Conceived and designed the experiments: DD AR. Performed the experiments: DD MS ANP VK RG. Analyzed the data: DD MS ANP VK RG AR. Contributed reagents/materials/analysis tools: VMK GM AR. Wrote the paper: DD AR.
The authors have declared that no competing interests exist.
Cancer stem cells exhibit close resemblance to normal stem cells in phenotype as well as function. Hence, studying normal stem cell behavior is important in understanding cancer pathogenesis. It has recently been shown that human breast stem cells can be enriched in suspension cultures as mammospheres. However, little is known about the behavior of these cells in long-term cultures. Since extensive self-renewal potential is the hallmark of stem cells, we undertook a detailed phenotypic and functional characterization of human mammospheres over long-term passages.
Single cell suspensions derived from human breast ‘organoids’ were seeded in ultra low attachment plates in serum free media. Resulting primary mammospheres after a week (termed T1 mammospheres) were subjected to passaging every 7th day leading to the generation of T2, T3, and T4 mammospheres.
We show that primary mammospheres contain a distinct side-population (SP) that displays a CD24low/CD44low phenotype, but fails to generate mammospheres. Instead, the mammosphere-initiating potential rests within the CD44high/CD24low cells, in keeping with the phenotype of breast cancer-initiating cells. In serial sphere formation assays we find that even though primary (T1) mammospheres show telomerase activity and fourth passage T4 spheres contain label-retaining cells, they fail to initiate new mammospheres beyond T5. With increasing passages, mammospheres showed an increase in smaller sized spheres, reduction in proliferation potential and sphere forming efficiency, and increased differentiation towards the myoepithelial lineage. Significantly, staining for senescence-associated β-galactosidase activity revealed a dramatic increase in the number of senescent cells with passage, which might in part explain the inability to continuously generate mammospheres in culture.
Thus, the self-renewal potential of human breast stem cells is exhausted within five in vitro passages of mammospheres, suggesting the need for further improvisation in culture conditions for their long-term maintenance.
Stem cells are found in many adult tissues and play a critical role in tissue homeostasis throughout life. Although few in an organism, these cells can repopulate an entire tissue when required, such as during injury or disease, by virtue of their long term self-renewal potential and the ability to generate heterogeneous progeny
Both normal and cancer stem cells have been identified in the mammary gland
Evidence for the existence of stem cells in human mammary gland came from studies of X chromosome inactivation. Adjacent patches of the epithelium had inactivation of the same X chromosome, indicating that these cells arose from a single primitive cell
The inability to maintain adult stem cells in an undifferentiated state in vitro has further marred their characterization. Recently, enrichment of human breast stem/progenitor cells as non-adherent mammospheres in serum-free media was demonstrated
In order to better comprehend the mammosphere system for its potential use in long-term maintenance of stem cells in culture, we have undertaken a thorough phenotypic and functional characterization of mammospheres. We show that the side population and sphere forming potential resides within two distinct subpopulations of CD24/44 dual stained mammospheres. In serial sphere formation assays, we find variations in size, number, proliferation, and differentiation status of mammospheres over passages. Importantly, we fail to detect generation of new mammospheres beyond the fifth passage despite the presence of live, proliferating and label retaining cells. Staining for senescence-associated β-galactosidase activity revealed a dramatic increase in the number of senescent cells with passage, which might in part explain the inability of these cells to continuously generate mammospheres in culture.
Normal breast tissue was obtained from Kidwai Memorial Institute of Oncology (KMIO) Bangalore, as per the Institutional Review Board, in compliance with the ethical guidelines of KMIO and the Indian Institute of Science. Patient consent was obtained in a written form prior to surgery, as per the protocol approved by the IRB of KMIO. Tissue was collected 6 cm away from or diagonally opposite to tumor site by pathologists at KMIO from mastectomy cases. H & E staining on the tissues confirmed their normal state. Tissues were collected in DMEM (minus phenol red) containing penicillin, streptomycin, gentamycin and fungizone and dissociated mechanically and enzymatically using 1 mg/ml Collagenase (Sigma Aldrich) and 100 U/ml Hyaluronidase (Calbiochem) at 37°C for 16–18 hours with rotation. Breast organoids were separated by differential centrifugation and seeded in 50-mm low attachment plate in serum free DMEM-F12 media containing 10 ng/ml hEGF, 1 µg/ml hydrocortisone, 10 µg/ml insulin, 20 ng/ml bFGF, 4 ng/ml heparin (Sigma Aldrich), B27 (Invitrogen) as described earlier
The organoids were enzymatically dissociated into single cells after 6–8 hrs in culture and typically 2.5×105 cells seeded per well in 6-well ultra low attachment plates (Corning). Mammospheres formed after 7 days were collected by centrifugation at 1000 rpm. For serial passaging, total number of mammospheres obtained at each passage was counted microscopically under a manually prepared ‘quadrant grid’, enzymatically dissociated into largely single cells, and seeded again in low attachment plates after live cell count. Sphere formation efficiency at each passage was calculated by dividing the total number of spheres formed by the total number of live cells seeded multiplied by hundred.
Staining of intact spheres was done, both in suspension and by fixation onto coated slides. Fixation was done with 1∶1 −20°C prechilled Methanol: Acetone for 10 mins. at room temperature (RT). Prior to permeabilsation, a short trypsin treatment was done to the fixed spheres for 3–5 mins., only for intracellular antigens. No such treatment was done for cell surface antigen like E-Cadherin. Permeabilisation was done in conjunction with blocking in 0.2% Fish Skin Gelatin using 0.5% Triton ×100 for 1 hr. This was followed by incubation with primary antibody for 2 hrs at RT. Primary antibodies were used at dilutions as specified by the manufacturer. Secondary antibody directly conjugated to a suitable flurophore was added and incubated for 45 min at room temperature in dark. Propidium Iodide (PI) was added as counter stain at a concentration of 0.5 ug/ml with the last PBS wash, incubated for 5 mins. at room temperature before mounting. DABCO was used as an antifade agent in the mounting media. Edges were sealed & the slides viewed using Zeiss 510 Meta confocal laser scanning microscope and analyzed using LSM Image Browser. For single cell staining, spheres were first trypsinised, fixed on slides, stained as above and viewed under fluorescence microscope (Leica). Antibodies against E Cadherin (BD Biosciences); ESA, Cytokeratins 14 and 18 (Sigma Aldrich); Cytokeratin 19 (Calbiochem) were used.
A Zeiss 510 Meta confocal laser scanning microscope was used to view the immunoflourescence staining. The 488, 543 or 660 nm laser lines were used for excitation of the fluorophores, while emissions were collected by specific band pass filters. Since the spheres are 3 dimensional specimens, optical sectioning was done along the XZ-plane to get a Z stack of the specimen.
Standardisation of the protocol for SP was done in HeLa cells and C57/black BL6 mouse bone marrow cells. After trypsinisation of cells, 5×105 cells was given a PBS wash and incubated with 1 ml of DMEM-F12 media and incubated at 37°C for 15 min. This was followed by addition of Hoechst 33342 at a final concentration of 2.5 µg/ml. To block Hoechst efflux, either Verapamil was added at a final concentration of 50 µM or Cyclosporin A was added at a concentration of 20 µM. The tubes were incubated in the dark at 37°C and 5% CO2 for 120 mins. with a constant and slow agitation. This was followed by two washes with PBS. For Rhodamine 123 staining, the dye was added upto a final conc. of 0.5 µg/ml and incubated at 37°C for 45 mins. This was followed by incubation in Rhodamine free medium for 30 mins. Cells were given a wash with cold PBS after staining and were analysed on the LSR-II machine (Becton Dickinson) . Hoechst 33342 was excited with a 355 nm UV laser. Emissions were collected in the blue and red channels using the 440/40 BP and 675 LP filters respectively. Immediately before analysis, Propidium Iodide (PI) was added at a final concentration of 2 µg/ml.
Mammosphere cells were incubated with anti CD24-FITC and anti CD44-PE antibodies (Becton Dickinson) in dark, on ice, for 45 mins, washed twice with cold PBS, resuspended in complete media and kept on ice till subsequent analysis on MoFlo (Dako). 50,000 cells from each quadrant were sorted in 1 ml complete media and seeded in 24 well Ultra low attachment plate (Corning). For dual staining of SP and CD markers, staining for SP was done at RT followed by staining for CD24 and 44 on ice as described above.
BrdU was added immediately after seeding cells to a final concentration of 10 µM. Cells were incubated with BrdU for a week at 37°C & 5% CO2 during which they formed spheres. These spheres were then dissociated into single cells and stained with Anti-BrdU antibody. Fixation was done using 70% ethanol. HCl treatment was done to remove the histones. Secondary antibody used was conjugated with FITC. The other steps of staining were the same as described earlier. PI was used as counterstain. Staining for BrdU was done for mammospheres at every passage. For each staining experiment, a negative control was included, in which primary antibody had not been added. Imaging of intact spheres stained for BrdU was done on the confocal laser scanning microscope & optical sectioning was done as described above to analyse the staining profile in all the layers of a mammosphere. Imaging of stained single cells was done in a regular fluorescence microscope (Leica). In order to quantify the proliferation status of cells within mammospheres of each passage, 10 random fields were captured during imaging. Total cell count and total positive cell count was taken over the random fields post imaging. A grand total of all the fields was obtained by adding up numbers from the random fields. This was done for 3 independent tissues at each passage and a graph was plotted subsequently to represent the proliferation status of mammosphere derived cells through multiple passages.
BrdU (10 µM) was added to 2.5×105 cells/ well of a 6-well ultra low attachment plate. At each passage (one week) a single well was harvested for BrdU staining, while the other wells were enzymatically dissociated and subjected to serial mammosphere assay without further addition of BrdU. Intact spheres from all the passages were subjected to BrdU staining and viewed by confocal microscopy.
Single cells obtained by enzymatic digestion of the spheres of each passage were given a wash with ice cold PBS followed by fixation with 0.2% glutaraldehyde at RT for 5 minutes. Post fixation, the cells were resuspended in buffer (pH 6.0) containing 1 mg/ml X-Gal and incubated in the dark at room temperature for 12 to 14 hours. The X-Gal was quenched by washing with ice cold PBS. The cells were viewed and counted under a regular phase contrast microscope.
Normal breast tissue was subjected to mechanical disruption followed by enzymatic treatment that resulted in the release of breast ‘organoids’ (
A: Representative phase contrast photomicrographs of organoids derived from primary human breast tissue; ×10 objective (a) and a primary mammosphere formed in suspension after 7 days; ×10 objective (b). B: Measurement of mammosphere size; ×10 objective (a) and cell size comprising the sphere (b) using LSM Image Browser from the Carl Zeiss website. C: Immunostaining of intact mammospheres viewed using a Zeiss 510 Meta confocal laser scanning microscope, and optical sectioning done along the XZ-plane to get a Z stack of the specimen. Photomicrographs represent negative control (minus primary antibody) (a), positive immunostaining for E-Cadherin (b) and Epithelial Specific Antigen (ESA) (c), but not for CD34 (d). RT-PCR analysis shows lack of nestin transcripts in mammospheres compared to U251 glioma cell line used as positive control (e). D: Photomicrographs represent immunostaining of single cells derived from primary mammospheres showing positivity for CK18 (a) CK14 (b) and CK19 (c). Insets show confocal images of intact spheres stained for the corresponding antigen. Hoechst 33342 and propidium iodide (PI) were used as nuclear counter stains for single cell and intact spheres, respectively (blue, Hoechst 33342; green, FITC; red, PI; scale bar = 20 µm). E: Gel picture shows hTERT expression in primary mammospheres (MS) by RT-PCR using HeLa as a positive control (a). Bar graph represents quantitative real time RT-PCR for hTERT expression in mammospheres compared to HeLa using primary human diploid fibroblasts (hF) as negative control (b). F: Mammospheres show few label retaining cells in T4 spheres. (a–d) represent photomicrographs of mammospheres immunostained for BrdU at each passage. (green, FITC signal; scale bar = 20 µm).
In order to understand the cellular architecture of mammospheres, we utilized confocal laser scanning microscopy to view and analyze cells in each plane through the entire sphere. Optical sections taken through propidium iodide (PI) stained spheres revealed that mammospheres are compact cellular structures (
Further analysis of the cellular composition of mammospheres by immunocytochemistry revealed that when compared with the negative control (
While most somatic cells are devoid of detectable telomerase activity, some adult stem cells exhibit telomerase activity
Adult stem cells have been identified in several systems by label retention assays wherein the slow dividing property of stem cells causes them to retain labels such as BrdU or tritiated thymidine, while the fast proliferating progenitors dilute them out with time
Many stem cells are endowed with the ability to efflux certain lipophilic drugs due to their cell surface expression of ABC family of membrane transporter proteins
When primary mammosphere-derived cells were subjected to SP analysis we observed a distinct side population comprising 0.5%–1% of the mammosphere cells, which disappeared on treatment with verapamil or cyclosporin A (
A: Side Population (SP) analysis of primary mammosphere-derived cells from two individual samples (a and b) reveal distinct SP. The top panels represent cells treated with Hoechst 33342 only (H), while the bottom panels represent cells treated with ABC transporter inhibitors verapamil (H+V) or cyclosporine (H+C). B: FACS profile of primary mammosphere-derived cells treated with Rhodamine 123 (a). In a simultaneous treatment of cells with Hoechst and Rhodamine, the SP (b) gated onto the Rhodamine profile (c) revealed that SP falls in the Rholow region. C: FACS profile of mammosphere-derived cells immunostained with anti CD24-FITC and anti CD44-PE antibodies (bottom) compared to unstained cells (top) (a). When 50,000 cells from different fractions were sorted and seeded in a 24-well ultra low attachment plate to assess sphere-forming ability (b), only the CD24lowCD44high fraction (4A and 4B) generated mammospheres. D: Mammosphere-derived cells, when simultaneously stained for CD24/44 and SP, revealed that SP falls in the CD24lowCD44low fraction. Dot plots represent unstained (a), CD24/44 profile (b), SP profile (c) and SP cells gated on to the CD24/44 profile (d).
When primary mammosphere derived cells were treated with Rhodamine123, we observed two distinct populations – Rhohigh and Rholow (
A major obstacle in breast stem cell biology is the absence of a specific marker(s) to identify this small population of cells. In some tissues, such as hematopoietic and neuronal tissues, normal and cancer stem cells have been shown to share the same cell surface markers
In order to understand the correlation between CD24/44 marker profile and SP in mammospheres, we undertook a simultaneous CD 24/44 and SP staining (
The ability of primary mammospheres/neurospheres to generate secondary spheres illustrates their self-renewal property
A: Bar graph represents the sphere forming efficiency (SFE) calculated by counting the number of mammospheres formed in a given well using a manually prepared grid and dividing this by the total number of cells seeded in the well, represented as percentage (n = 7). B: Bar graph represents cell death across different passages calculated by taking the difference in live cell count (Trypan blue exclusion) between each passage (n = 5). Statistical analysis for A and B was done using One Way ANOVA; ** signifies p<0.005 . Error bars represent standard deviation.
Since mammosphere formation involves proliferation, lack of proliferation could also result in the observed lack of mammosphere formation during serial passaging. Accordingly, we tested for the presence of proliferating cells through different passages using BrdU incorporation assay. Mammospheres from T1, T2, T3 and T4 passages contained on an average 3%, 6%, 13% and 10% proliferating cells, based on results obtained from five independent tissues (
A: Bar graph represents BrdU incorporation based proliferation analysis at every passage. Ratio of BrdU positive cell and total cell count was taken over random fields and represented as percentage. B: Graph represents comparison of proliferation potential and sphere forming efficiency from T1 to T4 (n = 5). C and D: Graphs represent differentiation status of cells within mammospheres from T1 to T4 passage as assessed by immunostaining for CK14 and CK18. (n = 4). Statistical analysis of this data was done using One Way ANOVA; * signifies p<0.05, ** signifies p<0.005. Error bars represent standard deviation.
Since only stem/progenitor cells, but not the differentiated cells, have the potential to generate mammospheres, an increase in the number of differentiated cells with serial passaging may also account for reduced mammosphere formation. To determine this, we analyzed the differentiation status based on expression of CK14 and 18 over different passages. The CK14 positive myoepithelial cell population increased from 7–30% from T1–T4, while the CK18 positive luminal epithelial cells population remained uniform, between 60–70%, across passages (
Primary cells in culture are known to undergo senescence after a few passages, which varies from one cell type to the other
A: Photomicrographs represent mammosphere-derived cells staining positive (blue color) for senescence associated β Galactosidase (SA β Gal) activity at different passages; ×20 objective. Trypsinized mammospheres were fixed with glutaraldehyde and incubated in X-Gal containing buffer (pH 6.0). B: Graph represents quantitative representation of the senescence profile of mammosphere-derived cells across passages, calculated by dividing the number of β Gal positive cells, as observed under the microscope, by the total number of cells and represented as percentage. (n = 6). Statistical analysis for B was done using One Way ANOVA. Error bars represent standard deviation.
The ability to grow in serum-free and anchorage-independent conditions as multicellular spheroids has been exploited for the isolation of stem/progenitor cells in several systems
Morphological characterization of mammospheres indicated that the spheres were of variable sizes ranging between 40 to 110 µm with an increase in smaller size spheres with passage. The difference in the size of the spheres may reflect the cell type of origin of mammospheres, with smaller spheres originating from progenitors, and larger spheres originating from stem cells. Also, the presence of cells of variable sizes within the mammospheres could represent the cellular heterogeneity found within the mammosphere; with smaller cells being the stem cells
The label retention experiment (
Based on our observations from serial sphere formation and label retention assays, differentiation and senescence analysis at each passage, we hypothesize that to begin with there is a small number of stem cells, many rapidly dividing progenitors along with a few differentiated cells within each mammosphere. With time, the progenitors differentiate, and the differentiated cells in turn either die or senesce. Some of the progenitors themselves might also senesce. The increasing senescent environment might negatively influence the sphere forming potential of the actual stem cells that may still be present in the senescent milieu at the end of a month, as shown by the label retaining cells.
The specific combination of CD24low and CD44high was found to mark breast cancer stem cells
Our results demonstrated a total lack of sphere formation after the fifth passage despite the presence of live (
Characterization of the senescence profile of mammospheres at every passage revealed a steady increase in the number of senescent cells from T1 to T4 (
Senescence is also triggered by non-telomeric signals such as, DNA damage, stress induced, culture conditions, or oncogenic stress
So, in the mammosphere culture system, the absence of telomerase in late passage spheres, and continuous exposure to high levels of oxygen, may result in widespread senescence. One possibility is that the stem/progenitor pool itself undergoes senescence. Another possibility is that the presence of a large number of senescent cells around the non-senescent stem/progenitor pool creates an unfavorable ‘niche’ for these primitive cells, thereby altering their self renewal and differentiation potential. The importance of the stem cell niche in regulating stem cell behavior is well recognized
Immunostaining and telomerase assay in mammospheres. A and B: Immunostaining of intact T1 mammospheres for CK 14 and 18 followed by optical sectioning of the stained spheres using confocal microscopy. C: TRAP assay done for the detection of functional telomerase in T2 and T4 mammospheres. When viewed on a 12% polyacrylamide gel, while one could observe a typical ladder pattern, which is representative of functionally active telomerase in T2 spheres (Lane 2), no such pattern was seen in T4 spheres (Lane 3). HEK 293T cells were used as positive control (Lane 1) and lysis buffer as negative control (Lane 4). SYBR green was used for detection. D: Detection of telomerase expression from T1 to T4 mammospheres by Real Time PCR. (HeLa cells were used as a positive control; human fibroblasts (hF) and lysis buffer as negative controls).
(1.16 MB DOC)
Single cell sorting of CD24-/lowCD44high cells and SP analysis of mouse bone marrow. Hoechst staining reveals a distinct side population in C56BL/6 mouse bone marrow cells used as positive control for SP staining of mammosphere derived cells. A: Staining of mouse bone marrow cells with Hoechst 33342 shows presence of SP in region R2. B: The disappearance of the side population (R2) in the presence of the transporter blocker Verapamil. C: Single CD24-/lowCD44high sorted single cell derived spheres formed in a 96-well ultra low attachment plate (a, b); Cells derived from the single mammospheres stained for the differentiation markers, CK 14 ( c ) and CK 18 (d).
(1.80 MB DOC)
Analysis of CD24 and CD44 expression. A: Quantitative analysis of expression of the surface markers CD24 and CD44 in three tissues. B: Analysis of expression of CD24 and CD44 by cells sorted from each depicted gates (refer to
(1.32 MB DOC)
Differentiation potential of mammosphere derived cells. A: Matrigel based 3D differentiation assay, results in formation of spherical (a) and tubular (d) structures within the gel after 15 days in culture. Optical section through the centre of one of the matrigel derived spheres, revealing majority of acinar structures (b and c). These structures stain positive for the differentiation markers, CK 14 (myoepithelial cells) and CK 18 (luminal epithelial cells) (e–f) (blue: Hoechst; green: FITC; scale bar for a and d is 25 µm; for b, c, e and f scale bar is 20 µm). B: Differentiation assay carried out in serum reveals CK 14 positive cells (a), CK 18 positive cells (b) and dual positive cells (c) Arrow indicates cells which are positive for CK 14 but not CK 18 (red: TRITC, green: FITC. Blue: Hoechst; scale bar represents 25 µm)
(2.91 MB TIF)
We would like to thank D Alaguraj, Dr. Omana Joy (IISc) and Dr. H Krishnamurthy (NCBS, TIFR Centre) for help with flow cytometry, and Naseer M for help with SP experiments. We also acknowledge the confocal, FACS, and IRIS Imaging facility of IISc.