|Home | About | Journals | Submit | Contact Us | Français|
Proteolytic degradation of extracellular matrix (ECM) components by cells is an important metabolic activity as cells grow, remodel, and migrate through the ECM. The ability to analyze ECM degradation can be valuable in the study of developmental processes as well as pathologies such as cancer. In this protocol we describe an in vitro live cell–based method to image and quantitatively measure the degradation of ECM components by live cells. Cells are grown in the presence of fluorescent dye-quenched protein substrates (DQ-gelatin, DQ-collagen I, and DQ-collagen IV) that are mixed with protein matrices. Upon proteolytic cleavage, fluorescence is released that directly reflects the level of proteolysis by the cells. Using confocal microscopy and advanced imaging software, the fluorescence is detected and accurate measurements of proteolytic degradation in three and four dimensions can be assessed.
In vivo, cellular microenvironments are constructed of complex combinations of extracellular matrix (ECM) proteins that are under constant turnover and modification. During most forms of cellular motility, cells remodel their immediate microenvironment via proteolytic degradation of ECM (Friedl and Wolf, 2003). Thus, the ability to examine the proteolytic degradation of ECM by living cells is crucial for understanding many aspects of cellular motility in normal processes and in pathologies such as cancer. The assay described here allows one to quantitatively analyze extracellular and intracellular proteolysis associated with degradation of ECM surrounding the cells. This is a microscopy-based assay in which proteolysis of fluorogenic dye-quenched protein substrates (DQ-substrates) by live cells is imaged. The cells are grown as either two dimensional (2-D) or three-dimensional (3-D) cultures on ECM or embedded in ECM, respectively. The DQ-substrates are mixed into the ECM and provide a protein substrate that can be cleaved by many proteases. This differs from the use of small molecule synthetic substrates that are cleaved selectively by a single protease or protease class.
Originally, the DQ-substrates were developed as non-selective, broad-spectrum substrates for in vitro applications such as analysis of protease activity in solution (Menges et al., 1997). We have adapted their use for live-cell confocal imaging in order to investigate the contribution of multiple classes of proteases to ECM degradation by migrating or invading cells. Visualization and accurate quantitative analysis of ECM proteolysis relies on the fluorescent characteristics of the DQ-substrates. The native, undegraded DQ-substrates are heavily conjugated with the fluorogenic dye, fluorescein. The close proximity of the dye molecules to each other renders native DQ-substrates fluorescently quenched due to a Fluorescence Resonance Energy Transfer (also known as Forster Resonance Energy Transfer or FRET) effect. Upon proteolytic hydrolysis of the DQ-substrate into smaller fragments, the fluorogenic dye molecules are separated, self-quenching is lost, and a strong fluorescence is emitted. The fluorescent signal is proportional to the proteolytic activity, and the pattern depends on the DQ-substrate itself. There is little background fluorescence in the absence of cells for either DQ-collagen I (Fig. 4.20.1A) or DQ-collagen IV (Fig. 4.20.1C). The cellular cleavage of DQ-collagen I (Fig. 4.20.1B) or DQ-collagen IV (Fig. 4.20.1D) differs with the former appearing as fluorescent dots along the collagen I fibrils and the latter as diffuse fluorescence. Depending on the cell lines and ECM, cells will grow as either 2-D monolayers, e.g., on gelatin and collagen I, or as 3-D structures on complex matrices such as reconstituted basement membrane (rBM) (Figure 4.20.2). The ECM can be mixed with any one of an assortment of DQ-substrates (DQ-albumin, DQ-gelatin, DQ-collagen I, or DQ-collagen IV) available through Invitrogen, thus creating customized combinations through which to analyze proteolysis in the context of a defined ECM. Recent studies from our laboratory and others employing this assay have shown that normal and tumor cells use several classes of proteases to actively degrade collagen IV and collagen I and, depending on the cell type, the degradation products are found intracellularly and/or extracellularly (Sameni et al., 2003; Cavallo-Medved et al. 2005; Podgorski et al. 2005; Tsai et al. 2005; Urbich et al. 2005; Sloane et al., 2006). We describe here the use of three combinations of matrices: (1) DQ-collagen I in collagen I; (2) DQ-gelatin in gelatin; and (3) DQ-collagen IV in recombinant basement membrane (rBM; Support Protocol 1).
Cells are grown on ECM containing DQ-substrates as described in Support Protocol 1. Although either an inverted or an upright laser scanning confocal microscope can be used for imaging live cells, this assay is most readily performed using an upright microscope. An upright microscope is ideally suited due to the inherent thickness (1–3 mm) of the culture sample (DQ-substrate in ECM + cells). Regardless of the microscope, the sample thickness requires the use of objectives with long working distances. For live-cell imaging at high magnification on an upright microscope, we routinely use water-immersion dipping objectives (40× and 63× Achroplan) that can be immersed in the cell culture medium. Note that these are different than the normal high-power Apochromat water-immersion objectives. The latter have very-short working distances (10–15 µM) and cannot be used with thick specimens to obtain a sharp focus through the entire sample. For imaging with an inverted microscope, one will use either a lower magnification lens or a long-working-distance lens (see Commentary). In our laboratory, we routinely use a Zeiss upright microscope equipped with Zeiss LSM 510 META confocal optics and argon, HeNe and two-photon lasers. Together with a microscope stage incubator, this setup allows a high level of flexibility in regard to sample dimensions, imaging timeframes, and fluorogenic dyes.
The key to quantitative imaging of proteolysis in three dimensions is the ability to acquire fluorescence without pixel saturation from single planes of focus encompassing the entire specimen. Confocal microscopy allows us to image sequentially above, through, and below the objects of interest. This optical z-sectioning and imaging of accumulated fluorescence gives a spatial 3-D output that represents proteolysis of DQ-substrates. Furthermore, the same field of view can be imaged repeatedly over a period of time to collect data in four dimensions (4-D), i.e., spatially and temporally. Regardless of whether data collection is over time or just a snapshot of one time point, accurate determination of proteolytic activity and localization of such activity requires the collection of data from all the cells and their surrounding 3-D volume (x,y,z-limits). As some cell types aggressively invade into the ECM, the optical sections may need to span several hundred micrometers in the z-direction. Therefore, prelabeling the cells (Support Protocol 1) and their nuclei (Basic Protocol 1) with fluorescent dyes will enable the user to not only define the entire relevant 3-D volume, but also to relate total DQ-substrate hydrolysis to cell number. In addition, prelabeling the cells with cell tracking dyes that occupy the entire cell enables differentiation of intracellular from extracellular proteolysis and their quantification (Basic Protocol 2). The additional data from nuclear and cytoplasmic staining allows normalization of proteolysis to cell number across samples. Visually, most advanced imaging applications are capable of transforming the collection of 2-D images from optical z-sections to graphical 3-D reconstructions of cells and areas of DQ-substrate degradation (Figure 4.20.3).
Following the acquisition of data in the form of z-stacks, image analysis software is used for quantification. The fluorescence data acquired due to the proteolytic cleavage of DQ-substrates can be used to visualize DQ-substrate proteolysis and quantify proteolysis in particular areas or entire volumes (Fig. 4.20.4).
The cell cultures being imaged must be grown on or within ECM containing the desired DQ-substrate. The ECM can exert a dramatically different effect on cellular morphology and physiology. The formation of unique structures such as monolayers, spheroids, acini, tubes, and branching clusters may depend on the composition and stiffness of ECM, media supplements, densities of cell seeding, etc. We have found that on gelatin and collagen type I cells primarily grow as 2-D monolayers like those seen on uncoated plastic and glass. In contrast, in complex matrices they form 3-D structures (e.g., see Fig. 4.20.2).
Prior to setting up cell cultures on ECM containing DQ-substrates, cells should first be prelabeled with a fluorescent cell tracking dye. Alternatively cells can be transfected or transduced to drive expression of a fluorescent protein such as monomeric Red Fluorescent Protein (mRFP), which we have found to remain cytosolic. There are two reasons to use fluorescently labeled cells: (1) to delineate the intracellular area taken up by cells in order to separate intracellular from extracellular degradation using imaging analysis software; and (2) to distinguish among the different cell types in co-cultures. Invitrogen provides several cell tracking fluorescent dyes with distinct excitation/emission spectra. This allows the user to label different cell types with unique colors, thus distinguishing each cell type in co-cultures. The investigator must be aware of the limitation of the microscope for handling multiple colors and availability of excitation and emission band-widths. After establishing cell cultures on ECM containing a DQ-substrate, the user can begin imaging immediately or at a later time. Cells will begin to degrade DQ-substrates shortly after seeding; however, sufficient hydrolysis for detection may require hours to accumulate. If long periods of culture time are required, as when the cells are slow to form 3-D structures or when degradation needs to be monitored in later stages of culture, the 3-D structures may need to be grown in the absence of DQ-substrates, harvested and then re-seeded onto coverslips coated with ECM containing DQ-substrate. Please see Support Protocol 2 for indications when reseeding should be considered.
When cells in rBM need to be grown over long periods of time before imaging, the cultures may first be grown without DQ-substrates and then transplanted into fresh ECM that has been mixed with the DQ-substrate. Harvesting the initial cultures from rBM requires gentle enzymatic digestion of the rBM so that the cellular structures can be removed with minimal disruption. Therefore, harvesting is only recommended for cell types that grow as tight spheroids or aggregates. Harvesting will significantly alter the 3-D morphology of structures composed of tubes, networks and branches.
The ability of cells to degrade ECM had been imaged previously by growing cells on FITC-labeled ECM and looking for a loss in fluorescence as a result of ECM degradation. With this method it is often difficult to assess loss of fluorescence, in particular small losses, because the background fluorescence is so intense. Furthermore, this method does not assess proteolysis in real-time by live cells since the FITC-labeled matrices and cells are fixed prior to imaging. As a result the areas of fluorescence loss may not be associated with cells. For example, we have found tracks in which there is loss of fluorescence when tumor cells are grown on FITC-labeled ECM (Sloane 1996). These tracks suggest that the cells are migrating and degrading the ECM as they migrate. We developed the DQ-substrate assay described here so that proteolysis by live cells could be imaged in real-time.
The fluorescence of the DQ-substrates is quenched as a result of extensive labeling, the close association of FITC molecules and a transfer of non-radiative energy, i.e., a FRET effect. Upon proteolytic cleavage, the distance between the FITC molecules increases resulting in emission of fluorescence (see Fig. 4.20.1B and D). This increase in fluorescence occurs on a non-fluorescent background (see Fig. 4.29.1A and C) and thus is readily observed. Furthermore, the gain in fluorescence can be imaged without fixation. When we use DQ-substrates for imaging of proteolysis by live cells, we see fluorescent fragments of DQ-substrates both extracellularly and intracellularly. At least some of the intracellular fluorescence represents intracellular degradation as cell-permeable inhibitors of lysosomal cysteine cathepsins reduce the amount of intracellular fluorescence (Sameni et al. 2003; Sameni et al. 2000). The presence intracellularly of fluorescent fragments of DQ-substrates requires endocytosis by live cells. This may be due to endocytosis of intact DQ-substrate and its degradation intracellularly or endocytosis of fluorescent fragments of DQ-substrate that had been degraded extracellularly. In either case, the intracellular fluorescence could not be visualized unless the cells were alive, a clear advantage of the DQ-substrate assay. A potential disadvantage is that the DQ-substrates are proteins so they do not allow one to directly assess the activity of any individual protease or protease class. As some DQ-substrates are proteins that would be encountered in vivo by migrating or invading cells, e.g., collagen IV or I, analyzing their degradation could allow one to identify proteases that are involved in normal developmental or pathological processes. Furthermore, more than one protease or protease class can degrade these protein substrates and thus their use should allow one to identify multiple proteases and potentially proteolytic pathways that are involved in these processes.
Another advantage of DQ-collagen substrates is that they allow one to study proteolysis in cells growing in a 3-D context that mimics the in vivo environment. There are several excellent reviews on the value of 3-D in vitro models (Schmeichel and Bissell, 2003; Debnath and Brugge, 2005; Yamada and Cukierman, 2007)
This protocol describes methods to set up and image cell cultures using an upright microscope equipped with long-working-distance dipping objectives. Although an upright microscope is ideal for the sample thicknesses of 3-D cultures, inverted microscopes can be used. Several microscope companies manufacture long-working-distance (2–3 mm) high-power dry objectives such as the Zeiss 40× LD and 63× LD Plan-neofluar, which can be used in inverted microscopes. Along with long-working-distance dry objectives, growing cells on glass-bottom culture dishes (coverslip thickness) will minimize the distance between the objective and the cells within the ECM. Furthermore, microscope-grade glass has optical characteristics that are optimal for high-quality imaging.
Fluorescence from DQ-substrates appears as large, amorphous deposits that are not associated with cells. The DQ-substrates may have been handled improperly following solubilization. Once in solution the DQ-substrates should not be frozen. Do not vortex any solution containing DQ-substrates and avoid excessive pipetting. Also do not let either the substrate or the ECM dry out. Make sure that the lot number of DQ-substrate has not expired or been recalled.
Collagen I does not adhere well to glass. We recommend establishing cell cultures on collagen I polymerized within a 35-mm plastic culture dish. Note that the increased surface area will also mean that higher cell numbers must be seeded.
Specific results will depend on the cells analyzed, the use of monolayer or three-dimensional cultures, and the DQ-substrates and ECM used for the assay. Degradation of the DQ-substrates can be seen within hours after plating and can be imaged in real-time in cultures growing in an incubator on the microscope stage. If stability of the set-up can be maintained, cultures can be imaged in real-time for long time periods; we have imaged for as long as 24 hr (Cavallo-Medved et al., submitted). This requires a stage incubator that can maintain temperature, humidity, oxygen tension, and CO2. Degradation of the DQ-substrates is seen within hours after plating. Not surprisingly, the denatured substrate DQ-gelatin is degraded faster than is either DQ-collagen I or DQ-collagen IV. Whether the DQ-substrates are degraded extracellularly, intracellularly or both may depend on the ECM in which they are mixed. The ECM can affect the morphology of the cells and whether they grow as two-dimensional monolayers or three-dimensional structures (e.g., see Fig. 4.20.2) and in addition can influence the ability of the DQ-substrates to be endocytosized.
3D cellular structures and cocultures require long imaging times. The thicker a 3-D structure is the more optical sections needed to image. It may take minutes to record an optical section into each fluorescent channel; this is particularly true for high-resolution images requiring long scan times. If one is imaging cocultures in which each cell type is labeled with a fluorescent dye/marker, then each color must be scanned into a separate fluorescent channel. Another consideration is that variability among structures requires imaging of a greater number of structures if the data are to be statistically significant.
Time at which one should perform fluorescent labeling. Since cytoplasmic dyes that can be used to prelabel cells will be diluted out with each cell division, these dyes can only be used to label short-term cultures or need to be added shortly before imaging, e.g., 2 hr (Support Protocol 1). The timing of staining with DNA binding dyes such as Hoechst 33342 may affect cell viability. Due to their interaction with DNA, these dyes will induce apoptosis of most cells within 12 hr.
This work was supported by National Institutes of Health (NIH) grant CA 56586, an NIH National Technology Center for Networks and Pathways grant U54 RR020843, a Department of Defense (DOD) predoctoral traineeship award to CAJ (BC051230) and a DOD Breast Cancer Center of Excellence Award (DAMD1702-1-0693). The Microscopy and Imaging Resource Center is supported, in part, by NIH Center Grants U54 RR020843, P30 CA 22453 and P30 ES 06639.