|Home | About | Journals | Submit | Contact Us | Français|
It is increasingly recognized that interactions between cancer cells and their surrounding stroma are critical for promoting the growth and invasiveness of tumors. For example, cancer cells alter the topography and molecular composition of stromal extracellular matrix by increasing paracrine regulation of fibroblastic stromal cells during early tumor development. In turn, these physical and biochemical alterations of the stroma, profoundly affect the properties of the cancer cells. However, little is known about the cross-talk between stroma and cancer cells and it is mainly due to the lack of a suitable in vitro system to mimic the stroma in vivo. We present an in vivo-like 3-D stromal system derived from fibroblasts harvested from tissue samples representing various stages of stroma progression during tumorigenesis. The chapter describes how to isolate and characterize fibroblasts from a plethora of tissue samples. It describes how to produce and characterize fibroblast-derived 3-D matrices. Finally, it describes how to test matrix permissiveness by analyzing the morphology of cancer cells cultured within various 3-D matrices.
One of the fundamental differences between transformed and normal cells is the manner in which they interact with their immediate environment. Benign epithelial tumors are constrained by a surrounding stroma (1, 2) consisting, in part, of fibroblastic cells and fibrillar extracellular matrix (ECM). This normal stroma inhibits or contains tumorigenicity (3, 4). However, at a critical point in transformation, tumors overcome this stromal barrier, inducing changes that promote rather than impede tumor progression (1, 5–13). Changes in the stroma accompanying tumor progression include appearance of discontinuities in the basement membrane surrounding the growing tumor, several immune responses, and the formation of new blood vessels (angiogenesis). Among these host responses are additional alterations to the mesenchymal connective tissue in the vicinity of the tumor (14). Mesenchymal alterations, known as ‘stromagenesis,’ occur in parallel to tumorigenesis and resemble tissue responses during wound healing or fibrosis (12, 14–19). Quiescent fibroblasts (also known as stellate cells in pancreatic and hepatic stromas (20, 21)) are the predominant cell type within a ‘normal stroma’, and secrete an ECM that is believed to provide a natural barrier that constrains tumor progression (1, 16, 22–25). In contrast, the ECM produced by a primed stroma either genetically or epigenetically modified, can provoke, stimulate, and support (instead of constraining) tumor progression (5, 8, 12, 25–29). The primed fibroblasts engage in paracrine and autocrine feedback signaling with the developing tumor cells (11, 30), causing the eventual loss of normal tissue homeostasis (11). In the course of this parallel progression, differentiated myofibroblastic stromal cells, now termed activated cancer- or tumor-associated fibroblasts (TAFs) begin to express a set of proteins including collagen-I, fibronectin (15, 31), desmin, α-smooth muscle actin (α-SMA) (16, 32), and others, grossly altering the protein constituents and architecture of the ECM. During this later ‘activated stroma’ phase or desmoplasia, the tumor becomes invasive and metastatic (33, 34).
As result of the interactions between stroma and cancer cells, the cancer cells modify their morphology and, thus, their migratory mechanism (35–37) Examples of these modifications include, among others, epithelial to mesenchymal and epithelial to amoeboid transitions (35, 38). The cancer cells that present amoeboid morphology present a ‘lymphocytic’ type of movement that is driven by weak interactions with the ECM, it is independent of proteases and is controlled by the small GTPase RhoA and its effector ROCK to generate cortical tension, stiffness and the maintenance of round cell morphology (36, 39). However, cells that undergo epithelial to mesenchymal transition present mesenchymal or spindle morphology and migrate guided by the matrix fibers or strands. The migration of cells with a mesenchymal morphology is dependent on integrin mediated adhesionand ECM degradation by proteases (40).
In this chapter, we describe protocols to isolate stromagenic fibroblasts from various tissue samples and to obtain three-dimensional (3-D) matrices derived from these fibroblasts. The chapter includes methods for characterizing both fibroblasts and their derived matrices in order to sort them as normal, primed or activated, also known as desmoplastic (tumor-associated). The last part of the chapter is dedicated to the analysis of the morphology that is acquired by cancer cells when cultured within the various fibroblast-derived 3-D matrices.
NOTE: All solutions and equipment coming into contact with tissue samples or living cells must be sterile. Therefore, aseptic techniques should be used accordingly.
GraPhpad Instant software (or any other simple statistical software).
NOTES: All solutions and equipment coming into contact with tissue samples or living cells must be sterile and aseptic techniques should be used accordingly. All cell-culture incubations are performed using a 37°C, 5–10% CO2 humidified incubator.
Fibroblasts synthesize and maintain the ECM of mesenchymal tissues. The main function known for fibroblasts is maintaining the structural integrity of all connective tissues by continuously secreting components (e.g., soluble cytokines, latent factors and matrix glycoproteins) and actively incorporating them to the ECM. The composition of a given ECM determines the specific physical and biochemical properties of each connective tissue.
This section describes two methods used to harvest primary fibroblasts from normal and/or neoplastic tissue samples (see Note 3). The first method, minced tissue method (Section 3.1.1), is based on the capability of fibroblasts to crawl out of tissue samples, thus facilitating their harvest. Minced tissue method provides fairly homogenous fibroblastic cell-cultures. Therefore, this method assures the isolation of a variety of different subpopulations of fibroblasts present in a given (normal or tumor-associated) tissue. The second method, enzymatic digestion method (Section 3.1.2), is faster and yields a higher recovery rate of cells from the tissue. Since this is a much faster method, contaminations are less common than in the first method. Fibroblast cultures obtained using the second method, are heterogeneous and probably better represent the fibroblastic population of a given tissue. Unfortunately, the heterogeneous aspect of this procedure often results in cultures that contain additional types of cells (e.g., epithelial). This method is adequate for fibroblasts that are impaired in their motile capabilities, or when heterogeneous cultures are needed.
In the minced tissue method (first method), the tissue samples are cut into small pieces and each of these pieces is separately placed in a tissue culture plate until fibroblasts migrate out of the tissue. In the enzymatic method (second method), tissue samples are actively digested and cells sorted by size exclusion. Once the fibroblast cultures reach confluence (in both methods) they can be frozen for later or expand for immediate experimental use.
NOTE: The isolated fibroblasts can be immortalized (e.g., using SV40 large T antigen). If the fibroblasts are immortalized, cultures would need to be re-characterized later on (see section 3.2).
NOTE: The isolated fibroblasts can be immortalized (e.g., using SV40 large T antigen). If the fibroblasts are immortalized, cultures would need to be re-characterized later on (see below).
Although most contaminant cells (epithelial and/or endothelial) perish after a few passages, it is always recommended to assure that the harvested cell population contains only fibroblastic cells. This can be assessed using specific cell markers, as well as by phenotypic analysis. Typically, fibroblasts are large, flat and relatively elongated cells with branched bodies that surround an oval and speckled nucleus. Fibroblasts express, among others, vimentin. In contrast, cytokeratin is an epithelial (and epithelial tumor) cell marker. In this section, we describe how to assess the homogeneity (under microscope) and the exclusivity (using a Western blot technique, see Note 12), of the harvested fibroblastic cell population. The isolated cells are plated onto 35-mm dishes and incubated overnight, their morphology and homogeneity is analyzed under a microscope, while cellular proteins are extracted and separated by SDS PAGE, followed by transfer into PVDF membranes and submitted to Western blot analyses using marker-specific antibodies.
NOTE: All solutions and equipment coming into contact with tissue samples or living cells must be sterile, and aseptic techniques should be used accordingly.
This section describes a method for generating cell-derived 3-D matrices produced by a plethora of primary fibroblasts. The resultant matrices can be used as substrates for culturing cells, since they closely resemble in vivo mesenchymal matrices (41, 42). Utilizing in vivo-like 3-D matrices as substrates allows to study, in a physiologically relevant manner, how cells interact with their natural ECMs, as well as their microenvironmental induced structures, functions, and signaling, which differ from observations obtained by culturing cells on conventional 2D substrates in vitro (43–48).
In the protocols described in this section, the fibroblasts used are the primary fibroblasts harvested from the assorted normal and tumor tissue-samples described in previous sections 3.1.1 and 3.1.2. Following the description for 3-D matrix production, we describe how to extract cellular debris from the matrices.
The basic approach is to allow fibroblasts to produce their own 3-D matrices. For this, fibroblasts are plated and maintained at a confluent state, while fresh ascorbic acid (stabilizing the secreted matrices) is added every other day for a period of five to nine days. Matrices are extracted, using an alkaline detergent treatment, thus removing cellular debris and leaving the 3-D matrices intact and stably attached to the culture dishes. The fibroblast-derived 3-D matrices contained within the dishes are then stored at 4°C for a couple of months.
NOTE: All solutions and equipment coming into contact with tissue samples or living cells must be sterile and aseptic techniques should be used accordingly.
OPTIONAL: For immunofluorescence experiments 22-mm coverslips are to be used. Coverslips need to be sterilized by flaming after dipping with absolute anhydrous ethanol. Proceed by placing the coverslips at the bottom of a 35-mm tissue culture dishes and rinse with PBS. Remove the PBS and continue to step 1.
Tumor-associated (e.g., desmoplastic) stroma has been associated with a variety of invasive cancers (15, 49). This stroma presents a scar-like phenotype, is highly fibrotic and can constitute more than 50% of the tumor mass. The desmoplastic stroma is characterized by the presence of activated myofibroblasts, which are highly proliferative and express alpha-smooth muscle actin (α-SMA). Three-dimensional matrices derived from primary fibroblasts harvested at different stages of tumor development differ in their orientation of fibronectin fibers, expression and organization of α-SMA and the morphology of both their cell body and nucleus (42). Therefore, characterization of the above-mentioned features can be used to sort unextracted 3-D matrix cultures as normal, primed or tumor-associated. For example, matrices derived from primary tumor-associated fibroblasts that are desmoplastic, present a parallel patterned matrix with high and homogenous α-SMA expression localized on stress fibers and elongated elliptical nuclei morphology (Figure 2). Matrices produced by primary primed fibroblasts present a more random organization of fibronectin fibers, relatively rounded nuclei and either lack α-SMA expression or express α-SMA at relatively low levels. Primary fibroblasts isolated from normal tissues normally produce very thin matrices and the majority of these do not overcome growth inhibition by contact, therefore, cultures are mono-layered. Nevertheless, some normal fibroblasts isolated from specific cites, such as normal ovaries, grow multi-layers when maintained in vitro as confluent cultures. Similarly to primed matrices, 3-D matrices obtained from normal primary fibroblasts are greatly disorganized and the unextracted nuclei also appear relatively round. However, in comparison to primed unextracted cultures, normal unextracted 3-D cultures vary on the expression of α-SMA from cultures that homogenously express high levels of α-SMA (e.g., normal ovarian-derived 3-D cultures (50)) to heterogeneous or low homogenous expression levels of this protein (e.g., skin and pancreas-derived 3-D cultures).
The classification of unextracted in vivo-like 3-D stromal matrices produced by assorted isolated fibroblasts (sections 3.1.1 and 3.1.2) as normal, primed, or tumor-associated (desmoplastic), can be questioned using indirect immunofluorescent staining. In this section, unextracted 3-D stromal matrices, prepared onto coverslips (see section 18.104.22.168), are fixed and permeabilized, and subjected to multi-channel simultaneous fluorescent labeling of matrices (e.g., fibronectin), nuclei and α-SMA. Following multi-channel fluorescent labeling, the unextracted 3-D cultures are analyzed under a scanning or spinning disk confocal microscope.
Listed below are some characteristics that will assist to classify fibroblastic unextracted 3-D cultures as “normal,” “primed” or “tumor-associated” (42) by using the three channel images acquired in the previous section (3.4.1) (see Note 28 and Figure 2):
Some examples of the above-mentioned characteristics for “tumor-associated” 3-D cultures are shown in Figure 2.
One of the main features of cancer progression resides in the fact that when tumors become invasive the basement membrane that normally isolates epithelium from mesenchyme becomes degraded and, therefore, invasive cancer cells directly interact with the mesenchymal stromal components both prior to intravasation and after extravasation at the secondary sites. The specific morphological phenotype acquired by invasive cancer cells while migrating within mesenchymal tissues is predictive of their invasive strategy behavior (35–37). For example, it is well known that tumor cell migration and metastasis can occur by multiple mechanisms (e.g., epithelial-mesenchymal transition, or mesenchymal-amoeboid transition (35–37). These various mechanisms require different signaling pathways, directly induced by the stroma and are clearly distinguished by specific cell morphologies (e.g., mesenchymal vs. amoeboid). Therefore, testing whether cells acquire differential morphologies within different stromagenic staged 3-D matrices could predict how the specific cells would invade within specific microenvironmental settings (e.g., normal, primed or tumor-associated stroma). Amoeboid cells are relatively rounded, while mesenchymal cells are spindled-shaped. Mesenchymal cell invasion requires the function of integrins and specific matrix-proteases while amoeboid invasion is independent of integrins and matrix-proteases functions and instead requires the activation of the ROCK pathway (39, 40).
In this section, we provide a method for evaluating 3-D matrix-induced epithelial cancer cell morphology. Prior of cell re-plating within the assorted 3-D matrices, the nuclear and cytosolic compartments of the epithelial cancer cells are fluorescently labeled. Then cells are re-plated within the assorted matrices overnight and their morphologies are measured following the acquisition of several representative double-channeled monochromatic images using an epifluorescence microscope equipped with filters for the acquisition of the specific fluorophores used.
The digital analysis can be carried out manually or automatically, this section describes how to measure cell morphology using the MetaMorph 7.01 software (Molecular Devices):
We thank M. Valianou for critical comments and K. Buchheit for assertive proof reading. This work was supported by the American Association of Cancer Research (AACR) Pennsylvania Department of Health (the Department specifically disclaims responsibility for any analyses, interpretations, or conclusions), the National Institutes of Health/National Cancer Institute (grants CA006927, and RO1-CA113451), the Ovarian Cancer Research Foundation (OCRF), the W.W. Smith Charitable Trust and an appropriation from the Commonwealth of Pennsylvania.
Note 1In order to stain the Nuclei, other dyes can be used (e.g., Hoechst or DAPI) but since most confocal microscopes do not have UV filters to detect the signal of these fluorescent dyes, we proposed to use the fluorescent SYBR Green dye, which absorbs blue light (λmax = 498 nm) and emits green light (λmax = 522 nm) and whose fluorescence staining can be analyzed using the 488 nm channel available in most of the confocal microscopes.
Note 2Instead of a confocal microscope, an epifluorescent microscope equipped with a motorized Z-motor and deconvolution software can be used.
Note 3Primary fibroblast cultures may be obtained from different tissues, normal or tumor, including but not limited to human tumor tissue from ovary, pancreas, lung or breast and from tissues of other species such as mouse and rat.
Note 4The tissue tumor sample can be cut into two pieces, one can be used to obtain fibroblasts and the other can be frozen for further analysis, such as protein localization using immunohistochemistry or immunofluorescence techniques. To freeze half of the sample, put the tissue sample inside a plastic mold and cover with embedding medium (e.g. Tissue-Tek). Using tweezers, freeze the sample by floating the mold on liquid N2, avoiding the liquid N2 to directly contact the sample, once frozen, store at −80°C.
Note 5If the tumor tissue is detached from the dish, it can be placed in a new scratched dish repeating steps 4–6 of section 3.1.1.
Note 6When subculture or recovery of the cells from a plate is required, remove the medium from the flask and rinse cells briefly using pre-warmed (37°C) trypsin-EDTA to remove trypsin inhibitors contained in the serum used for culturing the cells. Then, add enough trypsin-EDTA to slightly cover the cells on the flask and observe under an inverted microscope until the cells detach from the culture dish and become rounded and not clustered to each other (1 to 3 min). Once the cells have completely detached from the bottom of the flask, add 10ml of fibroblast medium to neutralize the trypsin and collect the cells. Pipette up and down carefully to mechanically disrupt the remaining cell aggregates. At this point, cells can be sub-cultured or frozen.
Note 7Before freezing, the fibroblasts should be actively proliferating therefore ensuring that no contaminations will be held within the frozen samples.
Note 8If the starting amount of tissue sample is large, the volumes can be scaled up; use a bigger Erlenmeyer assuring that the solution will not be spilled out while stirring. For instance, if the final volume of DMEM will be 40 ml, a 250 ml Erlenmeyer should be used and 4 ml Collagenase-3 (10X) should be added (for details see section 3.2.1).
Note 9When the pieces of the tissue are not dissolved after a period of 1h agitation, fresh collagenase-3 can be added or, alternatively, the tissue pieces can be mechanically dissociated by carefully pipetting up and down.
Note 10Prior to centrifugation, the centrifuge should be equilibrated by preparing another 50 ml polypropylene tube that weighs the same as the 50 ml polypropylene tube containing the digested sample and then, samples should be placed on opposite sides of the centrifuge.
Note 11This method renders an heterogeneous fibroblastic population. If homogenous clones are needed, then dilute cell concentration at this point, and plate them within multi-well tissue culture plate (96 or more wells), for sub-clonal selection.
Note 12The Western blot technique described to determine the cell expression of specific markers is based on the use of a modified secondary antibody linked to a reporter enzyme, which when exposed to an appropriate substrate drives a colorimetric reaction and produces a color or precipitate, or a luminescent reaction (ECL) allowing detection by exposing onto photographic films. Nowadays, a variety of conjugated secondary antibodies are commercially available and blotted proteins can even be detected by infrared emitted fluorescence (e.g., using the Li-cor’s Odyssey Infrared Imaging System and its pre-conjugated secondary antibodies).
Note 13From this point on, no sterile conditions are required. Instead, it is very important to keep all material on ice, at all times, in order to minimize protease activity thus avoiding protein degradation.
Note 14For larger dishes, scale up the volume of RIPA buffer. (e.g., for 60 mm dishes use 500 μl instead of the suggested 250 μl for 35 mm dishes).
Note 15To quickly freeze cell lysates, prepare a dry-ice/isopropanol bath by placing a 400 ml beaker containing about 100 ml isopropanol within an ice bucket filled with dry ice. For safety, do this inside the chemical hood. Allow the isopropanol to cool for 30 minutes. Place the tubes containing freshly prepared and aliquoted lysates to a tube-rack and slowly lower the rack into the isopropanol assuring that the lysate volume is immersed in the isopropanol. The lysates should be frozen almost immediately (smaller aliquots are better). Finally, quickly place the tubes on dry-ice for immediate transfer to a −80°C freezer. Samples should remain stable for 2 weeks.
Note 16In order to assure that the Western blot technique has properly worked, known fibroblastic and epithelial cell lysates should be loaded as positive controls (see Figure 1).
Note 17In these protocols, we propose to use commercially available pre-cast gels. Nevertheless, gels can be poured at the lab just prior to use. For more information about SDS-PAGE technique consult a student’s biochemistry text book or a molecular cloning laboratory manual.
Note 18Since the molecular weight of vimentin is similar to keratins, it is necessary to use two gels. One of them will be used to detect the presence of vimentin and GADPH and the other to detect the keratins. If Li-cor’s Odyssey Infrared Imaging System is used, only one gel will be required since vimentin and keratin could be labeled by different fluorophores. In that case, the anti-vimentin antibody recommended is one that was generated in rabbit (e.g., Biovision, # 3634-100).
Note 19Depending on the equipment used, different voltages, times and/or buffers may be required. Therefore, check the manufacturer’s instructions prior of using any SDS-PAGE or transfer equipment.
Note 20Following a 4°C overnight step, membranes should be brought back to room temperature before moving on to the next step.
Note 21The desired final number of plates to be used should be calculated before starting. The protocols describe the amount of volume needed for an individual 35-mm dish. Final volumes need to be calculated in respect to the final amount of desired 3-D matrix-coated plates and their types. For example, for 12-, 24- or 48-well tissue culture plates scale down the volumes of all reagents from 2 ml (35-mm dish) to 1, 0.5 and 0.250 ml per well, respectively. Alternatively, for the use of 60-mm or 10-cm dishes scale up the volumes of added reagents from 2 ml per dish to 4 and 10 ml, respectively.
Note 22Do not proceed to next step if cultures have not reached 100% confluence the following morning. If, after 24h, the dish does not appear completely confluent, change medium and wait until cultures reach 100% confluence. The lack of confluence prior to the next step can cause poor matrix production or quality or avoid matrix production all together.
Note 23If coverslips are used, transfer the coverslip to a 6-multiwell bacterial (instead of a regular tissue-culture) petri dish before adding the ascorbic acid. This will minimize the growth of fibroblasts on the plate area outside of the coverslip and will, especially, facilitate lifting the coverslip thus avoiding tearing off the matrices at the final steps.
Note 24Leave at least two dishes unextracted in order to use in section 3.4.1 for the characterization of matrices resulting in categorizing the fibroblasts (that produced the matrices) as normal, primed or tumor-associated.
Note 25At least two coverslips containing unextracted 3-D matrices should be analyzed for each isolated fibroblastic cell line.
Note 26Before adding the Block Vector Solution make sure that the coverslips are not touching the well walls. This will prevent loss of blocking solution due to capillarity, which could result in sample drying. If samples appear to be drying compensate with Block Vector Solution and continue with the incubation.
Note 27Before acquiring pictures make sure samples are at room temperature.
Note 28The classification described in this section is based on the characteristics of the unextracted cultures (42), which are matrix-dependent and, therefore, are not evident in 2-D cultures.
Note 29Matrix thickness can vary in different cultures of the same cell-line. It often depends on passage-number and quality of reagents used (e.g., FBS).
Note 30Make sure that matrices are pre-warmed to room temperature after storage at 4°C by placing the plates containing extracted matrices at room temperature for at least one hour prior to use.