This protocol efficiently studies mammalian cell division in 3D collagen matrices by integrating synchronization of cell division, monitoring of division events in 3D matrices using live-cell imaging technique, time-resolved confocal reflection microscopy and quantitative imaging analysis.
The study of how mammalian cell division is regulated in a 3D environment remains largely unexplored despite its physiological relevance and therapeutic significance. Possible reasons for the lack of exploration are the experimental limitations and technical challenges that render the study of cell division in 3D culture inefficient. Here, we describe an imaging-based method to efficiently study mammalian cell division and cell-matrix interactions in 3D collagen matrices. Cells labeled with fluorescent H2B are synchronized using the combination of thymidine blocking and nocodazole treatment, followed by a mechanical shake-off technique. Synchronized cells are then embedded into a 3D collagen matrix. Cell division is monitored using live-cell microscopy. The deformation of collagen fibers during and after cell division, which is an indicator of cell-matrix interaction, can be monitored and quantified using quantitative confocal reflection microscopy. The method provides an efficient and general approach to study mammalian cell division and cell-matrix interactions in a physiologically relevant 3D environment. This approach not only provides novel insights into the molecular basis of the development of normal tissue and diseases, but also allows for the design of novel diagnostic and therapeutic approaches.
Cell mitosis is a critical event in cellular life, the regulation of which plays crucial roles in tissue and organ development. Abnormal mitosis is implicated in natural genetic variations, human aging processes, and the progression of cancer1,2,3,4,5. The increased rate of proliferation of tumor cells compared with normal cells is one of the hallmarks of cancer, despite the fact that cell behaviors are quite heterogeneous among different types of tumors and even among patients. In spite of promising preclinical results, some newly-developed antimitotic drugs have not shown to be effective in clinical trials6,7,8,9,10,11.The relevance of experimental and preclinical models has to be considered. Many types of normal mammalian and cancer cells divide in three-dimensional (3D) matrices, such as fibroblasts and fibrosarcoma cells in collagen I-rich 3D connective tissues, and metastatic cancer cells in the 3D stromal extracellular matrix (ECM). However, the vast majority of mammalian cell division experiments and assays have been performed on cells cultured on two-dimensional (2D) substrates. An engineered 3D matrix could better recapitulate the microstructure, mechanical properties, and biochemical signals of the 3D ECM of both normal and pathologic tissues12,13,14,15,16,17.
The study of how mammalian cell division is regulated in 3D environments remains largely unexplored despite both the physiological relevance and the therapeutic significance18,19. Possible reasons include the technical difficulties and experimental challenges associated with studying cell division in 3D matrices. Cell mitosis constitutes a small temporal fraction in the whole cell cycle20. Previous work has shown that the proliferation rate of many mammalian cells, such as human breast adenocarcinoma MCF-7, human osteosarcoma U2OS, and human liver HepG2, is much lower in 3D matrices compared with their counterparts on 2D substrates21,22. Furthermore, cells embedded in 3D matrices move in and out of focus during live-cell imaging. All of these factors contribute to the extremely low efficiency of capturing cell-division events in 3D culture using imaging techniques.
Interactions between the ECM and cells play critical roles in regulating cell divisions. Here, we describe an approach to efficiently study mammalian cell division in 3D collagen matrices. The method includes the incorporation of mitotic markers to the cells, synchronization of cell division, as well as the monitoring of division events in 3D matrices using the live-cell imaging technique, time-resolved confocal reflection microscopy, and quantitative imaging analysis. Fluorescence-labeled histone protein H2B is first introduced into the cells as a marker to differentiate mitotic and interphase cells. Then the cells are synchronized using the combination of thymidine blocking and nocodazole treatment, followed by a mechanical shake-off technique. Synchronized cells are then directly encapsulated into 3D collagen matrices. Cell division events of multiple cells are monitored efficiently using low-magnification time-lapse live-cell imaging. The deformation of collagen fibers, which is an indicator of cell-matrix interaction, is monitored using confocal reflection microscopy at high-magnification.
We have previously used this technique to monitor and quantify cell-matrix interaction before, during and after the mitosis of two metastatic cancer cell lines, human invasive ductal carcinoma MDA-MB-231 and human fibrosarcoma HT1080 cells, in 3D collagen matrices19. The methods presented here provide an efficient and general approach to study both mammalian cell division in a 3D environment and cell-matrix interactions. The MDA-MB-231 cell line is used as an example throughout the paper. This protocol provides novel insights into the molecular basis of the development of normal tissue and diseases, and could also allow for the design of novel diagnostic and therapeutic approaches.
The protocol provided follows the guidelines of The Homewood Institutional Review Board (HIRB).
1. Stable Expression of H2B-mCherry as a Marker for Cell Mitosis
2. Synchronization of the Cells Stably Expressing H2B-mCherry
3. Incorporation of the Synchronized Cells into the Collagen I Matrices
NOTE: Type I collagen is the most abundant protein in the human body and in the ECM of connective tissues, and thus is widely used to investigate how eukaryotic cell functions are modulated by a 3D environment17,23,24. Collagen is soluble in acetic acid. After neutralizing and warming the collagen solution to 20 – 37 °C, collagen monomers polymerize into a meshwork of collagen fibrils.
4. Live Cell Imaging of the Cells Dividing in the 3D Collagen Matrices (Low Magnification)
NOTE: Images of cells are collected at 2 min intervals using a charge coupled device (CCD) camera mounted on a phase contrast microscope that is equipped with a 10X objective and controlled by imaging software.
5. Collagen Network Deformation During Cell Division (High Magnification Microscopy)
The goal of this article is to present an imaging-based method to study mammalian cell division processes in 3D matrices, and to quantify the interactions between the cell and the 3D extracellular matrix during and after cell division. To facilitate the imaging of cell mitosis, we incorporated H2B-mCherry into MDA-MB-231 cells using lentiviral transduction. H2B conjugated with fluorescent proteins is used as a mitotic marker to distinguish mitotic cells from interphase cells, and to define different stages during cell mitosis19,20,26. Using this method, we were able to monitor the entire division process of MDA-MB-231 cells stably expressing H2B-mCherry in 3D collagen matrices (Figure 1A). Mitotic phases started with the dissolution of the nuclear membrane (prophase; frame 1, Figure 1A); the re-organization of the chromosomes (prometaphase: frame 2); the alignment of the chromosomes in the middle of the cell body (metaphase; frame 3); the separation of the chromosomes (anaphase; frame 4); the reorganization of the chromosomes and nuclear membrane, and the separation of the bodies of the two daughter cells (telophase/cytokinesis, frame 5).
The proliferation rate of cells in a 3D matrix is usually much lower than their counterparts on a 2D substrate, which renders the monitoring of cell division in 3D less efficient19. To increase the efficiency of studying 3D cell division, we employed a method combining thymidine, nocodazole, and a shake-off technique to synchronize and select MDA-MB-231 cells that are at the mitotic phase. Synchronized cells were then embedded into collagen matrices. The division process of multiple cells was monitored in real-time using live-cell imaging at 10X magnification. About 70% of the cells divided within the first 2 h after the formation of collagen matrix (Figure 2), thus allowing for efficient monitoring of mitosis in 3D (Figure 2).
To monitor the interaction between cells and their surrounding collagen matrices, we combined confocal reflection microscopy to image collagen fibers, and fluorescence microscopy to image the cells. A 60X lens allows for the capture of high-quality images of collagen fibers. Imaging using a high-magnification lens, which captures fewer cells in each field of view, is much lower-throughput compared to the use of low-magnification at 10X or 20X. Successful synchronization of the cells greatly enhances the efficiency and throughput of such an experiment, since most of the synchronized cells divide within the first 2 h after the formation of the collagen matrix. The matrix deformation during and after cell division is visualized (representative snapshots are shown in Figure 1B) and quantified using a custom PIV software. We quantified and compared matrix deformation during interphase and mitotic phase for cells synchronized into the mitotic phase. We observed that matrix deformation changed very little during the mitotic phase (Figure 3A), and is less than the deformation observed in post-mitotic phases (Figure 3B). This result shows that mammalian cells have minimal attachment and interactions with the surrounding matrix while they are in the mitotic phase.
Figure 1: Representative micrographs obtained from a high-magnification live-cell imaging video of a MDA-MB-231 cell embedded in a collagen matrix that stably expresses H2B-mCherry. (A) Different phases of the mitotic progression of the MDA-MB-231 cell are defined by H2B-mCherry as labeled in red. (B) The collagen fibers (white) during the mitotic process are visualized by time-dependent confocal reflection microscopy. Scale bar = 20 µm. Please click here to view a larger version of this figure.
Figure 2: Division of the synchronized cells in 3D collagen matrices. Cells were synchronized to G2/M phase and embedded in a collagen matrix. About 70% of synchronized cells divide within the first 2 h, whereas control cells without synchronization divide randomly. Error bar = SEM (standard error of the mean). Please click here to view a larger version of this figure.
Figure 3: Quantification of the matrix deformation for MDA-MB-231 cells during interphase and mitosis. (A) The change in the magnitude of matrix deformation for matrix-embedded MDA-MB-231 cells. The green arrow indicates prophase and red arrow indicates telophase. (B) Quantification of matrix deformation for MDA-MB-231 cells during interphase and mitotic phase, indicating that matrix deformation is minimal during cell division. Error bar = SEM (standard error of the mean). * p <0.05. Please click here to view a larger version of this figure.
The previous study of cell division in 3D was not efficient due to experimental limitations and technical challenges18,19. The critical steps for efficient study of mammalian cell division in 3D collagen matrices are: (1) the incorporation of fluorescence-labeled mitotic markers to the cells; (2) the synchronization of cell division; and (3) the monitoring of division events in 3D matrices using live-cell imaging technique, time-resolved confocal reflection microscopy, and quantitative imaging analysis.
Mitotic cells on 2D substrates can be distinguished from the interphase cells based on their morphology, i.e. mitotic cells are round and barely attach to the substrate whereas interphase cells spread out and attach firmly to the substrate. In 3D matrices, however, the cell morphology is not a reliable marker for mitotic cells since some cells barely spread out and remain round in the matrix27,28. Thus, it is essential to introduce a mitotic marker to the cells for the study of cell division in 3D matrices. We stably expressed H2B-mCherry in MDA-MB-231 cells, which served as a reliable marker for different mitotic phases including prophase, prometaphase, metaphase, anaphase, and telophase/cytokinesis. We previously used this approach to distinguish mitotic cells from interphase cells in 3D matrices. With the help of this marker, we were also able to measure the length of the mitotic phase for another cell line, HT1080 cells, dividing on 2D substrates and in 3D matrices19.
There are multiple ways to synchronize cells, including serum starvation29, mitotic shake-off30,31, double thymidine blocking32, and nocodazole32. We combined the thymidine treatment and the nocodazole treatment to efficiently synchronize MDA-MB-231 cells to the G2/M phase. Cells exposed to thymidine are arrested at the G1/S transition and throughout S phase due to the inhibition of DNA synthesis by thymidine. The release of the cells from the thymidine exposure let the cells progress to G2/M phase for cells arrested at G1/S phase, and to G1 for cells arrested at S phase. All the cells exposed to nocodazole are arrested at the G2/M phase. The rounded-up cells entering mitotic phase were then shaken-off from the plate and directly encapsulated into the collagen matrices. We showed that about 70% of the cells divide within 2 h after they are incorporated into the collagen matrices (Figure 2). Alternatively, cells could be embedded in collagen matrices before synchronization, however, the cross-linked network of the collagen matrix presents a physical barrier and reduces the rate of biomolecular diffusion and convection33,34. Indeed, we attempted to synchronize cells in collagen matrices using thymidine and nocodazole, but failed to obtain efficient synchronization. This result might be due to the inefficient diffusion and convection of the drugs through the collagen matrix.
Reflection is an intrinsic optical property of many biopolymers, including collagen. The confocal reflection microscopy technique visualizes and quantitates the microtopography of porous biomaterials prepared from synthetic polymers and 3D collagen matrices23,24,35,36,37. In our lab, we have established techniques to monitor changes in collagen fiber polarity as the concentration of collagen varies38. Here, we describe the method to monitor the deformation of collagen fibers based on the time-lapse video of the reflective confocal images to denote the cell-matrix interaction. The representative results presented here show that the matrix deformation for the mitotic MDA-MB-231 cells is significantly smaller than those cells in interphase, which suggests that mammalian cells have minimal attachment and interactions with the surrounding matrix when they enter the mitotic phase19.
Previously, we used confocal reflection microscopy to monitor and quantify cell-matrix interaction before, during and after the mitosis of HT1080 cells. We also monitored the matrix deformation by β1-integrin knock-down HT1080 cells during both interphase and mitotic phase. Depleting β1-integrin significantly reduces the matrix deformation by the cell during interphase. However, there is no difference in matrix deformation during the mitotic phase of the round β1-integrin knockdown (KD) cells and the HT1080 wild type cells19.
An alternative approach to visualize collagen fibers is to employ fluorescence-conjugated type I collagen. We previously used this approach to image tracks in the collagen matrices generated by cells19. This approach, however, requires the labeling of collagen with fluorescent dye such as fluorescein isothiocyanate (FITC), which is both time consuming and less efficient. On the other hand, confocal reflection microscopy can be directly applied to unmodified collagen to save time and resources, and exclude the issues associated with fluorescence photobleaching. Moreover, this method does not require an individual fluorescent channel, and is therefore compatible with all fluorescent dyes.
The method presented in this paper can be potentially applied to any type of mammalian cells that divide in a 3D collagen matrix. The incorporation of the H2B-mCherry marker into other mammalian cells by lentiviral transduction will follow exactly the same procedures as described in the paper, although different types of cells may have varied efficiencies for transduction by lentivirus39. Both the density of the cells upon transduction and the titer of the virus could be optimized for efficiency. If high transduction efficiency cannot be achieved, cells could be selected by fluorescence assisted cell sorting (FACS). Thymidine blocking and nocodazole are applied to successfully synchronize several other types of mammalian cells, such as HeLa40. Mechanical shake-off could be applied to any cell types that round up and barely attach to the substrate during the mitotic phase30,31. Moreover, the imaging of collagen fibers using confocal reflection microscopy, and the quantification of the matrix deformation can be directly applied to all other types of dividing mammalian cells.
The method presented here is an efficient and general approach to study mammalian cell division and cell-matrix interactions in a 3D environment. The approach facilitates our probe into the molecular basis of the development of normal tissue and diseases, and potentiates the design of novel diagnostic and therapeutic approaches in the future.
The authors have nothing to disclose.
This work was supported by NIH grants R01CA174388 and U54CA143868. The authors would like to acknowledge the PURA award from the Johns Hopkins University for support of Wei-tong Chen. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. 1232825.
Human embryonic kidney 293T | ATCC | ||
MDA-MB-231 | Physical Sciences Oncology Center, NIH | ||
DMEM | Corning | 10-013-CV | |
DMEM powder | ThermoFisher Scientific | 12100-046 | |
Fetal bovine serum | Hyclone | SH30910.03 | |
Penicillin-Streptomycin 100X | Sigma-Aldrich | P0781 | |
Fugene HD | Promega | E2311 | |
Lipofectamine 2000 | Life technologies | 11668-07 | |
Plasmid encoding H2B-mCherry in a lentiviral vector | Addgene | plasmid 21217 | |
Thymidine | Sigma-Aldrich | T1895 | |
Nocodazole | Sigma-Aldrich | M1404 | |
Opti-MEM | Life Technologies | 31985-070 | |
Sodium bicarbonate | GibcoBRL | 11810-025 | |
HEPES | Sigma-Aldrich | 113375-100 | |
Collagen | Corning | 354236 | |
NaOH | J.T. Bake | 3722-01 | |
Millex-HV syringe filter unit, 0.45-μm, PVDF, 33 mm | Millipore | SLHVM33RS | |
Nikon TE2000E epifluorescence microscope | Nikon | TE2000E | |
Cascade 1K CCD camera | Roper Scientific | ||
NIS-Elements AR imaging software | Nikon | ||
Nikon A1 confocal microscope | Nikon | A1 |