Time-lapse imaging is used to assess behavior of primary preneoplastic mammary epithelial cells derived from genetically engineered mouse models of breast cancer risk to determine if there are correlations between specific behavioral parameters and distinct genetic lesions.
Time-lapse imaging can be used to compare behavior of cultured primary preneoplastic mammary epithelial cells derived from different genetically engineered mouse models of breast cancer. For example, time between cell divisions (cell lifetimes), apoptotic cell numbers, evolution of morphological changes, and mechanism of colony formation can be quantified and compared in cells carrying specific genetic lesions. Primary mammary epithelial cell cultures are generated from mammary glands without palpable tumor. Glands are carefully resected with clear separation from adjacent muscle, lymph nodes are removed, and single-cell suspensions of enriched mammary epithelial cells are generated by mincing mammary tissue followed by enzymatic dissociation and filtration. Single-cell suspensions are plated and placed directly under a microscope within an incubator chamber for live-cell imaging. Sixteen 650 μm x 700 μm fields in a 4×4 configuration from each well of a 6-well plate are imaged every 15 min for 5 days. Time-lapse images are examined directly to measure cellular behaviors that can include mechanism and frequency of cell colony formation within the first 24 hr of plating the cells (aggregation versus cell proliferation), incidence of apoptosis, and phasing of morphological changes. Single-cell tracking is used to generate cell fate maps for measurement of individual cell lifetimes and investigation of cell division patterns. Quantitative data are statistically analyzed to assess for significant differences in behavior correlated with specific genetic lesions.
Genetically engineered mouse models are tools to study and understand how different genetic lesions contribute to the risk of developing breast cancer. For example, genetically engineered mice have shown that the combination of three factors: loss of the full-length breast cancer 1, early onset (Brca1) gene in mammary epithelial cells, tumor protein p53 (Tp53) germ-line haploinsufficiency, and mammary epithelial cell targeted up-regulated Estrogen receptor alpha (ERa) expression results in the development of mammary cancer in 100% of Brca1floxed (f)11/f11/Mouse Mammary Tumor Virus (MMTV)-Cre/p53+/-/tetracycline-operator (tet-op)-ER/MMTV-reverse tetracycline transactivator (rtTA mice by 12 months of age in comparison to the lower percentages reported in Brca1f11/f11/MMTV-Cre/p53+/– mice without ERa over-expression (~50-60%) and Brca1f11/f11/MMTV-Cre mice without Tp53 haploinsufficiency (<5%).1
Dynamic time-lapse imaging of the behavior of preneoplastic primary mammary epithelial cells reveals differences in cell behavior that are less easily appreciated in static tissue sections. Alterations in proliferation and differentiation are observed in primary mammary cells from human BRCA1 mutation carriers.2 Creation of single-cell suspensions of primary mammary epithelial cells from normal and genetically engineered mice are generated through enzymatic disassociation of resected mammary gland tissue.3 Time-lapse images are viewed to assess mechanism and timing of cell colony appearance and incidence of morphological changes in cells including epithelial-mesenchymal transition (EMT) and apoptosis. Generation of cell fate maps, quantification of the length of time between cell divisions (cell lifetimes), and determination of patterns of cell division are facilitated by use of single-cell tracking. Timm’s Tracking Tool (TTT) is publically available software used to generate single-cell fate maps. Its utility in elucidating mechanisms of cell fate has been established4,5 examining normal hematopoietic stem cell development6-9 and the generation of neurons.10
1. Overall Scheme
2. Generation of Primary Mammary Epithelial Cell Cultures
3. Live-cell Imaging
4. Direct Viewing of Time-lapse Images to Assess Timing and Mechanism of Initial Epithelial Cell Colony Formation, Incidence of Apoptosis, and Phasing of Morphological Changes
5. Image File Modification
6. Generation of Cell Fate Maps of Single Cells Using TTT
7. Statistical Analyses
Epithelial and fibroblast cells can be distinguished by cell morphology. Epithelial cells have a cuboidal shape (Figure 1A-B) and form cell colonies (Figure 1A). Fibroblasts, a type of stromal cell, have an elongated morphology (Figure 1C).
Cells were rounded and floating at the onset of imaging (Figure 2A-D). After attachment to the plate they became flat and demonstrated a cuboidal-type appearance (Figure 2E, H). By 4 days of culture the majority of epithelial cells elongated into an EMT-like morphology (Figure 2I-L). This change occurred with the same chronology in all genotypes studied. Some digital images were blurry because Kohler illumination and phase ring alignment was not set correctly (Figure 2D, H, L).
The floating cells developed into defined individual epithelial cell colonies by 24 hr after plating (Figure 3A, B). Serial image analysis revealed that these colonies were generated by cell aggregation rather than cell division. While disruption of Brca1 by itself did not alter the number of colonies formed, addition of Tp53 haploinsufficiency significantly reduced the number of colonies formed and addition of ERα over-expression significantly increased the number of colonies formed (Figure 3C).
It is critical to monitor the digital images acquired and adjust the focus map frequently during the first 24 hr as cells attach to the plate and then at least twice daily to ensure the cells remain in focus and cultures are not contaminated. Accuracy of analyses is compromised if imaged cells are not in focus (Figure 4).
In the representative experiment presented here, a critical question was to determine if changes in expression levels of genes known to impact mammary cancer risk (Brca1, Tp53,and estrogen receptor 1 (Esr1)1) altered the number of hours between cell divisions (cell lifetime) or impacted the patterns of division (symmetric versus asymmetric) in non-cancerous primary mammary epithelial cells. To accomplish this, individual cell fate maps (trees) were generated using TTT. Examples of representative cell fate maps for each of the four genotypes tested are illustrated (Figure 5A-D). The four genotypes of the cells tested were wild-type (no genetic manipulation) (Figure 5A), loss of full-length Brca1 (Brca1f11/f11/MMTV-Cre ) (Figure 5B), loss of full-length Brca1 in combination with Tp53 haploinsufficiency (Brca1f11/f11/MMTV-Cre/p53+/-) (Figure 5C), and loss of full-length Brca1 in combination with Tp53 haploinsufficiency and gain of Esr1 (Brca1f11/f11/MMTV-Cre/p53+/-/MMTV-rtTA/tet-op-ER) (Figure 5D). Generation of cell fate maps was stopped when cells became confluent (~3-4 generations) since unequivocal tracking of single cells was no longer possible after that. Generation 0 (time to first cell division) was not included in the analyses of cell lifetimes because duration of generation 0 was longer and more variable than subsequent generations. Generations 1, 2, 3, and 4 were grouped together for the final analyses because there were no significant differences in mean cell lifetimes or Standard Error of the Mean (SEM) between these generations. Comparisons of mean cell lifetimes between the different genotypes (Figure 5E) revealed that loss of full-length Brca1 reduced the number of hours between cell divisions from 21.3 ± 1.6 hr (wild-type) to 16.5 ± 1.0 hr (Brca1f11/f11/MMTV-Cre ). Notably, although loss of one Tp53 allele in addition to loss of full-length Brca1 is associated with a significant increase in mammary cancer development1, mean cell lifetime was unchanged (16.3 ± 1.2 hr) compared to mice lacking Brca1 only. Interestingly gain of Esr1 de Brca1f11/f11/MMTV-Cre/p53+/-/MMTV-rtTA/tet-op-ER mice restored cell lifetime durations to near wild-type levels (19.3 ± 2.1 hr), albeit these mice have the highest incidence of mammary cancer development of all the models studied here.1 These findings indicated that loss of full-length Brca1 was not invariably associated with a shortened cell lifetime, instead this was modified by over-expressed ERα. Possible next-step experiments utilizing this technology could be dose-response experiments using different estrogens, other growth factors, or candidate therapeutics to measure their impact on cell lifetime in the different genetic backgrounds. In contrast to the impact on cell lifetime duration, none of the genetic alterations studied here altered cell division patterns.
Figure 1. Appearance of an epithelial cell colony (A), epithelial cell (B), and fibroblast (C) from time-lapse imaging. Epithelial cells appear more cuboidal while fibroblasts appear more elongated. Size bars=50 μm.
Figure 2. Representative serial time-lapse images from time 0, 2 and 4 days demonstrate changes in cellular morphology over time and differences in appearance with and without Kohler illumination. At time 0 the plated cells are floating (arrowheads A-D), by two days in culture they are adherent and may exhibit a cuboidal-type morphology (thick arrows E, H) and by four days in culture they elongate and demonstrate an EMT-like morphology (thin arrows I-L). Kohler illumination is present in panels A-C, E-G, and I-K. Panels D, H and L illustrate images without Kohler illumination. Images shown are from the same imaging point over time. Size bars=200 μm. Click here to view larger figure.
Figure 3. Numbers of epithelial cell colonies at 1 day varied by genotype. (A) Rounded floating cells at the beginning of time-lapse imaging. (B) Two representative epithelial cell colonies (arrows). (C) Numbers of epithelial cell colonies/frame formed at 24 hr varied by genotype. Bar graphs illustrate the mean and standard error of the mean. *p<0.05, ANOVA. Size bars=200 μm. Cells isolated from mammary glands without palpable tumor from 10- to 12-month-old Brca1f11/f11/MMTV-Cre (n=4), Brca1f11/f11/ MMTV-Cre/ p53+/- (n=3), Brca1f11/f11/p53+/-/MMTV-Cre/Tet-op/-ER/MMTV-rtTA (n=3), and wild-type (n=3) mice were imaged for the studies. Five independent imaging experiments composed of different combinations of wild-type and genetically engineered mice were performed. Media contained phenol red. No estrogen was added. Click here to view larger figure.
Figure 4. Out-of-focus images cannot be accurately analyzed. Before imaging the plate, the plate should sit in the imaging incubator for 15 min to equilibrate. During the first day of imaging, the focus should be checked and adjusted every few hours. After that it should be checked twice daily but generally remains more stable. Arrow indicates an out-of-focus epithelial cell colony.
Figure 5. Cell fate maps generated from single-cell tracking using TTT for (A) wild-type, (B) Brca1f11/f11/MMTV-Cre, (C) Brca1f11/f11/MMTV-Cre/p53+/-, and (D) Brca1f11/f11/MMTV-Cre/p53+/-/MMTV-rtTA/tet-op-ER mice. Cells that cannot be tracked due to loss out of frame or into a confluent cell layer are indicated by a question mark. When no question mark indicated, tracking was deliberately ended due to cell confluence and inability to unequivocally follow single-cell fate. (E) Mean cell lifetimes varied by genotype. Bar graphs illustrate the mean and standard error of the mean cell lifetimes (time between cell divisions) over generations 1-4 for each genotype. Mean cell lifetimes were significantly shorter in mammary epithelial cells derived from Brca1f11/f11/MMTV-Cre and Brca1f11/f11/MMTV-Cre/p53+/- as compared to wild-type mice. *p<0.05, ANOVA. Cells isolated from mammary glands without palpable tumor from 10- to 12-month-old Brca1f11/f11/MMTV-Cre (n=4), Brca1f11/f11/ MMTV-Cre/ p53+/- (n=3), Brca1f11/f11/p53+/-/MMTV-Cre/Tet-op/-ER/MMTV-rtTA (n=3), and wild-type (n=3) mice were imaged for the studies. Five independent imaging experiments composed of different combinations of wild-type and genetically engineered mice were performed. Media contained phenol red. No estrogen was added. Click here to view larger figure.
Critical steps
It is important to ensure that mammary glands are harvested from mice of the same age to control for age-related variability in mammary epithelial cell behavior. When plating the cells, the same number of cells should be plated in each well for every experiment. Cells should be relatively sparse when plating so that cultures do not become confluent too quickly making it possible to follow multiple individual cells through serial images. It is essential that the plate be equilibrated in the incubator before imaging because the focus will change after the plate is equilibrated. Once imaging has begun, the focus should be checked frequently and adjusted as needed, especially within the first 24 hr. It is important to have the temperature of the incubator regulated and pre-warmed. The number of people going in and out of the room should be limited as much as possible. We recommend practicing all aspects of setting up the experiment beforehand to ensure reproducibility and quality control. As with all cell culture experiments, sterility is an important issue and standard sterile cell culture practices should be used at all times.
Saving and manipulating the large image files requires a substantial amount of memory space. A shared network drive or portable hard drives can be used. Once files are converted it is important to make sure the images are consistently named correctly and placed in corresponding folders. Frequent saves of digital data should be performed throughout the process.
Limitations
This method uses a 2-D culture system that does not allow for analysis of behavioral differences that might become apparent in a 3-D culture system. Durations of cell lifetimes may only be reproducibly measured after the first observed cell division as the number of hours to first cell division following initial plating of the cells is variable.
Possible modifications
Depending on the requirements of the experiment, time-lapse images can be taken more or less frequently. For example, if transient cellular structures are to be quantified, a 5 sec interval may be more appropriate. Other procedures to generate primary mammary epithelial cell cultures may be used. Any imaging equipment may be used to create the time-lapse imaging, as long as it is able to meet the requirements of the experiment. Alternative systems for single-cell tracking can be employed. For the procedure described here, a standard plastic flat bottom 6-well plate was adequate. Plastic dishes can be used for all long working distance air lenses such as 4x, 10x, and 20x. It is important to choose regions near the center of the plastic dishes when performing phase contrast imaging since the plastic is curved near the edges in most dishes and distorts the light. Coverslip thickness glass bottomed wells would be required for normal working distance immersion lenses (oil, water, glycerol, etc.), which are typically at higher magnification.
Trouble-shooting
When cells are plated, media should be sufficient to maintain cell growth for 5 days to decrease the need for manipulating the plate on the stage; however, if necessary, media can be added. It is recommended that evaporation be kept to a minimum. Evaporation should be managed by placing three to four dishes of sterile deionized water inside the incubator and refilling them when necessary. It is possible to fill the unused wells of the 6-well plate with sterile deionized water.
If images are not loading into the TTT program, first check and make sure the files and folders are arranged and named correctly. Next, check if the log file is in the correct folder (see step 5.5).
Future application or direction after mastering technique
The same general procedure could be used to analyze behavior of other cell types including human primary mammary epithelial cells as well as breast cancer cells. For example, genotypic-specific behavior and/or response to candidate treatments can be analyzed. Alternatively, imaging of fluorescence-activated cell (FAC) sorted cell populations could be performed or fluorescent labels added to monitor specific cell types.
Significance of this technique with respect to existing methods
Addition of time-lapse imaging and analyses of cell behavior over time to traditional cell culture and labeling techniques provides quantitative data from single cells over time rather than only whole population dynamics. This technique allows single cells to be observed continuously, unlike traditional methods that permit observation only at discrete timepoints. Moreover, traditional methods require the cells to be disturbed by taking them in and out of the incubator to view them under a microscope whereas the approach described here allows the cells to be observed without being transferred. Finally, time-lapse imaging provides an enduring record of the cells under observation that can be returned to for further analyses. The use of TTT to track individual cell behavior permits an unequivocal analysis of multiple parameters and does not require the use of a fluorescent label to localize the specific cells under observation. The technique reveals differences in mammary epithelial cell behavior that are difficult to measure in static tissue sections of mammary gland. Cell lifetime measurements quantify the number of hours between mitotic events. This measurement provides specific data on the actual number of hours between cell divisions indicating the length of the cell cycle in hours. A researcher can use this data to determine how specific genetic modifications or culture conditions impact cell cycle length.
The authors have nothing to disclose.
The authors wish to thank Bofan Wu and Christian Raithel for technical assistance and Michael Rieger for his introduction to live-cell imaging. Supported by NCI, NIH RO1CA112176 (P.A.F.), NCI, NIH. R01CA89041-10S1 (P.A.F.), Deutscher Akademischer Austaush Dienst e.V. A/09/72227 Ref. 316 (R.E.N.), Department of Defense W81XWH-11-1-0074 (R.E.N.), WCU (World Class University) program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (R31-10069) (P.A.F.), NIH IG20 RR025828-01 (Rodent Barrier Facility Equipment), and NIH NCI 5P30CA051008 (Microscopy and Imaging and Animal Shared Resources).
Name of Reagent | Company | Catalog Number | Comments |
EpiCult-B Basal Medium Mouse | StemCell Technologies | 05610 | |
EpiCult-B Proliferation Supplements Mouse | StemCell Technologies | 05612 | |
recombinant human Epidermal Growth Factor (rhEGF) | StemCell Technologies | 02633 | |
Collagenase/Hyaluronidase | StemCell Technologies | 07912 | |
Disposable Scapels | Feather | 2975#10 | |
Hanks’ Balanced Salt Solution | StemCell Technologies | 37150 | |
Ammonium Chloride | StemCell Technologies | 07800 | |
Tryspin-EDTA | StemCell Technologies | 07901 | |
Dispase | StemCell Technologies | 07913 | |
DNase I | StemCell Technologies | 07900 | |
40 μm cell strainer | StemCell Technologies | 27305 | |
FBS | StemCell Technologies | 06100 | |
PenStrep | Gibco | 15140 |