination as a common reference to correlate CSC with normal stem/progenitors. We have characterized the normal breast Ep-CAM/CD49f epithelial fractions and their subpopulations then compared them with breast cancer cells.
Breast epithelial stem/progenitor cells are limited to CD49f + cell fractions
At First, we have re-established the stem/progenitor ability of the Ep-CAM/CD49f cell populations. After exclusion of stromal cells, Ep-CAM/CD49f staining displayed three distinct epithelial cell populations designated here for simplicity as A,B and C: Ep-CAM-/low/CD49f + (A), Ep-CAMhigh/CD49f + (B), and Ep-CAMhigh/CD49fneg (C) (Figure 1A). Epithelial cell fractions were sorted and their stem/progenitor cell features were assessed using mammosphere and colony forming assays. Figure 1B shows that population A (Ep-CAM-/low/CD49f + cells) were the most efficient in forming mammospheres, consistent with being the source of mammary stem cells, followed by population B (Ep-CAMhigh/CD49f + cells). On the other hand, population C (Ep-CAMhigh/CD49fneg) cells did not form mammospheres. Consistently, in colony forming assays, population C did not form colonies while populations A and B gave typical basal (or mixed) and luminal colonies respectively (Additional file 3: Figure S3).
Altogether, these results confirm that basal Ep-CAM-/low/CD49f + (A) and luminal progenitor Ep-CAMhigh/CD49f + (B) cell fractions contain the stem/progenitor epithelial cells while Ep-CAMhigh/CD49fneg (C) contain only differentiated cells. Therefore, we focused thereafter on CD49f + populations (A and B).
CD44high/CD24low epithelial cells showed the highest progenitor ability
Previous reports have used either one or two markers to identify stem/progenitor cells in the human mammary gland. However, to date, the relationship between different markers has not been well characterized. In order to link these stem/progenitor markers together we have simultaneously tested their expression levels in relation to Ep-CAM/CD49f profile using multi-parametric (up to 9 colors) cell sorting.
CD44high/CD24low phenotype within Ep-CAM-/low/CD49f + cells (A) enriches for basal progenitors
Basal Ep-CAM-/low/CD49f + cells expressed: CD10 (57 ± 6%), CD44high/CD24low (21 ± 4%) and Ep-CAM+/MUC-1neg (25 ± 7%), but were ALDHneg/or low (Figure 2A). In order to relate these markers to stem/progenitor cell function we sorted each stem/progenitor subpopulation within the Ep-CAM-/low/CD49f + group and examined their mammosphere and colony forming abilities (A representative gating strategy is presented in Additional file 1: Figure 1). All three stem/progenitor cell markers were effective in selecting for mammosphere forming cells. However, CD44high/CD24low cells formed more mammospheres than other subpopulations, although the difference was not significant except between CD44high/CD24low cells and CD10+ fractions (Figure 2B). Interestingly, when mammosphere size, which measure cell progenitor ability , was considered, CD44high/CD24low formed significantly larger mammospheres than the other stem/progenitor subpopulations of Ep-CAM-/low/CD49f cells (Figure 2C). Consistently, colony forming assays showed cells with the CD44high/CD24low phenotype have the highest number of basal (or mixed) colonies, while CD10neg formed the lowest number of these colonies within the Ep-CAMlow/CD49f + cell population (Figure 2D). Cells expressing all three stem/progenitor cell markers simultaneously (ALL) within Ep-CAM-/low/CD49f + cells did not form more mammospheres or colonies than CD44high/CD24low cells suggesting that CD44high/CD24low phenotype was accurate enough to select for basal progenitors (Figure 2C&2D).
These data indicate that within the basal Ep-CAM-/low/CD49f + (A) population, CD44high/CD24low cells have the highest progenitor ability while, CD10neg cells have the least progenitor ability (i.e. differentiated myoepithelial cells).
CD44high/CD24low phenotype within Ep-CAMhigh/CD49f + cells (B) selects for luminal progenitors
Luminal Ep-CAMhigh/CD49f + (B) cells expressed: ALDHHigh (34 ± 4%), CD44high/CD24low (7 ± 1%) and Ep-CAMhigh/MUC-1neg (67 ± 5%) markers, while cells in this population were mostly CD10neg (Figure 3A). These stem/progenitor marker(s) positive subsets, within the main Ep-CAMhigh/CD49f + population, were sorted to evaluate their mammosphere and colony forming abilities. Results show that cells with ALL followed by cells with CD44high/CD24low phenotype demonstrated the highest mammosphere forming ability when both the number (Figure 3B) and the size of mammospheres were considered (Figure 3C), while ALDHneg/low showed the least ability to form mammospheres or colonies (Figure 3B&C). Likewise, ALL and CD44high/CD24low cells gave the highest percentages of colony forming cells (almost all luminal), although the difference was not statistically significant from other subpopulations (Figure 3C).
This demonstrates that luminal cells having the CD44high/CD24low phenotype have the highest progenitor ability when used in combination with CD49f expression status.
CD44high/CD24low cells within Ep-CAMhigh/CD49f + population correlates with ER negative status
Luminal mammary epithelial cells have a population of estrogen receptor (ER) positive cells that become the majority in hormone receptor positive breast cancers. On the other hand, in the normal breast, proliferating luminal cells are estrogen receptor negative . Therefore, in order to identify these ER negative cells (presumably proliferating progenitor cells) we investigated the ER status of cells expressing stem/progenitor cell markers within the luminal Ep-CAMhigh/CD49f + (B) population. Immunohistochemistry of sorted cells revealed that within population B, cells with CD44high/CD24low phenotype had the lowest number of ER positive cells, consistent with their progenitor ability (Figure 3E). Cells from Ep-CAMhigh/CD49fneg population (population C), known to have the highest percentage of ER positive cells, were used as a positive control while cells from the Ep-CAM-/low/CD49f + population were used as a negative control . These data show correlation between CD44high/CD24low and estrogen receptor negativity in normal mammary epithelial cells within Ep-CAMhigh/CD49f + cells (B), which is consistent with their progenitor ability.
CD44high/CD24low epithelial cells within Ep-CAMlhigh/CD49fnegcells (population C) lack stem/progenitor ability
Finally, we measured the expression level of the breast stem/progenitor cell markers in the Ep-CAMhigh/CD49fneg cell population (C). Both CD44high/CD24low and Ep-CAM+/MUC-1neg phenotypes were expressed at 11 ± 3% and 16 ± 3% respectively (Additional file 4: Figure S4). On the other hand, there were neither ALDHhigh (majority of cells were ALDHlow) nor CD10+ cells in this population. Even though Ep-CAMhigh/CD49fneg cells expressed CD44high/CD24low and Ep-CAM+/MUC-1neg markers, they did not form mammospheres or colonies in vitro. This emphasizes that CD44high/CD24low and Ep-CAM+/MUC-1neg, previously described stem/progenitor markers, could not select for progenitor cells when used alone, and that they should to be used in combination with Ep-CAM/CD49f profile.
These results altogether, demonstrate that there are multiple subpopulations of progenitor cells within each Ep-CAM/CD49f cell group. The CD44high/CD24low cells in both basal and luminal CD49f + fractions had the highest progenitor ability in each cell type respectively. Ep-CAMhigh/CD49fneg, although expressing some of the stem/progenitor cell markers, were differentiated cells as they failed to form colonies and mammospheres in vitro (data summarized in Table 3).
Characterization of Ep-CAM/CD49f populations and the stem/progenitor cell markers they express
Has colony forming cells
Expressed "stem cell" markers
Other “stem/progenitor” cell markers do not show any stem/progenitor enrichment ability over CD49f + cells
Besides the above examined common breast stem/progenitor cell markers, we evaluated additional markers that have been associated with stem/progenitor cells in the breast (CD133, CXCR-4, SSEA-4, c-kit, EPCR, ABCB1 and ABCG2), within the Ep-CAM and CD49f cell fractions. All the examined stem/progenitor cell markers were expressed by Ep-CAMhigh/CD49f + luminal progenitor cells, or Ep-CAM-/low/CD49f + basal cells. Functionally, none of these markers could further enrich for stem/progenitor cells over the main Ep-CAMlow/CD49f + basal or Ep-CAMhigh/CD49f + luminal populations, as assessed by mammosphere formation (Additional file 5: Figure S5) and colony forming assays (data not shown). This indicates that these putative stem/progenitor cell markers did not enrich for stem/progenitor cells above CD49f + alone.
Cancer stem cells can best be enriched using combination of CD44high/CD24low and Ep-CAMhigh/CD49+ markers
Subsequently, we sought to compare the stem/progenitor cell populations between normal mammary epithelial cells and breast cancer cells.
The majority of breast cancer cells have luminal phenotype
While the standard profile for normal human mammary epithelial cells depends on the expression of Ep-CAM and CD49f, such data is currently not available for breast cancer cells. Therefore, we compared the Ep-CAM/CD49f expression patterns of normal mammary epithelial cells with primary tumor cells obtained from breast cancer patients. Our results show a clear drift in primary breast cancer cells towards population C (Ep-CAMhigh/CD49fneg), which almost doubled, while population A (Ep-CAM-/low/CD49f+) decreased dramatically in cancer cells - to less than one fourth of its normal counterpart (Figure 4A). As population A was barely present among the breast cancer cells, and existed in few patient samples, we focused on population B and C (i.e. Ep-CAMhigh/CD49f + and Ep-CAMhigh/CD49fneg respectively) as they constituted the vast majority, if not all, of the tumor cells in primary breast cancer samples.
Both CD49fneg and CD49f + cancer cells express stem/progenitor markers
We examined more closely the expression of the stem/progenitor cell markers in population C (Ep-CAMhigh/CD49fneg) and population B (Ep-CAMhigh/CD49f+) of tumor cells. We found no significant difference in the percentage of CD44high/CD24low tumor cells among population C (Ep-CAMhigh/CD49fneg) and population B (Ep-CAMhigh/CD49f+) (Figure 4B). In contrast, the vast majority of ALDHhigh cells were among the CD49f + stained cells. This clearly shows a phenotypic similarity in the distribution of stem/progenitor cell markers between primary breast cancer cells and normal epithelial cells. In contrast, unlike normal epithelial cells, there was no statistically significant difference in the percentage of Ep-CAM+/MUC-1neg cells between Ep-CAMhigh/CD49fneg and Ep-CAMhigh/CD49f + cells, implying an alteration in the MUC-1 expression upon carcinogenesis (Figure 4B). We further stratified the 16 breast cancer samples analyzed into the four main subtypes of breast cancer: ER (luminal A), ER/Her2 (luminal B), Her2 and Basal. There was no statistically significant difference in the expression of the stem/progenitor markers between the four types of breast cancer, probably due the small number of samples analyzed (data not shown). However, despite a small sample size, there was significant difference (P = 0.01) in CD44high/CD24low expression among Ep-CAMhigh breast cancer cells between ER and Basal subtypes of breast cancer. This difference was only present among CD49f + cancer cells (Figure 4C). This suggests that CD49f, if used in combination, with CD44high/CD24low might be able to link stem/progenitor cell markers with breast cancer subtypes.
Cancer stem cells are abundant in primary CD44high/CD24low/Ep-CAMhigh/CD49+ cancer cells
We then functionally tested the stem/progenitor cell ability of the above examined subpopulations. We assessed the mammosphere formation of sorted CD44high/CD24low or ALDHhigh cells that were further fractionated from either population C (Ep-CAMhigh/CD49fneg) or population B (Ep-CAMhigh/CD49f+) cancer cells. Stem/progenitor cell subpopulations within CD49f + fractions of cancer cells formed more mammospheres than their CD49fneg cell counterparts (Figure 4D). In addition the size of the mammospheres formed in CD49f + cancer cells were considerably larger than CD49fneg cancer cells (Figure 4E). The CD44high/CD24low cancer cells tended to form more mammospheres than ALDHhigh-although the difference was not statistically significant. This demonstrates that CSC can best be enriched by selecting for tumor cells with the CD44high/CD24low or ALDHhigh phenotypes within Ep-CAMhigh/CD49f + cancer cells.
Breast cancer cell lines are mostly Ep-CAMhigh/CD49+
Breast cancer cell lines are frequently used as a model to study breast cancer cells. Therefore, we have examined the phenotype of 9 commonly used breast cancer cell lines using the Ep-CAM/CD49f reference markers. Interestingly, similar to primary breast cancer cells 6 out of 9 cell lines had the Ep-CAMhigh luminal phenotype (Additional file 6: Figure S6A). We then examined the expression levels of the stem/progenitor cell markers in these cell lines. Importantly, there was a correlation between the percentage of cells with CD44high/CD24low phenotype and ER negativity status (correlation coefficient 0.63) (Additional file 6: Figure S6B).
Among all tested cell lines: MDA-MB-468 remarkably expressed the four major stem/progenitor cell markers previously examined i.e. CD44high/CD24low, ALDHhigh, Ep-CAM+/MUC-1neg and CD10 (Additional file 6: Figure S6B). We therefore decided to use this cell line as a model to compare the four stem/progenitor markers. We sorted then injected low cell numbers from each stem/progenitor subpopulation of MDA-MB-468 cells into NOD/SCID mice. After 4 weeks only CD44high/CD24low and ALDHhigh cells formed tumors (in 4/4 of the injected mice). This shows that CD44high/CD24low were comparable to ALDHhigh in enriching for cancer stem cells in breast cancer cell lines.
These results collectively show that CD44high/CD24low were comparable to ALDHhigh phenotype in selecting for cancer stem cells both in primary, as well as established breast cancer cell lines. However, while these stem cell markers can be used alone in cell lines (vast majority are CD49f+), the primary CD44high/CD24low breast cancer cells existed in both CD49fneg and CD49f + cancer cell fractions. Therefore, in primary breast cancer samples, stem cell markers should be used in combination with Ep-CAM/CD49f antibodies.
The presence of stem/progenitor cell populations in the human breast has been well documented in several reports [8, 9, 18, 19]. Several markers have been used to identify this population including Ep-CAM-/low/CD49f + , ALDH + , CD44high/CD24low, CD10+ , or Ep-CAM+/MUC-1neg. However, to date there is no study with detailed comparison between these markers. Our approach was to compare phenotypically and functionally most of the previously reported stem/progenitor cell markers side-by-side in reference to Ep-CAM/CD49f profile. We found for the first time that CD44high/CD24low mammary cells exhibited the highest stem/progenitor ability, both in normal and malignant breast cells, when combined with Ep-CAM/CD49f markers. We have used multi-parametric (up to 9 colors) fluorescence-activated cell sorting (FACS) coupled with several in vitro and in vivo assays to compare the progenitor/tumorigenic ability of the different stem/progenitor subpopulations of the human breast. Importantly, in this study, we have used uncultured/unmanipulated cells in contrast to several previous reports based on cells cultured for 3 days prior to analysis [8, 10].
Since the pioneering work of Al-Hajj et al  on the phenotype of CSC as CD44high/CD24low, multiple studies have suggested that these markers did not correlated with the survival of breast cancer patients [21–23]. We have demonstrated in this study that CD44high/CD24low from CD49f + cancer cells formed more mammospheres than CD49fneg cancer cells. This indicates that it is necessary to use CD49f in combination with CD44high/CD24low panel. In agreement, Cariati et al  have shown that only CD49f + MCF-7 form mammospheres and induce tumors in mice, and not CD49fneg cells. Interestingly, very recent clinical data (utilizing large sample size of breast cancer patients) by Ali et al  have shown that CD44high/CD24low or ALDHhigh in combination with the CD49f positivity correlate with patient survival.
We have reported an abundance of CD44high/CD24low in the human mammary gland of 21 ± 4% in Ep-CAMlow/CD49f + and 7 ± 1% in Ep-CAMhigh/CD49f + cell populations. A recent study has reported that CD44high/CD24low subpopulation is restricted to Ep-CAMlow/CD49f + fraction of cells . This discrepancy is most likely due to the CD44high gating used. To our knowledge, there is no standard criterion in the literature to describe CD44high and CD24low gates (The criteria we used is clearly described in the methods and materials section). However, regardless of the gate used, we have established here that the relative fraction CD44high/CD24low among Ep-CAMhigh/CD49f + cell population was able to enrich for colony forming cells more than 2 times the bulk of Ep-CAMhigh/CD49f + cells. This indicates that the gate for CD44high/CD24low does not have to be the same for Ep-CAMhigh and Ep-CAMlow in order to select for epithelial progenitors. This further supports the importance of using CD44high/CD24low phenotype in combination with Ep-CAM/CD49 reference markers.
In this report, we have demonstrated that the majority of cancer cells showed a luminal Ep-CAMhigh phenotype, with very small percentage of cancer cells of Ep-CAM-/low/CD49f + phenotype. Similarly, 6 out of 9 breast cancer cell lines had Ep-CAMhigh phenotype. These findings are consistent with the previously described luminal phenotype of cancer cells based on strong correlative evidence showing breast cancer cells express luminal markers (such as MUC-1, Keratins 18 and 19 ), and lack basal markers (like CD10 and α-SMA ). In addition, Ince et al  established that transformed luminal cells (BPLER) were able to form tumors from as little as 102 cells, while as many as 106 cells were needed to form tumor from (HMLER) basal enriched transformed cells. This indicates that luminal cells are more tumorigenic than basal cells. In addition, Fillmore et al  have shown that CD44high/CD24low within Ep-CAM + and not Ep-CAMneg breast cancer cell lines have cancer stem cell features including colony formation and tumorigenicity in NOD/SCID mice. Unfortunately, we could not characterize Ep-CAMlow primary cancer cells although they are interesting population as they correspond to the mammary stem cell enriched population in the normal mammary gland. This is because cancer cells, with Ep-CAMlow/CD49f + phenotype, were present only in some patients and represented small percentages making them practically infeasible for us to study them.
The cell of origin of most of breast cancers still remains unknown. Keller et al  have demonstrated that transforming Ep-CAMhigh luminal cells produce breast cancers commonly seen in the clinic, while transforming normal mammary basal Ep-CAMlow/CD10+ cells produces a rare undifferentiated metaplastic type of breast cancer. Our results illustrate a similarity of phenotype between CSC and normal luminal progenitors. In addition, the phenotype of cancer cell lines being Ep-CAMhigh/CD49f + might further suggest the origin of CSC from Ep-CAMhigh/CD49f + normal luminal progenitors, at least in some cases of breast cancer. In agreement, Lim et al  and Molyneux et al  have demonstrated that the cell of origin of BRCA1 defective breast carcinomas, a predominantly triple negative type of breast cancer, is also in the Ep-CAMhigh/CD49f + luminal progenitor cells. Similarly, Lo et al  has shown the Ep-CAMhigh/CD49f + cells are the cell of origin of Her2/neu mouse model.
On the other hand, histological observation of the abundance of CD44high/CD24low cells in the normal basal layer of the breast (anatomically in close contact with basement membrane)  has encouraged others to speculate that CD44high/CD24low cancer cells might originate from the Ep-CAMlow normal basal layer . Our results suggest that the luminal layer might be an additional source for CD44high/CD24low cancer cells. This is further supported by our finding that only the luminal fraction of CD44high/CD24low cells overlaps with ALDHhigh cells. Overall, underscores the importance of using stem cell markers CD44high/CD24low in combination with Ep-CAM/CD49f.
In the normal human mammary gland we have validated and compared side-by-side many breast stem/progenitor cell markers and found that among them only ALDHhigh, CD10+, CD44high/CD24low Ep-CAM+/MUC-1neg can enrich for stem/progenitor cells over CD49f + alone. CD44high/CD24low had the highest ability to enrich for cell progenitors when used in combination with Ep-CAM/CD49f antibodies in order to differentiate between basal or luminal progenitors. Similarly, in breast cancer CD44high/CD24low (as well as ALDHhigh) showed the highest ability to enrich for CSC. When normal breast stem/progenitor populations are compared with their counterparts in breast cancer, there were similarities and differences between stem/progenitor cells in normal and malignant breast. In both normal and cancer cells there was a correlation between CD44high/CD24low phenotype and estrogen receptor negative status. In addition, there were cells with the CD44high/CD24low phenotype in both Ep-CAMhigh/CD49f + and Ep-CAMhigh/CD49fneg cell populations. Furthermore, ALDH was highly expressed by Ep-CAMhigh/CD49f + cells in both normal and malignant cells. On the other hand, breast cancer cells had mainly luminal phenotype, with an increase in the CD49fneg fraction compared with normal breast which exhibited balanced populations of luminal (both differentiated and progenitor) and basal cells. The CD49fneg cells in normal breasts could not form mammospheres, while in malignant breast they formed mammospheres, albeit to a much lesser extent than CD49f + cells (summarized in Figure 5). CD44high/CD24low is expressed by both CD49f + and CD49fneg cancer cells. However, CD44high/CD24low/CD49f + had significantly higher stem/progenitor ability as measured by mammosphere formation thus proposing that these cells are the best phenotype to identify breast CSC.
These findings may provide a better understanding of how CSC evolve, and which population to target and monitor during therapy, a leading step to eradicate this disease at its root.
HG: Conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, GS: collection and/or assembly of data, data analysis and interpretation, PM, AA, EB and FA: Collection and/or assembly of data, KA (plastic surgeon): Design and coordination/selection of patients' samples, CA: Conception and design, supervision of data analysis and interpretation and final approval/revision of manuscript. ALL authors read and approved the manuscript. All contributing authors approve the submission of this version of the manuscript and assert that the document represents valid work. All contributing authors have no disclosures to make.
Colony forming assay
Cancer stem cell
Forward light scatter
Side light scatter
Forward light scatter width.
We acknowledge the support of Dr. Sultan T. Al-Sedairy, our Executive Director for his guidance and vision. We wish to acknowledge the support from the Logistics and Facilities Management Office and Research Centre Administration at King Faisal Specialist Hospital & Research Centre. This work was supported by grants from King Faisal Specialist Hospital and Research Centre and King Abdulaziz City for Science and Technology (RAC# 2080-045/AT-24-29 and RAC# 2080-023/MED 483-20). We would like to thank Drs. Monther Al-Alwan, Ameera Gaafar, Andrew Wetzig and Ayodele Alaiya for reviewing the manuscript.
Stem Cell & Tissue Re-engineering Program, King Faisal Specialist Hospital & Research Centre
Stem Cell & Tissue Re-engineering Program, Departments of Comparative Medicine, King Faisal Specialist Hospital and Research Centre
Stem Cell & Tissue Re-engineering Program, Departments of Pathology, King Faisal Specialist Hospital and Research Centre
Stem Cell & Tissue Re-engineering Program, Department of Surgery, King Faisal Specialist Hospital and Research Centre
College of Medicine, Al-Faisal University
Transplantation Research Center (TRC), Brigham & Women’s Hospital and Children’s Hospital Boston, Harvard Medical School
Stem Cell & Tissue Re-engineering Program, Research Centre, King Faisal Specialist Hospital and Research Centre
Frank NY, Margaryan A, Huang Y, Schatton T, Waaga-Gasser AM, Gasser M, Sayegh MH, Sadee W, Frank MH: ABCB5-mediated doxorubicin transport and chemoresistance in human malignant melanoma.Canc Res 2005,65(10):4320–4333.View Article
Li X, Lewis MT, Huang J, Gutierrez C, Osborne CK, Wu MF, Hilsenbeck SG, Pavlick A, Zhang X, Chamness GC, et al.: Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy.J Natl Canc Inst 2008,100(9):672–679.View Article
Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN: Glioma stem cells promote radioresistance by preferential activation of the DNA damage response.Nature 2006,444(7120):756–760.PubMedView Article
Visvader JE, Lindeman GJ: Cancer stem cells in solid tumours: accumulating evidence and unresolved questions.Nat Rev Canc 2008,8(10):755–768.View Article
Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF: Prospective identification of tumorigenic breast cancer cells.Proc Natl Acad Sci USA 2003,100(7):3983–3988.PubMedView Article
Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown M, Jacquemier J, Viens P, Kleer CG, Liu S, et al.: ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome.Cell Stem Cell 2007,1(5):555–567.PubMedView Article
Eirew P, Stingl J, Raouf A, Turashvili G, Aparicio S, Emerman JT, Eaves CJ: A method for quantifying normal human mammary epithelial stem cells with in vivo regenerative ability.Nat Med 2008,14(12):1384–1389.PubMedView Article
Dontu G, Abdallah WM, Foley JM, Jackson KW, Clarke MF, Kawamura MJ, Wicha MS: In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells.Genes Dev 2003,17(10):1253–1270.PubMedView Article
Stingl J, Raouf A, Emerman JT, Eaves CJ: Epithelial progenitors in the normal human mammary gland.J Mammary Gland Biol Neoplasia 2005,10(1):49–59.PubMedView Article
Lim E, Vaillant F, Wu D, Forrest NC, Pal B, Hart AH, Asselin-Labat ML, Gyorki DE, Ward T, Partanen A, et al.: Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers.Nat Med 2009,15(8):907–913.PubMedView Article
Alexander CM, Puchalski J, Klos KS, Badders N, Ailles L, Kim CF, Dirks P, Smalley MJ: Separating stem cells by flow cytometry: reducing variability for solid tissues.Cell Stem Cell 2009,5(6):579–583.PubMedView Article
Fillmore CM, Kuperwasser C: Human breast cancer cell lines contain stem-like cells that self-renew, give rise to phenotypically diverse progeny and survive chemotherapy.Breast Canc Res 2008,10(2):R25.View Article
Ghebeh H, Mohammed S, Al-Omair A, Qattan A, Lehe C, Al-Qudaihi G, Elkum N, Alshabanah M, Bin Amer S, Tulbah A, et al.: The B7-H1 (PD-L1) T lymphocyte-inhibitory molecule is expressed in breast cancer patients with infiltrating ductal carcinoma: correlation with important high-risk prognostic factors.Neoplasia 2006,8(3):190–198.PubMedView Article
Ghebeh H, Tulbah A, Mohammed S, Elkum N, Bin Amer SM, Al-Tweigeri T, Dermime S: Expression of B7-H1 in breast cancer patients is strongly associated with high proliferative Ki-67-expressing tumor cells.Int J Canc 2007,121(4):751–758.View Article
Wang CS, Goulet F, Lavoie J, Drouin R, Auger F, Champetier S, Germain L, Tetu B: Establishment and characterization of a new cell line derived from a human primary breast carcinoma.Canc Genet Cytogenet 2000,120(1):58–72.View Article
Clarke RB, Howell A, Potten CS, Anderson E: Dissociation between steroid receptor expression and cell proliferation in the human breast.Canc Res 1997,57(22):4987–4991.
Stingl J: Detection and analysis of mammary gland stem cells.J Pathol 2009,217(2):229–241.PubMedView Article
Bachelard-Cascales E, Chapellier M, Delay E, Pochon G, Voeltzel T, Puisieux A, Caron De Fromentel C, Maguer-Satta V: The CD10 enzyme is a key player to identify and regulate human mammary stem cells.Stem Cells 2010,28(6):1081–1088.PubMedView Article
Shipitsin M, Campbell LL, Argani P, Weremowicz S, Bloushtain-Qimron N, Yao J, Nikolskaya T, Serebryiskaya T, Beroukhim R, Hu M, et al.: Molecular definition of breast tumor heterogeneity.Canc Cell 2007,11(3):259–273.View Article
Abraham BK, Fritz P, McClellan M, Hauptvogel P, Athelogou M, Brauch H: Prevalence of CD44+/CD24-/low cells in breast cancer may not be associated with clinical outcome but may favor distant metastasis.Clin Canc Res 2005,11(3):1154–1159.
Mylona E, Giannopoulou I, Fasomytakis E, Nomikos A, Magkou C, Bakarakos P, Nakopoulou L: The clinicopathologic and prognostic significance of CD44+/CD24(-/low) and CD44-/CD24+ tumor cells in invasive breast carcinomas.Hum Pathol 2008,39(7):1096–1102.PubMedView Article
Ahmed MA, Aleskandarany MA, Rakha EA, Moustafa RZ, Benhasouna A, Nolan C, Green AR, Ilyas M, Ellis IO: A CD44(-)/CD24(+) phenotype is a poor prognostic marker in early invasive breast cancer.Breast Canc Res Treat 2012,133(3):979–995.View Article
Cariati M, Naderi A, Brown JP, Smalley MJ, Pinder SE, Caldas C, Purushotham AD: Alpha-6 integrin is necessary for the tumourigenicity of a stem cell-like subpopulation within the MCF7 breast cancer cell line.Int J Canc 2008,122(2):298–304.View Article
Ali HR, Dawson SJ, Blows FM, Provenzano E, Pharoah PD, Caldas C: Cancer stem cell markers in breast cancer: pathological, clinical and prognostic significance.Breast Canc Res 2011,13(6):R118.View Article
Shehata M, Teschendorff A, Sharp G, Novcic N, Russell IA, Avril S, Prater M, Eirew P, Caldas C, Watson CJ: Phenotypic and functional characterisation of the luminal cell hierarchy of the mammary gland.Breast Canc Res 2012,14(5):R134.View Article
Gusterson BA, Warburton MJ, Mitchell D, Ellison M, Neville AM, Rudland PS: Distribution of myoepithelial cells and basement membrane proteins in the normal breast and in benign and malignant breast diseases.Canc Res 1982,42(11):4763–4770.
Kalof AN, Tam D, Beatty B, Cooper K: Immunostaining patterns of myoepithelial cells in breast lesions: a comparison of CD10 and smooth muscle myosin heavy chain.J Clin Pathol 2004,57(6):625–629.PubMedView Article
Ince TA, Richardson AL, Bell GW, Saitoh M, Godar S, Karnoub AE, Iglehart JD, Weinberg RA: Transformation of different human breast epithelial cell types leads to distinct tumor phenotypes.Canc Cell 2007,12(2):160–170.View Article
Keller PJ, Arendt LM, Skibinski A, Logvinenko T, Klebba I, Dong S, Smith AE, Prat A, Perou CM, Gilmore H, et al.: Defining the cellular precursors to human breast cancer.Proc Natl Acad Sci USA 2012,109(8):2772–2777.PubMedView Article
Molyneux G, Geyer FC, Magnay FA, McCarthy A, Kendrick H, Natrajan R, Mackay A, Grigoriadis A, Tutt A, Ashworth A, et al.: BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells.Cell Stem Cell 2010,7(3):403–417.PubMedView Article
Lo PK, Kanojia D, Liu X, Singh UP, Berger FG, Wang Q, Chen H: CD49f and CD61 identify Her2/neu-induced mammary tumor-initiating cells that are potentially derived from luminal progenitors and maintained by the integrin-TGFbeta signaling.Oncogene 2012,31(21):2614–2626.PubMedView Article
Bankfalvi A, Terpe HJ, Breukelmann D, Bier B, Rempe D, Pschadka G, Krech R, Lelle RJ, Boecker W: Immunophenotypic and prognostic analysis of E-cadherin and beta-catenin expression during breast carcinogenesis and tumour progression: a comparative study with CD44.Histopathology 1999,34(1):25–34.PubMedView Article
Fillmore C, Kuperwasser C: Human breast cancer stem cell markers CD44 and CD24: enriching for cells with functional properties in mice or in man?Breast Canc Res 2007,9(3):303.View Article
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.