Systems microscopy approaches to understand cancer cell migration and metastasis
- 1.4k Downloads
Cell migration is essential in a number of processes, including wound healing, angiogenesis and cancer metastasis. Especially, invasion of cancer cells in the surrounding tissue is a crucial step that requires increased cell motility. Cell migration is a well-orchestrated process that involves the continuous formation and disassembly of matrix adhesions. Those structural anchor points interact with the extra-cellular matrix and also participate in adhesion-dependent signalling. Although these processes are essential for cancer metastasis, little is known about the molecular mechanisms that regulate adhesion dynamics during tumour cell migration. In this review, we provide an overview of recent advanced imaging strategies together with quantitative image analysis that can be implemented to understand the dynamics of matrix adhesions and its molecular components in relation to tumour cell migration. This dynamic cell imaging together with multiparametric image analysis will help in understanding the molecular mechanisms that define cancer cell migration.
KeywordsCancer Cell migration Matrix adhesions Dynamic imaging Multiparametric analysis
Cell migration, both single and collective, is a highly integrated multistep process that is essential in embryonic morphogenesis, tissue homeostasis and immune surveillance. While collective migration requires the movement of cohesive groups of cells , the single migrating cell is highly polarised with complex regulatory pathways that are spatiotemporally controlled . Migration contributes to several important pathological processes, including cancer progression and metastasis formation. Metastasis, dissemination of malignant tumours to a distant organ, is the major cause of cancer mortality. Tumour cell motility is the hallmark of invasion and is an essential step in metastasis.
Cell migration can be seen as a cyclic process. The initial response of a cell to a chemotactic signal is to polarise and extend protrusions in the direction of movement. These protrusions are usually driven by actin polymerisation, and are stabilised by adhering to the extracellular matrix (ECM). These adhesions serve as traction sites for migration as the cell moves forward on top of them, and they are disassembled at the cell rear, allowing it to detach. This process depends on the cell type and environment. Matrix adhesion turnover is clearly visible in slow-moving cells such as fibroblasts or epithelial cells which show large protrusions, and is less visible in fast-moving cells such as neutrophils or cancer cells, which display small protrusions with a rapid turnover . Interestingly, the movement of cell sheets shows some features of single-cell migration; however, the polarisation extends over the entire sheet . In addition to typical matrix adhesions, typically referred to as focal adhesions, cells can form another variety of adhesive structures, namely podosomes and invadopodia, also termed podosome-type adhesions (PTAs) . Those unique actin-rich adhesion structures are typically associated with sites of proteolytic degradation of the extracellular matrix components. This is thought to contribute to cellular invasiveness in physiological and pathological situations. Cell types that form podosomes include monocytic, endothelial and smooth muscle cells, whereas invadopodia are mostly observed in carcinoma cells [4, 5, 6]. In this review, we will focus on the study of matrix adhesions in relation to tumour cell migration and invasion; however, most of what we will discuss here is also relevant for the study of PTAs [7, 8].
Although matrix adhesion dynamics is very important for understanding cell migration behaviour, little is known about the molecular mechanisms that regulate adhesion dynamics and signalling during tumour cell migration and invasion. Advances in microscopic imaging techniques including spatiotemporally resolved imaging, fluorescent reporter reagents and multiparametric image analysis will contribute to a better insight into these processes. In this review, we will focus only on single cell migration and discuss the current and emerging imaging technologies that can be implemented to study adhesion dynamics and signalling in migrating cells. Tables 1 and 2 summarise all the existing techniques to study, respectively, protein dynamics and modes of tumour migration and can be used as guidelines. In parallel, we discuss the different multiparametric image analysis tools that can be applied after image acquisition to generate more detailed and reliable cellular and molecular measurements. Table 3 gives a non-exhaustive list of challenges and potential solutions for improving the ‘throughput’ of this systems microscopy approach. Finally, studying the fundamental mechanisms underlying cell migration and invasion using physiologically relevant experimental models together with dynamic imaging and multiparametric image analysis will help in discovering new interesting targets and therapeutics for inhibition of tumour invasion and consequently cancer metastasis formation.
Matrix adhesion complexes are cytoplasmic structures that have been originally identified by electron microscopy  or by interference reflection microscopy . They are the closest site of contact between the cell and the underlying substratum. Integrins are the principal cell surface adhesion receptors mediating cell–matrix adhesions . Integrins are heterodimeric receptors that directly bind extracellular matrix molecules and couple them to the actin cytoskeleton (for reviews, see [12, 13]). Integrin cytoplasmic domains form multi-molecular complexes with proteins involved in cell adhesion signalling and with adaptors that provide a connection to the cytoskeleton . Approximately 150 proteins, which form the so-called ‘integrin adhesome’, have been retrieved to date to be part of the matrix adhesions including kinases, phosphatases and structural proteins (for review, see ). Even a larger number of over 600 proteins are defined to be involved in the spatiotemporal regulation of matrix adhesions (van Roosmalen et al., unpublished data). Upon attachment, integrins will cluster and promote local recruitment of structural proteins like vinculin, paxillin, talin, α-actinin and tensin, and signalling molecules including tyrosine kinases such as focal adhesion kinase (FAK), serine/threonine kinases and various adapter proteins. The molecular complexity of cell–matrix adhesions enables them to fulfill their dual role as modulators of both mechanical cell anchorage and transmembrane signalling . Integrins are not the unique transmembrane receptors that have been described to control and regulate cell adhesion co-signalling; syndecans, discoidin domain receptors (DDR), CD44 and dystroglycan can also bind directly to the ECM, while growth factor receptors such as the epidermal growth factor receptor (EGFR) can crosstalk with the integrins to regulate the recruitment of cytoplasmic proteins to adhesion sites (for reviews, see [15, 16]). Syndecans as well as DDRs can bind to ECM molecules but most of all they synergise with integrins . DDRs are a novel classs of receptor tyrosine kinases that bind to several collagens and stimulate matrix metalloproteinases (MMP) production. CD44, ICAM-1 and ICAM-2 and L-selectin are other adhesion receptors that are able to interact with ECM ligands as well as with ezrin, radixin, moesin (ERM) family proteins. Dystroglycan is a ubiquitously expressed heterodimeric adhesion receptor that can localise at matrix adhesions . Dystroglycan in concert with integrins is involved in the spatial and temporal remodelling of adhesion numbers, types and distribution. It has emerged as a multifunctional adhesion platform with many interacting partners (for review, see ). One emerging area is the role of dystroglycan in cancer. Dystroglycan function appears to be disrupted in numerous epithelial-derived cancers . A lot is known on the role of integrins in cell migration; in contrast, the mechanistic contribution of distinct adhesion systems such as syndecans, DDRs, CD44 and dystroglycan still remains incompletely understood. Advanced methods are necessary to systematically study the mechanism of regulation of these diverse adhesion systems.
Adhesion types in 2D and 3D
Adhesion turnover during cell migration
Cell migration is an integrated process that requires the coordinated regulation of various structural and signalling molecules, including distinct kinases and phosphatases [27, 28]. Cell migration requires the establishment of cell polarity to create a leading edge and a trailing edge. The leading edge undergoes membrane protrusive activities driven by actin polymerisation that establish new matrix contacts, whereas at the trailing edge, cell adhesions are disassembled to promote retraction of the cell rear and forward cell movement. The rate of cell migration can be limited by the rate of rear retraction, and thus the dynamic formation and disassembly of cell–matrix adhesions are critical to cell migration [2, 28].
Formation of adhesions
The mechanism by which adhesions assemble in migrating cells is still under investigation. Some cells, particularly rapidly migrating ones such as leukocytes, have few visible integrin clusters, and thus very small submicroscopic adhesions are probably important for their migration. In other cells, small adhesions known as focal complexes can be observed at the leading edge. Formation of these adhesions depends on Rac- and Cdc42-GTPases, and these adhesions stabilise the lamellipodium by mediating attachment to the ECM, thereby contributing to efficient migration. However, cells with large integrin clusters (“focal adhesions”) are tightly adherent and are typically either non-migratory or move very slowly. The assembly of focal adhesions involves Rho-GTPase as well as myosin-induced contractility. During their formation, some protein components enter adhesions with similar kinetics, which suggests that they exist in preformed cytoplasmic complexes. However, other components enter adhesions with very distinct kinetics, which is consistent with a model in which a regulatory event initiates the serial addition of different proteins .
Adhesion disassembly at the front and the rear
Adhesion disassembly is observed both at the leading edge, where it accompanies the formation of new protrusions, and at the cell rear, where it promotes tail retraction. At the front of migrating cells, adhesions at the base of a protrusion disassemble as new adhesions form at the leading edge (for review, see ). However, some adhesions persist and mature into larger, more stable structures. Little is known about adhesion disassembly versus maturation; however, targeting of microtubules has been implicated as one factor that promotes adhesion disassembly (for review, see ). Both protein kinases and phosphatases also appear to be central to the regulation of adhesion turnover and stability . For example, cells lacking the tyrosine kinases FAK or Src have more and larger adhesions and migrate poorly . Adhesion turnover in migrating cells is also regulated by a complex of Rac-associated proteins [32, 33]. Also, the calpain family of intracellular proteases is implicated in regulating cell migration . Calpain activity is tightly controlled in vivo as this intracellular protease irreversibly cleaves numerous signalling and structural proteins with widespread impacts on cell functioning and viability. The ubiquitous calpains can cleave a large number of adhesion complex components including talin, paxillin, vinculin, ezrin, alpha-actinin, FAK, E-cadherin, and the cytosolic tails of some of the beta-integrins [28, 34, 35, 36]. Calpain activity which relies on the removal of a protein or elimination of an enzymatic activity is a key regulator of matrix adhesion dynamics. For instance, the study of Franco and co-workers demonstrates that calpain-mediated proteolysis of talin is a rate-limiting step in adhesion disassembly . In conlusion, functional studies that will sytematically determine the role of individual proteins involved in cell–matrix adhesions using dynamic imaging of the protein localisation, kinetics and interactions will shed a more complete light on the mechanisms of cell migration. RNAi (live) cell imaging-based screening approaches are key to such an analysis.
Cell migration/matrix adhesion and cancer
Metastasis is the most frequent cause of death for patients with cancer. Tumour cell motility is the hallmark of invasion and is an initial step in metastasis. In order to metastasise, cancer cells must first detach from the primary tumour, migrate, invade through tissues, and attach to a second site. Invasive carcinoma cells acquire a migratory phenotype associated with increased expression of several genes involved in cell motility such as matrix adhesion-associated genes. Many metastatic human carcinomas are characterised by the overexpression or constitutive activation of ErbB tyrosine kinase receptor (EGFR) family members involving activating mutations of the receptor kinases or an autocrine loop with EGF family ligands. High expression of EGFR and ErbB2, as well as another tyrosine kinase receptor c-Met, are associated with poor prognosis of breast cancer patients. In addition to stimulating cell differentiation and proliferation, EGF promotes tumour cell motility, invasion and metastasis [38, 39, 40]. Recent data provide evidence for a requirement for certain focal adhesion protein expression (e.g. integrin, FAK/Pyk2, paxillin, ILK, Ezrin) in metastatic dissemination. Integrins are known to regulate a diverse array of cellular functions crucial to the initiation, progression and metastasis of solid tumours. The importance of integrins in several cell types that affect tumour progression has made them an appealing target for cancer therapy . Increased activity or overexpression of Src is a frequent occurence in many types of human cancer. Indeed, a prominent role of Src is played in invasion, epithelial-mesenchymal transition (EMT) and development of metastasis. Inhibitors of Src are promising drugs for cancer therapy (for reviews, see [42, 43]). Focal adhesion kinase (FAK), another well-studied non receptor tyrosine kinase, plays a key role in cancer progression. FAK expression and activity are enhanced in metastatic tumours of diverse origin [44, 45, 46]. FAK is considered as a promising gene candidate, and inhibitors of FAK are currently in clinical trial (for reviews, see [47, 48]). Not only activity of kinases but also activity of proteases such as calpain is enhanced upon transformation induced by the v-Src, v-Jun, v-Myc, k-Ras, and v-Fos oncoproteins. Elevated calpain activity commonly promotes focal adhesion remodelling, disruption of actin cytoskeleton, morphological transformation, and cell migration, although proteolysis of target substrates (such as focal adhesion kinase, talin, and spectrin) is differently specified by individual oncoproteins [34, 35]. Therefore, studying the motility mechanisms used by cancer cells would clarify some of the key events influencing metastasis in cancer. In addition, identification of the molecular pathways that play a role in cancer cell motility will provide new diagnostic approaches and targets for the treatment of metastatic cancer.
Studying migration and adhesion dynamics in living cells
The above indicates a general requirement for advanced technologies to systematically study both cell adhesion dynamics and tumour cell migration. Recent advances in fluorescence probes and microscopy technologies have provided powerful approaches that present advantages over the traditional biochemical approaches. In particular, the newly developed recombinant fluorescent proteins (FPs) and genetically encoded biosensors are useful tools for imaging protein distribution, dynamics and interactions in live cells with high temporal and spatial resolutions enabling the elucidation of molecular mechanisms behind adhesion turnover responsible for the cell migration [49, 50, 51, 52, 53, 54].
Phase contrast/Normaski imaging
The starting point for many investigations of cell movment is to characterise the behaviours of cells, by recording shape changes, direction and persistence of movement and the dynamics of subcellular membrane structures such as lamellipodia, filopodia, protrusions and membrane ruffling. This can be done with wide-field microscopy in either phase-contrast or Normarski Interference Contrast modes which is fast, not photo-toxic to cells and can be performed over a long period of time. Many software packages are now available that automatically track moving cells visualised in this way: e.g. CellTrack, ImageJ (plug-in MTrack2), Imaris, Image Pro and in-house-developed cell track systems [55, 56]. The tracking rates of those different softwares are very good, although application of fluorescence microscopy instead of phase contrast can improve the tracking efficiency . Furthermore, due to high signal to noise ratio in phase-contrast or Normaski imaging, it remains difficult to analyse other cell features. Fluorescent imaging is an answer to that problem.
Fluorescent live cell imaging
Cell migration: wide-field
Confocal laser scanning microscopy (CLSM)
While epi-fluorescence microscopy provides the high degree of time resolution which is required to visualise fast processes, confocal laser scanning microscopy (CLSM) offers several advantages over conventional optical microscopy, including controlled depth of field, the elimination of image degrading out-of-focus information, and the ability to collect serial optical sections from thick specimens. It provides high spatial resolution, but still many physiological processes and events take place faster than they can be captured by most CLSMs, which have image acquisition rates typically in the order of one frame per second. CLSMs using acousto-optical devices and a slit for scanning are faster than the galvanometer-driven point-scanning systems, and are more practical for physiological studies. These quicker designs combine good spatial resolution with good temporal resolution, which may be 30 frames per second at full screen resolution, or near video rate. The slower point-scanning microscope systems can achieve the best temporal resolution only by scanning a much reduced area on the specimen. If full spatial resolution is required, the frames must be collected less frequently, losing some temporal resolution. The confocal systems using disk-scanning or oscillating mirror-scanning methods are also capable of imaging rapid physiological or other transient events. Nowadays, new developments in confocal microscopy provide both high-speed and high-resolution imaging to capture intracellular biological processes. A resonant confocal scan head dramatically improves time, spatial and spectral resolutions and allows high-speed imaging up to 230 frames per second (512 × 64 pixels) which make it an ideal technique for bleaching and protein kinetic experiments (photo-activation and photo-bleaching, see below). A major drawback of confocal microscopy is photo-bleaching and photo-damage from the illuminating laser beam that can be cumulative over multiple scans. So the exposure to the beam should be kept to the minimum necessary to acquire the image. Another recent development is controlled light exposure microscopy (CLEM) which should be ideal for live imaging as it helps to reduce photo-bleaching and photo-toxicity; the two main limitations in live-cell confocal microscopy . In CLEM illumination excitation light is reduced using two strategies. The first is based on the principle that if there is no signal, then no illumination is required (for example, when imaging the background). The second detects whether there is sufficient signal to acquire an image. If so, illumination is stopped.
Total internal reflection fluorescence (TIRF) microscopy
Studying protein dynamics and interactions in adhesion
Imaging techniques to study protein dynamics and interactions in adhesion
What is next
Measure high resolution diffusion, trafficking and stability of protein
Actin dynamics, protein movement, integrin receptor trafficking
For in vivo imaging, to track in the long term photoconverted cells and study protein dynamics
FRAP or FLIP
Measure t 1/2, k on, k off, diffusion coefficient but individual measurements
Integrin, FAK, paxillin, zyxin, vinculin, and actin
Comined FRAP and/or FLIP with TIRF and/or FRET
Measure t 1/2, k on, k off, diffusion coefficient and protein mobility parameters. Include all bleached and unbleached matrix adhesions
FAK, paxillin, vinculin, zyxin and actin
Comined FRAP-FLIP with TIRF and/or FRET
Le Dévédec et al., unpublished data
FCS, ICS, RICS
Determine rates of diffusion, degree of aggregation, number of fluorescent entities and flow velocities (mainly used in solution)
α5-integrin, α-actinin, FAK, paxillin and actin
In living cells to study protein distribution, dynamics and interactions at high time and spatial resolution
Movement of structure, assembly dynamics, and subunit turnover
Actin, integrin, FAK, talin and α-actinin
Protein–protein interaction and protein activity
Src, FAK, Rho GTPases, and matrix adhesion proteins
Combined with FRAP and/or TIRF and in vivo
Photo-activation, Fluorescence recovery after photobleaching (FRAP) and fluorescence loss in photo-bleaching (FLIP)
Many fluorescent proteins (FPs) have been designed by mutagenesis to change their fluorescence intensity (photo-activation) or colour (photo-conversion) upon illumination by lights with specific wavelength, intensity and duration (for review, see ). Those FPs can be grouped into three classes: (1) irreversible changes in colour upon illumination which includes photo-activatible (PA)-GFP  that can be photo-activated with UV illumination and emits green fluorescence; (2) irreversible change in colour from green to red upon UV illumination which includes Kaede  and Dendra ; and (3) reversible change in intensity/colour upon UV illumination such as Dronpa  which is a very valuable marker for photo-conversion studies and fast cellular processes. Until now, mainly PA-GFP has been successfully used to investigate actin dynamics , protein adhesion movement  and trafficking of integrin receptors . Those photo-switchable fluorophores are valuable probes than can be used in live cells to monitor with high resolution diffusion, trafficking and stability of protein targets.
In a FLIP experiment, the loss of fluorescence in a region or structure far from the bleached region is monitored. FRAP and FLIP can also be combined (FLIP–FRAP): two regions are monitored simultaneously after bleaching only one of them. We successfully developed a FLIP–FRAP bleaching procedure so that in one experiment we can analyse all the focal adhesions distributed over the whole cell body (Le Dévédec et al., unpublished data). Together with Monte Carlo simulation, we observed that FAK and paxillin had an equal diffusion rate but a differential residence time that is related to adhesion size and strength. With this technique, we are able to extract mobility parameters of adhesion proteins as well as a mapping of the protein kinetics according to focal adhesion size, type and localisation in the cell (Fig. 4b).
Fluorescence speckle microscopy (FSM)
Another fluorescent technique that detects protein dynamics, turnover and interaction is a method called fluorescent speckle microscopy (FSM) that uses a very low concentration of fluorescent subunits, conventional wide-field fluorescence light microscopy and digital imaging with a low-noise, cooled charged coupled device (CCD) camera . In FSM, the fraction of fluorescently labelled molecules in the cell, relative to the level of endogenous unlabelled molecules, has to be very low (typically 0.5% or less). Labellled and unlabelled molecular subunits stochastically co-assemble into structures, giving a random and sparse distribution of fluorescent subunits with a ‘speckled’ appearance in high-resolution fluorescence images. The low level of fluorescent subunits reduces background fluorescence. Translation of the fluorescent speckle distribution indicates movement of structures whereas changes in speckle intensity and pattern reveal assembly dynamics and subunit turnover. Keys to successful FSM are the ability to image diffraction-limited regions (~0.25 mm) containing few (2–10) fluorophores, and the capacity to inhibit photo-bleaching which is only possible with sensitive imaging system that includes a low noise/high quantum efficiency camera. Extensive characterisation of actin dynamics using FSM has revealed two spatially, kinetically and kinematically distinct actin networks, with the local expansion of the lamella network being a source of persistent cell protrusion [83, 84]. There is also evidence that the actin network is dynamically coupled to adhesions . A recent study using FSM measured the coupling of focal adhesion proteins to actin filament. Their FSM analysis of the dynamic interactions between matrix adhesion components and F-actin in living cells revealed that there is a hierarchy of motion from fast to slow, from actin-binding proteins to adhesion proteins within matrix adhesions and to integrins . Those FSM-mediated molecular measurements provided considerable knowledge on the mechanism behind matrix adhesion dynamics.
Fluorescence correlation spectroscopy (FCS) and variants
Fluorescence correlation spectroscopy (FCS) analyses concentration fluctuations as a function of time to determine kinetic parameters, molecular associations and concentrations. This technique requires laser excitation of small focal volume and measure fluctuations in fluorescence intensity over many time intervals. Single or cross-correlation analysis (in the case of two different fluorophores) is applied in local areas across a time course to determine rates of diffusion, degree of aggregation, number of fluorescent entities and flow velocities. When two different fluorescent proteins are used, the cross-correlation function provides estimates of their fractional association and rates of co-transport. FCS is used most widely to study molecules in solution; only a few groups have applied FCS to analyse adhesion protein movements in intact living cells. A variant of FCS, image correlation spectroscopy (ICS), was implemented by Wiseman and co-workers  and allowed them to investigate the distribution, dynamics and interactions of α5-integrin, and α-actinin in the context of the formation and disassembly of adhesions during cell migration. Integrins are clustered throughout the cell and in nascent adhesions get 1.4 times more concentrated and 4.5 times more clustered and less mobile than in surrounding regions. Although FCS has a high temporal resolution (microseconds) but low spatial resolution, ICS has a low temporal resolution (seconds) but a high spatial resolution. A new analysis method, termed raster imaging correlation spectroscopy (RICS), can be applied on any confocal microscope  and bridges the timescales of FCS and ICS, as well as providing spatially resolved dynamic information such as the diffusion and binding of paxillin-GFP stably expressed in CHOK1 cells  and of FAK-GFP in MEFs . Another variant of FCS, a general velocity-mapping technique termed spatio-temporal image correlation spectroscopy (STICS) has been described and provided new insights into the protein mobilities within the focal adhesions: while integrins were mostly immobile, paxillin and FAK immobile fractions were equal (74%), and actin was more rapidly diffusing (24%) . Finally, a recent study that combined different fluorescence fluctuation approaches demonstrated that paxillin-GFP shows a heterogeneous dynamic within the cell . In the cytoplasm, paxillin is uniformly distributed and diffuses freely as a monomer. Near adhesions, paxillin binds to protein partners and so its dynamics is reduced. These dynamic were different from assembling to disassembling adhesion regions, even within a single adhesion . The implementation of complementary fluctuation methods will provide new data on the dynamics of protein adhesions during cell migration.
Fluorescence resonance energy transfer (FRET)
A powerful imaging method to study protein–protein interactions in living cells is fluorescence resonance energy transfer (FRET) [93, 94]. FRET is the non-radiative transfer of energy from a donor fluorophore in an excited state to a nearby acceptor fluorophore to allow energy transfer if within only 10 nm. Because this distance is in the range of protein sizes, FRET can also be used to study conformational changes of proteins tagged with a FRET donor and FRET acceptor. The most frequently used FRET methods are sensitised emission, ratio imaging, and acceptor photo-bleaching FRET  but the latter is not appropriate for studying rapid changes of protein interactions over time. The sensitised emission approach detects the emission of the acceptor fluorophore (often cyan fluorescent protein, CFP) while the donor fluorophore (often yellow fluorescent protein, YFP) is excited. Although still widely used, sensitised emission requires careful data processing  and due to signal-to-noise ratio can be poorly sensitive. Cross-talk and bleed through from one fluorophore to another makes the analysis highly dependent on control measurements of cells in which only one of the two fluorophores is present. An alternative approach to determine FRET is acceptor/donor ratio imaging (e.g. YFP/CFP) where both donor and acceptor emission are detected simultaneously when excited at the excitation wavelength of the donor. However, this method can be only applied when donor and acceptor are equally expressed in a cell system which is always the case when using FRET biosensors. In the study of adhesions, a few FRET biosensors have been designed to monitor in live cells the activity of a number of kinases, e.g. Src and FAK [97, 98] and GTPases, e.g. Rho, Rac and Cdc42 [99, 100, 101, 102]. A fourth method to detect FRET is based on the reduced lifetime of excited donor molecules when they are in the proximity of acceptors [103, 104]. This technique is considerably more sensitive and accurate than intensity-based methods, but is slower and requires specific detector and often a pulsed light source such as a two-photon laser. Consequently, this has limited the application of FLIM in live cell studies. However, improvements in microscope design detector technology have reduced the time for data acquisition. Indeed, the group of P. French has developed a time domain optically sectioned FLIM microscope developed for high-speed live cell imaging. This single photon excited system utilises a spinning Nipkow disc microscope and can acquire fluorescence lifetime images of live cells at up to 10 frames per second, permitting high-speed FLIM of cell dynamics and protein interactions with potential for high throughput cell imaging and screening applications . When correctly applied, FRET is a useful tool for investigating the molecular mechanisms that regulate integrin-mediated signalling in migrating cells [93, 97, 102, 106].
Optical imaging towards understanding tumour cell migration and intravasation
Modelling distinct modes of tumour migration
Insights into the organisation of molecular machineries underlying cell adhesion and migration
Wide-field, confocal, confocal reflection microscopy, SHG
Distinguish aspects of cell movement/invasion (collective/individual; mesenchymal/amoeboid). Visualise interactions cell. ECM (in particular, collagen fibers I)
On/in collagen gel
Confocal, confocal reflection, Multiphoton (SHG and FLIM)
Aspects of cell movement in the primary environment. Visualise interaction between tumour cells and tumour environment (ECM, host cells and blood vessels). Visualise intravasation event when blood vessels are counterstained
Imaging adhesions and tumour invasion in 3D culture systems
Imaging migration and adhesions in vivo
Studying adhesion dynamics in migrating cells in an in vitro 3D matrix is already technically challenging. Obviously, it is even more challenging to do so within intact organisms. Suitable in vivo models for both cancer progression and high resolution imaging are necessary.
Zebrafish model for metastasis analysis
Mouse metastasis model for imaging of tumour cell migration
The availability of multi-photon intravital microscopy has allowed researchers to visualise the dynamic behaviour of cancer cells in vivo [119, 120, 129, 130]. Multiphoton microscopy uses longer wavelengths (up to 1,200 nm) that are able to penetrate deeper into tissues and allow us to visualise more than 100 μm deep into the primary tumour. Multi-photon excitation also causes less photo-damage, permits good optical sectioning and 3D resolution  and non-invasive visualisation of the extra-cellular matrix thanks to the second harmonic generation phenomenon [132, 133]. In the past years, intravital imaging has been mainly used to follow up individual or group of cells fluorescently labelled within the primary tumour and study the interaction of moving cells with their microenvironment such as collagen matrix (Fig. 6b) [130, 134]. Recently, a new technical development, the mammary imaging window (MIW) has been shown to be advantageous for studying cell movement and adhesion with high resolution. The use of photo-convertible fluorophores such as Kaede or Dendra2 allows a precise monitoring of cellular movement in vivo no longer just over hours but over days , . Despite all these technological advances, it was still not possible to visualise adhesions in migrating cells in vivo. Just recently, for the first time, a study on E-cadherin dynamics in living animals has been reported . Photo-bleaching and photo-activation was used to compare the mobility of cell adhesion and plasma membrane probes in vitro and in tumours grown in mice and consequently demonstrate critical differences in molecular dynamics in vitro and in vivo.
Future directions and concluding remarks
FRAP/FRET, FRAP/TIRF and FRET/TIRF
An interesting alternative to the FRAP method is the combination of both FRAP and FRET that has already been applied in the nucleus [95, 138]. In the acceptor photo-bleaching FRET methodology, the bleaching of the acceptor results in an increase of the donor intensity. By applying FRAP on cells expressing adhesion proteins tagged with both FRET partners, e.g. CFP and YFP, and simultaneously recording in the bleached region the acceptor recovery and the redistribution of the increased donor signal, it is possible to compare the mobility of the interacting proteins (donor redistribution) relative to the mobility of the total pool of proteins (the YFP recovery as in a conventional FRAP experiment). Of course, it would be even more attractive to apply our previously described FLIP/FRAP methodology together with FRET to understand the complexity of both protein dynamics and interactions. Another very attractive combination of imaging technique is TIRF together with FRAP and/or FRET, a technique that is currently experiencing rapid growth in application [139, 140, 141]. This combination of imaging technologies will ensure an improved insight into adhesion protein dynamics and complexation in migrating cells since TIRF microscopy enormously enhances spatial resolution of fast-moving matrix adhesions.
High through put techniques (2D and 3D) for target identification
High-throughput fluorescence microscopy using fixed assays
The Geiger laboratory published two different screens which provide helpful methodologies and data on cell adhesion and migration: the first used high resolution microscopy to profile the effect of a library of natural extract on cell adhesion , while the second used a modified phagokinetics tracks with MCF7 and identified novel pro-migratory, cancer-associated genes . Very recently, a third screen using high-throughput, high-resolution, microscopy-based assay together with human kinases, phosphatases and adhesome libraries was also performed and it provided a model for the molecular hierarchy of FA formation . Another very elegant study used the traditional wound-healing assay with MCF-10A breast epithelial and screened siRNAs targeting 1,081 human genes encoding phosphatases, kinases and proteins predicted to influence cell migration and adhesion . Extensive validation of all the hits yielded 66 high confidence genes that, when downregulated, either accelerated or impaired migration; 42 of these high confidence genes were not previously associated with motility or adhesion . Although, the results of these screens are very promising and provide new data on cell migration, the analysis was performed with fixed samples.
High-throughput time-lapse fluorescence microscopy
Challenges and potential solutions to increase throughput of imaging techniques
Development of predictive screening assays (→ use of primary or embryonic stem cells (ESCs) instead of easy to culture tumour cell-lines but genetically aberrant)
Development and validation of relevant 3D scaffolds (→ characterise ECM of patient tumour material)
Improve 3D cell culture techniques for automated liquid handling robotics (→ collaboration between academia and pharmaceutical companies)
Automated microinjection of tumour cells in ZF (→ automatic microinjector based on pattern recognition)
Automated filling of the microwells plates with ZF preferably all similarly orientated (→ make use of adapted mould)
Automated image acquisition
Autofocus combined with z-scans for 3D imaging (→ image-based or reflection-based autofocusing)
Pre-optimisation of the acquisition settings (→ autoexposure algorithm to adjust integration time of detectors)
Automated object (cells of matrix adhesion) localisation (→ autoexposure algorithm to adjust integration time of detectors)
Higher throughput kinetic imaging microscopes suitable for automated 3D invasion studies (→ see commercially available kinetic imaging systems such as Incucyte or Cell-IQ )
Higher throughput kinetic imaging microscopes suitable for FRET, FRAP or FCS (→ in the future, intelligent microscope that recognises the object to be visualised)
Storing terabytes of data (→ storage area network (SAN) which has multiterabyte to tens of terabytes capacity; commonly, data on the SAN are backed up on tape as well)
Data management (→ development of databases retrieved [157, Table 2])
Image segmentation (→ depending on imaging quality, choose between region-based, edge-based or region-growing method)
Multiparametric image analysis (→ phenotypic profiling which involves computer vision methods)
Object tracking (→ high time resolution for imaging; adapted tracking algorithms for 3D imaging  and HTS data)
Data mining and modeling
Screening reproducibility and estimators (→ quality standards, e.g. coefficient of variations (CVs) and zscores should not exceed 5% and should be higher than 0.5, respectively)
Significant behaviour changes detection
Automated classification (→ supervised machine learning)
Development of computational models
Perspectives and challenges to drug discovery and clinical treatment of cancer metastas formation
The above systems microscopy approach integrates different imaging techniques and relevant models for all stages of target-based drug discovery. This involves the study of cells, including target discovery, drug screening in cell-based assays, early safety evaluation, mode-of-action studies and in vivo studies to monitor cell fate. A prerequisite for reaching conclusions with low failure rates in drug discovery is the improvement of throughput that provides reliable data. The level of complexity comprising both changes in cellular morphology and macromolecular subcellular localisation and expression level makes HTS ideal among current approaches used for drug screening [155, 156]. Automated microscopy platforms that run in medium- or high-throughput mode to screen thousands of cDNAs, siRNAs, aptamers, proteins and antibodies using endpoints 2D migration and 3D invasion would benefit target discovery and validation [122, 148]. For the lead identification and optimisation, endpoint and time-lapse microscopy combined with multiparametric kinetic analysis incorporating both cell invasion assays and assessment of target activity with FAST/FLIM would be the next step . However, as previously mentioned, kinetic HTS is still a technical and intellectual challenge which requires adapted fluorescent labelling techniques, automated ‘intelligent’ microscopes and well-developed software platforms [157, 158] (Table 3). In the later stages of the drug discovery trajectory, intravital imaging of tumour invasion in various relevant in vivo models can be used to evaluate the response of the drug in development and assess its activity when combined with FRET/FLIM. In the future, the combination of other microscopy techniques such as FRAP or FCS will help to obtain additional mechanistic information so that the drug development process can be more efficient. Nonetheless, a number of significant challenges remain related to the use of more biology- and disease-relevant predictive cell systems and the development of biosensors and bioreporters to measure and manipulate cellular events. Finally, the identification of relevant biomarkers specific to distinct invasion mechanisms will facilitate the development of clinical and preclinical studies and help administrate appropriate anti-invasive therapies to cancer patients .
Here, we presented a systems microscopy approach which integrates different imaging techniques developed over the past years to study adhesion signalling in relation to tumor cell migration and cancer metastasis . In Table 1, we summarised the different techniques to study protein dynamics and interactions in matrix adhesions. In Table 2, an overview of the techniques used to model the various modes of tumour migration was given. In Table 3, we recapitulated the challenges and potential solutions to increase the throughput of imaging and multiparametric analysis. Understanding the movement of single cells and cells in tissue requires the analysis of complex processes under normal and perturbed conditions. Ideally, one would like to quantitatively determine movement and shape changes of cells and correlate these with the spatio-temporal dynamics of the cytoskeletal elements and the extra- and intracellular signalling pathways controlling these behaviours in isolated cells as well as in cells in tissue. Imaging together with multiparametric analysis and data modelling is the most appropriate method to reach that ultimate goal. Optimisation of current techniques and systems that combine diverse techniques will improve both our spatial and temporal resolutions of the role of matrix adhesions in migrating cells. Also, more and more bioinformatics tools will be generated for rapid and detailed image analysis and data processing. The additional knowledge obtained will hopefully provide insights into the molecular mechanisms behind tumour cell migration and help in developing new anticancer therapies.
The authors would like to thank Leo Price for technical support for the 3D invasion assays. This work was financially supported by grants from the Dutch Cancer Society (UL 2007-3860), the EU FP7 Health Program Metafight (Grant agreement no.201862) and ZF Cancer (Grant agreement no.201862), and the Netherlands Organization for Scientific Research (902-21-229 and 911-02-022).
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- 23.Zamir E, Geiger B, Kam Z (2008) Quantitative multicolor compositional imaging resolves molecular domains in cell-matrix adhesions. PLoS ONE. doi: 10.1371/journal.pone.0001901
- 38.Xue C, Wyckoff J, Liang F, Sidani M, Violini S, Tsai KL, Zhang ZY, Sahai E, Condeelis J, Segall JE (2006) Epidermal growth factor receptor overexpression results in increased tumor cell motility in vivo coordinately with enhanced intravasation and metastasis. Cancer Res 66:192–197PubMedCrossRefGoogle Scholar
- 48.Yafeng M, van de Water B, Le Dévédec SE (2009) Focal Adhesion Kinase inhibition as a potential strategy for anticancer therapies. Drugs Future 34:477–484Google Scholar
- 78.Wolfenson H, Lubelski A, Regev T, Klafter J, Henis, YI, Geiger B (2009) A role for the juxtamembrane cytoplasm in the molecular dynamics of focal adhesions. PLoS ONE. doi: 10.1371/journal.pone.0004304
- 105.Grant DM, McGinty J, McGhee EJ, Bunney TD, Owen DM, Talbot CB, Zhang W, Kumar S, Munro I, Lanigan PM, Kennedy GT, Dunsby C, Magee AI, Courtney P, Katan M, Neil MA, French PM (2007) High speed optically sectioned fluorescence lifetime imaging permits study of live cell signaling events. Opt Express 15:15656–15673PubMedCrossRefGoogle Scholar
- 133.Kedrin D, Wyckoff J, Sahai E, Condeelis J, Segall JE (2007) Imaging tumor cell movement in vivo. Curr Protoc Cell Biol 19.7.1–19.7.17Google Scholar
- 136.Gligorijevic B, Kedrin D, Segall JE, Condeelis J, van Rheenen J (2009) Dendra2 photoswitching through the Mammary Imaging Window. J Vis Exp. doi: 10.3791/1278
- 150.Naffar-Abu-Amara S, Shay T, Galun M, Cohen N, Isakoff SJ, Kam Z, Geiger B (2008) Identification of novel pro-migratory, cancer-associated genes using quantitative, microscopy-based screening. PLoS ONE. doi: 10.1371/journal.pone.0001457
- 153.Neumann B, Walter T, Heriche JK, Bulkescher J, Erfle H, Conrad C, Rogers P, Poser I, Held M, Liebel U, Cetin C, Sieckmann F, Pau G, Kabbe R, Wunsche A, Satagopam V, Schmitz MH, Chapuis C, Gerlich DW, Schneider R, Eils R, Huber W, Peters JM, Hyman AA, Durbin R, Pepperkok R, Ellenberg J (2010) Phenotypic profiling of the human genome by time-lapse microscopy reveals cell division genes. Nature 464:721–727PubMedCrossRefGoogle Scholar
- 154.Talbot CB, McGinty J, Grant DM, McGhee EJ, Owen DM, Zhang W, Bunney TD, Munro I, Isherwood B, Eagle R, Hargreaves A, Dunsby C, Neil MA, French PM (2008) High speed unsupervised fluorescence lifetime imaging confocal multiwell plate reader for high content analysis. J Biophotonics 1:514–521PubMedCrossRefGoogle Scholar