Für neue Kunden:
Für bereits registrierte Kunden:
Doktorarbeit / Dissertation, 2014
142 Seiten, Note: 1,0
Table of figures
Table of tables
Chapter 1: Introduction
1 Introduction to drug development in oncology
1.1 Public and privately funded efforts in cancer research
1.2 Early drug discovery process
1.3 Screening approaches for drug discovery
1.4 Phenotypic High-Content Screening (HCS)
1.5 3D cell culture model for oncology
1.6 Application of 3D cell cultures in oncology
1.7 Targeting dormant cancer cells with 3D cell culture models
1.8 Combination of High-Content Analysis (HCA) and 3D cell cultures to identify sensitizers of dormant tumor regions
2 Introduction to 3D cell co-culture models for invasion processes
2.1 Invasion processes in cancer and other diseases
2.2 Invasion processes in idiopathic pulmonary fibrosis
2.3 Monitoring tissue interactions in 3D cell co-cultures
2.4 Aim of this work
Chapter 2: Materials and methods
3.1.1 Laboratory equipment
3.1.2 Chemicals and solutions
3.1.3 Fluorescent dyes and kits
3.1.4 Molecular biology kits, antibodies and enzymes
3.1.6 Other consumables
3.1.7 Cell lines
3.1.8 Applied Biosystems TaqMan® Gene Expression Assays
3.2.1 Preparation of Agarose coated imaging plates
3.2.2 Cell culture, drug treatment, and dead cell staining procedure
3.2.3 Fibroblast staining
3.2.4 Image acquisition
3.2.5 Light sheet based image acquisition (on mDSLM microscopy system)
3.2.6 2D cytotoxicity assay
3.2.7 Cell Titer Glo Assay
3.2.8 Immunocytochemistry (whole spheroids)
3.2.10 Mesoscale AMPK activation
3.2.11 Quantitative Reverse Transcription PCR
3.2.12 Seahorse measurements
3.2.13 Metabolomics with Nuclear Magnetic Resonance (NMR) spectroscopy
3.2.14 Statistical analysis and hit evaluation
Chapter 3: Results
4 Results - 3D cell cultures for oncology
4.1 A 3D cell culture model for high-throughput screening
4.2 Visualization of multicellular tumor spheroids (MCTS) for High-Content Analysis
4.3 High-Content Analysis on 3D data sets
4.4 Fast and accurate automated image analysis workflow for compound profiles in oncology
4.5 Evaluation and characterization of different cell lines for 3D cell culture setup
4.6 T47D breast cancer multicellular tumor spheroids (MCTS) as a model for cancer dormancy and drug resistance
4.7 Dormant multicellular tumor spheroid core regions are resistant to cytostatic- based chemotherapy
4.8 High-Content Screen to identify compounds that specifically kill dormant MCTS cells
4.9 Identified compounds act by inhibition of respiratory chain complexes
4.10 Further mechanistic assays to decipher effect of respiratory chain inhibitors
4.11 Benefits of combination treatment of respiratory chain inhibitors with chemotherapeutic agents in vitro
4.12 Conclusion - Use of 3D cell cultures to model tumor dormancy in vitro
5 Results - heterotypic 3D cell co-cultures
5.1 3D cell co-culture models to reproduce invasive processes
5.2 3D cell co-culture model for fibroblast invasion in pulmonary fibrosis and adaptation of imaging and analysis protocols
5.3 High-Content Screen for inhibitors/modulators of lung fibroblast invasion
5.4 Hit classification
5.5 Prostaglandin EP receptor and Rho-associated protein kinase (ROCK) pathway inhibitors could be classified as target pathways
5.6 Conclusion - Use of 3D cell co-cultures as an in vitro model for fibrosis
Chapter 4: Discussion
6 Discussion - 3D cell cultures for drug discovery
6.1 Cell cultures and their use in oncology
6.2 Metabolic adaptations in multicellular tumor spheroids
6.3 Multicellular tumor spheroids as a model for tumor dormancy
6.4 Identification of respiratory chain inhibitors that target dormant MCTS regions
6.5 Proposed model of glucose sensitivity and energy homeostasis in dormant MCTS regions
6.6 Translation into in vivo activity and clinical strategies for combinational treatments with respiratory chain inhibitors
6.7 Conclusions and Outlook
7 Discussion - 3D cell cultures for invasion processes
7.1 3D cell co-cultures as a model for idiopathic pulmonary fibrosis (IPF)
7.2 High-Content Screening for the identification of compounds inhibiting fibrotic processes prevalent in idiopathic pulmonary fibrosis
7.3 Targeting prostaglandin and Rho-associated protein kinase (ROCK) signaling in IPF treatment
7.4 Relevance of identified compounds and translation into in vivo activity
7.5 Conclusions and Outlook
Chapter 5: Appendix
9 Supplementary data
Erklärung gemäß §5 Absatz 6 der Promotionsordnung der Technischen Universität Berlin
Hiermit erkläre ich, dass ich die vorgelegte Dissertation selbstständig angefertigt und nur die angegebenen Quellen und Hilfsmittel verwendet habe. Wörtlich oder inhaltliche übernommene Stellen sind als solche gekennzeichnet.
Carsten Wenzel Berlin,
Förderung: Diese Arbeit wurde vom Bundesministerium für Bildung und Forschung unterstützt (BMBF Kennzeichen 13N11115 (ProMEBS)).
Die Anfertigung der Arbeit erfolgte in der Arbeitsgruppe von:
Dr. Patrick Steigemann
Lead Discovery Berlin / High-Content Analysis Bayer Pharma AG
Die Betreuung wurde von Dr. Steigemann übernommen.
Prof. Dr. Peter Neubauer
Prof. Dr. Roland Lauster
Prof. Dr. Juri Rappsilber
Dr. Patrick Steigemann
Tag der wissenschaftlichen Aussprache:
3D microtissue for drug discovery in oncology
Established a novel method for 3D cell culture based High-Content Screening Validated 3D spheroids as an in vitro model for tumor cell dormancy
Produced a first report on a High-Content Screen for the identification of compounds that specifically target cells in dormant tumor spheroid regions
Identified respiratory chain inhibitors to specifically induce cell death in dormant tumor spheroid regions and to enhance cytostatic-based therapy in vitro
Showed that glucose is a major determinant that influences efficacy of respiratory chain inhibitors
3D microtissue to monitor fibrotic processes in pulmonary fibrosis:
Established 3D co-culture assay for tissue interaction to monitor fibroblast invasion and migration
Discovered new compounds by High-Content Screening that prevent or modulate invasion in idiopathic pulmonary fibrosis (IPF)
Identified compounds relevant for IPF treatment in vivo
Despite all efforts to discover novel therapeutic options to treat cancer the disease remains devastating causing more than 14 million deaths in 2012 - and the trend is rising. Some parts of the preclinical drug discovery process rely on cell culture models to reproduce the pathophysiological features of cancer in in vitro experiments. These cellular models are then used in screening campaigns that aim to identify substances such as small molecules, peptides, or interfering RNA that alter the phenotype of a cancer cell in a desired manner.
One of the main properties of cancer cells is their uncontrolled cell division. Accordingly, the cell cycle is a major target for chemotherapy, but resistance to chemotherapy frequently causes treatment failure in patients with advanced and inoperable cancer. As the distance from supplying blood vessels increases, oxygen and nutrient concentrations decrease and cancer cells react by halting cell cycle progression and entering a dormant state. As cytostatic drugs mainly target proliferating cells, cancer cell dormancy is considered a major resistance mechanism to this class of anti-cancer drugs. Therefore, substances that target cancer cells in poorly vascularized tumor regions have the potential to enhance cytostatic-based chemotherapy in solid tumors.
Multicellular tumor spheroids (MCTS) allow for three-dimensional growth conditions, thus reproducing several parameters of the tumor microenvironment, including oxygen and nutrient gradients as well as the development of dormant tumor regions. Here, the evaluation of a 3D cell culture high-content screening system led to the identification of nine substances from two commercially available drug libraries that specifically target cells in inner MCTS core regions. Subsequently, the mode of action of the identified compounds could be identified as respiratory chain inhibitors. Ultimately, benefits in combinational treatment with commonly used cytostatics in MCTS were proven. The data suggest a rationale to find and evaluate new substances that target the altered metabolism of tumor cells in dormant tumor regions to enhance cytostatic- based therapies.
In a second approach the combination of different cell types in a 3D cell co-culture model was evaluated, which allows monitoring of invasion processes in different malignancies. Among them idiopathic pulmonary fibrosis (IPF) currently affects 50.000 people per year with high mortality rates. The disease is marked by increased scarring and fibroblast invasion into normal lung tissue with currently no treatment options available. The developed 3D invasion model mimics invasion processes of fibroblasts in tissue surrogates in vitro. This screening-compatible assay allowed for the identification of compounds that inhibit or modulate uncontrolled fibroblast invasion into tissue. Finally, 16 compounds from a screening library could be identified that strongly decreased fibroblast invasion. Two target pathways were determined that included substances currently under investigation in animal models or even used in clinical studies for the treatment of IPF, proving the validity of the developed in vitro 3D co-culture system.
In summary, this work provided novel 3D cell (co-) culture approaches for indications for which no appropriate screening compatible in vitro assays in preclinical drug discovery yet exist. New compounds and targets in different disease models could be identified by the developed assay. Finally, by providing increased physiological relevance the in vitro assays help to predict compound efficacy with higher probability of in vivo activity and can therefore reduce attrition rates in preclinical drug development.
Trotz großer Fortschritte in der Krebsforschung und zahlreicher neuer Therapieansätze bleibt Krebs eine oftmals tödliche Erkrankung, die mehr als 14 Millionen Menschen im Jahr das Leben kostet. Ein großer Teil der präklinischen Forschung findet mit in vitro Zellkulturen statt, welche einige Aspekte der Krankheit im Labor abbilden können. Diese zellbasierten Modelle werden genutzt, um aus großen Substanz-Bibliotheken wirksame Substanzen zu identifizieren, die das Überleben oder Wachstum der Krebszellen verhindern können. Eine Haupteigenschaft von Krebszellen ist ihre gesteigerte Zellteilungsrate. Substanzen, die schnell teilende Krebszellen töten werden daher oft in der Chemotherapie verwendet. Auftretende Resistenzen führen dabei jedoch häufig zum Scheitern der Therapie und der Tumor kommt nach anfänglich erfolgreicher Behandlung zurück. Mit steigender Distanz zu versorgenden Blutgefäßen nimmt der Mangel an Sauerstoff und Nährstoffen im Tumor rapide ab. In der Folge nimmt die Zellteilung der Tumorzellen ab und eine ruhende Zellpopulation im Tumor entsteht. Da zellteilungshemmende Zytostatika vornehmlich schnell wachsende Zellen töten, wird angenommen, dass die ruhende Population möglicherweise verschont wird und zum Rückfall der Erkrankung führt. Die Entdeckung von Substanzen, die schlecht versorgte, ruhende Zellen im Tumor töten können, ist eine vielversprechende Möglichkeit, um die Wirksamkeit von zytostatika-basierter Chemotherapie zu erhöhen. Multizelluläre Tumorsphäroide (MZTS) ermöglichen ein 3-dimensionales Wachstum von Krebszellen und können damit einige Parameter der Tumorumgebung in vitro abbilden. Durch Nährstoff- und Sauerstoffgradienten innerhalb der MZTS, wie sie auch im in vivo Tumor vorhanden sind, entsteht eine ruhende Zellpopulation. Die Evaluierung dieser 3D Zellkulturen mit automatisierter Mikroskopie und Bildanalyse (sog. High-Content Analysis oder High-Content Screening) erlaubte es 9 Substanzen in einem Pilotscreen (aus 1120 Substanzen) zu identifizieren, die spezifisch Zellen in der ruhenden Population von MZTS abtöten. Zusätzlich konnte der Wirkmechanismus der Substanzen als Inhibitoren der Atmungskette charakterisiert werden und eine verbesserte Wirksamkeit in der Kombination mit Zytostatika konnte in vitro bestätigt werden. Die Daten legen nahe, dass die Nutzung von Atmungsketteninhibitoren in Kombination mit Zytostatika einen sinnvollen therapeutischen Ansatzpunkt liefert.
In einem zweiten Ansatz wurde ein 3D Ko-Kulturmodell zur Abbildung von fibrotischen Prozessen evaluiert, wie sie in Krankheitsbildern wie der idiopathischen pulmonalen Fibrose (IPF) auftreten. IPF wird jährlich bei 50.000 Menschen diagnostiziert und hat eine hohe Mortalitätsrate. Das Einwachsen von Fibroblasten in Lungengewebe lässt das Gewebe vernarben und reduziert die Sauerstoffaufnahme drastisch. Bisher existieren keine erfolgreichen Therapiemöglichkeiten. Das entwickelte 3D Ko-Kulturmodell imitiert den Einwanderungsprozess von Fibroblasten in 3D Sphäroidgewebe in vitro. Dieses hochdurchsatz-fähige Screening-Modell hat es ermöglicht 16 Substanzen zu identifizieren, welche das Einwandern von Fibroblasten verhindern können. Dabei konnten 2 Signalwege identifiziert werden, die ebenfalls für die Entstehung von IPF in vivo relevant sind, was die Validität dieses 3D Ko-Kulturmodells bestätigen konnte.
Zusammengefasst bildet diese Arbeit damit die Grundlage komplexe 3D Zellkulturen in der frühen Wirkstofffindung im Hochdurchsatz-Screening anzuwenden, um für verschiedene Erkrankungen neue Wirkstoffkandidaten zu entdecken. Neue Substanzen und Signalwege konnten dabei identifiziert werden. Durch eine physiologischere Abbildung der Erkrankungen ist die Wahrscheinlichkeit höher, dass die Substanzen auch in vivo aktiv sind und damit bei der Medikamentenentwicklung die hohe Fehlerrate in der präklinischen Entwicklung verringern können.
Figure 1: Early preclinical stages in drug discovery
Figure 2: Comparison of 3D cell culture techniques
Figure 3: Hypoxia and proliferation gradients in solid tumors
Figure 4: Various gradients exist in multicellular tumor spheroids (MCTS)
Figure 5: Schematic comparison of normal lung architecture and impairment by idiopathic pulmonary fibrosis (IPF)
Figure 6: Workflow of semi-automated agarose coating for microplates
Figure 7: Integration of 3D cell cultures in a High-Content Screen (HCS) setup
Figure 8: Custom-build mask generation and artifact removal allows identification of localized phenotypes
Figure 9: Pivotable incubation chamber of a custom-build light sheet microscope
Figure 10: Comparison of non-adherent cell culture microplates and agarose-coated imaging microplates
Figure 11: Growth characteristics compared between ultra-low-attachment plates and agarose plates
Figure 12: Agarose-coating procedure was established for 96-well, 384-well and 1536-well microplates
Figure 13: Summary of High-Content Analysis (HCA) readouts
Figure 14: Comparison of marker strategies for the analysis of MCTS in a High-Content Analysis setup
Figure 15: Problems in optical sectioning of biological 3D samples with confocal or light sheet based microscopy systems
Figure 16: Workflow and preparation of 3D spheroids for chemical clearing
Figure 17: Chemical clearing allows imaging 3D MCTS with diameters of up to 800 μm
Figure 18: 3D projection of cleared T47D spheroids and subsequent 3D data analysis
Figure 19: Evolution of acquisition and analysis protocol to decrease processing and analysis time
Figure 20: 2D image acquisition and 2D image analysis workflow for fast and accurate compound profiling on 3D spheroids
Figure 21: Impairment of actin dynamics, but not tubulin impairment, inhibits spheroid formation in a time-lapse experiment
Figure 22: Optimal seeding number has to be evaluated for every cell line to guarantee a sufficient 3D microenvironment while preserving the ability to quantify drug effects in a high dynamic range
Figure 23: Comparison of hypoxia and proliferation in 8 different human and murine cancer cell lines
Figure 24: Different metabolic adaptations between 2D cell cultures and 3D MCTS is revealed by gene expression analysis and NMR spectroscopy of metabolites
Figure 25: Multicellular tumor spheroids mimic several parameters of the tumor microenvironment
Figure 26: Cytostatic-resistant spheroid core remains after treatment with commonly used chemotherapeutical agents
Figure 27: High-Content Screening (HCS) on multicellular tumor spheroids (MCTS)
Figure 28: Scatter plot visualization of screened compounds
Figure 29: Exemplary visualization of dose-response curve of diphenyleneiodonium
Figure 30: Identified compounds that induce cell death in core regions act as respiratory chain inhibitor
Figure 31: Death in MCTS core regions after respiratory chain inhibition in different cancer cell lines
Figure 32: AMPK and caspase activation under respiratory chain inhibition
Figure 33: Response to Compound C (Dorsomorphin) (an AMPK inhibitor) suggests involvement of AMPK in energy homeostasis
Figure 34: Glucose concentration in the extracellular environment is a major determinant for the sensitivity of MCTS cells against inhibitors of the respiratory chain
Figure 35: Combination therapy with cytostatics increased efficacy of respiratory chain inhibitors
Figure 36: Evaluation of live cell dye Vybrant DiD that can monitor invasion processes of fibroblasts in spheroid tissue surrogates
Figure 37: Cryosections show distribution of fibroblasts in T47 tissue surrogates which is traceable by automated image analysis on fluorescently labeled MRC-5 fibroblasts
Figure 38: Strategy to identify compounds that inhibit invasion into 3D spheroids
Figure 39: Fibroblast invasion assay with 3D cell co-cultures
Figure 40: Dose-response curve of the prostaglandin E2 receptor specific hit butaprost
Figure 41: Prostaglandin receptor subtype EP2 mediated inhibition of lung fibroblast invasion
Figure 42: Stained cryosections showing rho kinase inhibitors decrease invasion of fibroblasts
Figure 43: Schematic drawing of gradients present in multicellular tumor spheroids (MCTS)
Figure 44: Schematic drawing of metabolic adaptations under OXPHOS inhibition
Figure 45: Schematic drawing of induction of cell death in inner spheroid core regions involving AMPK and intra-spheroidal glucose gradient
Figure 46: Possible clinical strategies for combinational treatments with respiratory chain inhibitors
Table 1: List of cell lines tested for their capacity to form spheroids
Table 2: Stress and hypoxia-regulated genes induced under oxygen or nutrient depletion with corresponding proteins and their respective function
Table 3: Hit list and hit expansion of respiratory chain inhibitors that induce cell death in inner core regions of MCTS. AC50: Concentration of the compound eliciting 50% of the maximal cell death in inner core regions (54)
Table 4: Fibroblasts cell lines evaluated to reproduce invasion in idiopathic pulmonary fibrosis (IPF). IPF-derived and skin-derived cell lines showed no or only weak invasiveness into T47D tissue surrogates (red) rendering them not suitable to mimic pathophysiological invasion processes. Normal and fetal lung fibroblasts were suitable to mimic invasion processes into tissue surrogates (green)
Table 5: Hitlist of compounds that inhibit invasion of MRC-5 fibroblasts into T47D tissue surrogates including hit expansion, (compounds marked with * = hit expansion). IC50 value is defined as the concentration of the compound required to achieve half maximal inhibition of fibroblast invasion into tissue surrogates
Immense financial, personal and scientific investments in almost all areas of cancer research have advanced our understanding of the molecular processes and treatment options in various types of cancer. This has led to the development of new treatment strategies like targeted anti-cancer therapies that interfere with patient-specific signaling oncoproteins. They provide reasonable benefits for patient groups and proved to show increased specificity while reducing side effects (1). However, cancer often remains an incurable disease and mortality remains high due to the emergence of therapy-resistant tumors that escape or survive treatment (2,3). As cancer is mainly an age-related disease it is expected that annual cancer cases will rise from 14 million in 2012 to 22 million within the next 20 years (4) corresponding with increased life expectancy. These prospects show the persistent high medical need which has been early addressed by the establishment of large public research funds and institutes stretching back to the end of the 19th century, for example the Imperial Cancer Research Fund (UK) founded in 1902 or the Memorial Sloan Kettering Cancer Center (US) in 1884 (5). Hand in hand with basic research, pharmaceutical companies strive to translate basic discoveries about the molecular basis of cancer into clinical applications.
Historically, new medicines have been discovered by observing phenotypic changes in biological systems such as animals or cells after treatment with plant extracts. Subsequent purification of the extracts enabled further characterization of the novel drug candidate along with attempts to unveil its molecular mode of action (MMOA) responsible for the pharmacological effect (6). This phenotype-based drug discovery was nearly completely replaced by target-based drug discovery since the beginning of the genomics era in the 1990’s. The increased knowledge of proteins that are relevant for a pathophysiological disease state led to a focus on protein-based drug targets (7). This target-based drug discovery facilitated a more rational drug development process, which was able to rapidly identify molecular binding partners that modify target activity and therefore provide a rational basis for new medicines (6) (depicted in Figure 1). These binding partners can be chemically engineered small-molecules or “biologics”, such as monoclonal antibodies. Biochemical or simple, cell-based assays use compound libraries in an attempt to inhibit the enzymatic activity or binding properties of a purified target protein (8).
The resulting drug candidates can then be subsequently improved for potency, selectivity, and pharmacokinetic parameters before they enter clinical trials and end in clinical studies.
illustration not visible in this excerpt
Figure 1: Early preclinical stages in drug discovery. The diagram illustrates the early phase of drug discovery, in which the aim is to identify target and lead molecules. In the phenotype-based approach, lead molecules are obtained first, followed by target deconvolution to identify the molecular targets that underlie the observed phenotypic effects. In the target-based approach, molecular targets are identified and validated before lead discovery starts; assays and screens are then used to find a lead (9). These compounds are subsequently improved for potency, selectivity, or pharmacokinetic parameters before they can enter clinical trials. Adapted from Georg C. Terstappen, Christina Schlüpen, Roberto Raggiaschi & Giovanni Gaviraghi, Nature Reviews Drug Discovery 6, 891-903 (November 2007) (10).
However, these efforts face major challenges with the rising cost of research and development activities, decreasing health care budgets, and finally decreasing drug approval rates. These challenges arise in part from the high attrition rate in preclinical and clinical phases of the drug development process (11). Considering these drawbacks, better preclinical strategies are needed to help improve the success rate of the drug discovery process.
Recent studies (12,13) revealed that phenotypic-based drug discovery is a successful strategy for the identification of new medicines in comparison to target-based drug discovery. The reason for the lack of clinical success in the target-centric approach could be due to the fact that although the molecular mode of action (MMOA) of the drug candidate is well described little is known about the capability of the target to achieve a desired biological effect in vivo. In addition, biological complexity and compensatory mechanisms are not easily addressable in target-based assays.
On the contrary, in phenotypic-based drug discovery no target hypothesis or knowledge of the MMOA is needed to discover new drug candidates. Such screens can aid in the identification of molecules with multiple targets or that modify a disease phenotype by acting on a previously undescribed target which might not be found in a focused target-based screening campaign (14). Therefore, the rationale for the success of phenotypic assays is the unbiased assumption of an involved MMOA and the potential for direct readout of a desired effect in a living system. However, the lack of understanding of the MMOA may slow progression of the drug candidate because subsequent studies will be needed for target deconvolution. Significant progress has been made in target deconvolution to overcome this obstacle, for example by using RNA interference that will aid in the identification of the drug target and subsequently foster the optimization of a drug candidate (10,15).
Hence, the resurgence of phenotypic cell-based screening assays at early preclinical stages has become an integral component in drug discovery (13).
Considerable progress has been made in recent years to automate phenotypic assays and increase their throughput (16-18). A technology that enables a multiparametric readout in high-throughput phenotypic assays is High-Content Analysis (HCA). HCA is the automation of fluorescence microscopy whereby the manual interpretation of images is replaced by computer algorithms (fully or semi- automated image analysis routines) that can quantify staining profiles in different areas of the image (9). To fully employ multiparametric HCA, diverse marker strategies are accessible for a broad range of cellular structures and other biological readouts such as apoptosis and invasion to quantify a desired phenotype. Combined with robotic handling and automated microscopy systems the technique enables for High-Content Screening (HCS) of large compound libraries. Phenotypic cell-based assays provide more information about the interaction of the drug candidate with its binding partner (one target or even multiple targets) in a complex cellular environment. Thus, they show a higher probability for the translation of results to an in vivo situation (19). To further increase the significance of discoveries from phenotypic screens, cell-based assays have to reflect physiological disease conditions as closely as possible, like growth rate or oxygen supply. In an attempt to enhance phenotypic assays, 3D cell cultures have come into focus because they can reproduce the physiological conditions and morphology of tumors in vitro.
3D cell culture models have gained interest in oncology due to their potential to mimic the complex three dimensional organization of tumor tissue in vitro (20,21). Furthermore, they are able to mimic some physiological constraints like limitations in oxygen or nutrient supplies (22-25). Moreover, direct cell-to-cell contact under 3D culture conditions reflects physiological organization in tissue or tumors. This is also important for cellular behavior and crucial for their growth rate, invasiveness, and response to anti-cancer therapies (26,27).
In general, two main principles for the generation of 3D cell cultures exist: scaffold-based and scaffold-free systems (Figure 2):
Scaffold-based systems | Scaffold-based systems provide structures in which the cells are able to grow in 3D dimensions and interact with each other and/or the scaffold. Many of the scaffolds are based on hydrogels like collagen, other cross-linked biopolymers, or collagen-rich mixtures (e.g. BD Matrigel/basement membrane). These gels are termed natural scaffolds and are able to mimic the natural extracellular matrix (ECM) and allow cells to grow in 3D structures within the gels (28). Synthetic scaffolds use adhesive biomaterial to enforce 3D growth. The techniques vary from polystyrene or ceramics to more advanced electro spinning fiber technologies. These techniques help the cells to attach in 3-dimensions by providing a solid scaffold (29).
Scaffold-free systems | The most commonly used scaffold-free technique makes use of the self- organizing ability of cancer cells to build multicellular tumor spheroids (MCTS). They produce their own ECM and start to build 3D cellular networks. This is achieved by limiting cell adhesion to traditional cell culture lab ware by different techniques. Hanging-drop cultures are widely used, in which cells are seeded upside down in a hanging drop of liquid media. This prevents adhesion and the cells accumulate on the liquid-air surface and start to form compact 3D MCTS structures (30).
Alternatively, nano-imprinted surfaces (e.g. Scivax plates), polysaccharides (e.g. agarose), or hydrogel coated (ultra low attachment) microplates efficiently prevent cell adhesion and, given the right shape, foster cellular concentration and condensation in 3D aggregates (31). 3D structures that are covered (or overlayed) by media are also termed liquid overlay techniques (LOT). These LOT-based microplates yield sufficient numbers of reproducible and compact 3D MCTS for screenings assays. Advantages and disadvantages of these techniques are strongly dependent on the context in which the MCTS are used. The possibilities and limitations of the available techniques are thoroughly discussed in the literature (20,21,28,30-36).
illustration not visible in this excerpt
Figure 2: Comparison of 3D cell culture techniques. (A) Natural hydrogels or (B) synthetic scaffolds support cancer cells to interact and grow in 3D networks. (C) Techniques that support self-assembly of multicellular tumor spheroids (MCTS). Cancer cells growing as MCTS build 3D structures by producing own extracellular matrix and direct cell-cell contacts. Image copyrights see Supplementary data 9.1.
As almost all scaffold-based techniques somehow interfere with microscopic readouts and therefore High-Content Analysis, the scaffolds have to be removed prior to imaging. In addition, most of the scaffold-based 3D cell culture systems are not yet available in microplate formats, which is a prerequisite for image-based phenotypic High-Content Screening (HCS) applications (32,37). In contrast, scaffold-free systems were adapted to produce compact and reproducible spheroids from single cell suspensions that would allow of direct use in imaging setups.
One of the main properties of cancer cells is sustained proliferative growth. Accordingly, the cell cycle is a major target for chemotherapy. Cytostatic drugs show strong anti-cancer efficacy in conventional in vitro assays. However, findings from 2D cell culture based experiments can only be partially translated to experimental outcomes in vivo and resistance to chemotherapy is still a frequent cause for treatment failure in patients with advanced and inoperable cancer (38). Several factors confer resistance to standard treatment regimens including, but not limited to: pharmacokinetic properties, genetic heterogeneity, and drug clearance by cancer cells (39-42).
As commonly used cytostatics mainly target proliferating cells, tumor cell dormancy could be a factor in the limited response to these compounds (43-45). Tumor cell dormancy is influenced by regional differences in oxygen and nutrient supply within the neoplastic tissue, depending on the amount and quality of (neo-) vascularization (i.e. the distance from supplying blood vessels). As tumor growth requires high amounts of energy and nutrients, tumor cell proliferation is therefore mainly restricted to regions adjacent to blood vessels and human tumor tissue can show relatively low proliferative indices in poorly perfused areas (46).
Cancer tissue can be subdivided, depending on vascularization, into well-supplied, proliferating cancer cells in the vicinity of blood vessels and mostly dormant cells in poorly vascularized tumor regions (41) (Figure 3). Cells distinct from blood vessels show a variety of stress responses that promote survival pathways and the reorganization of metabolic pathways that enable them to survive hypoxic conditions. It has been reported that these hypoxic cells confer both radiotherapy and chemotherapy resistance (47-49). A strategy to limit oxygen and nutrients supply in the developing tumor targets new blood vessels by blockade of the vascular endothelial growth factor (VEGF) pathway. However, side-effects and development of resistance limit the therapeutic use of anti-angiogenesis therapy (50,51).
In most cases cancer cells under hypoxic conditions respond by stopping proliferation and, later on, using metabolic adaptations to ensure survival. Consequently, tumor dormancy develops (44) and a dormant cancer cell population could potentially lead to disease relapse after cytostatic-based chemotherapy. Indeed, limited clinical efficacy of recent cancer therapies have been shown to be linked to hypoxic regions in solid tumors (49,52,53).
Therefore, it is of great clinical interest to target dormant cancer cells in order to enhance cytostaticbased chemotherapy or anti-angiogenic therapy (54)..
illustration not visible in this excerpt
Figure 3: Hypoxia and proliferation gradients in solid tumors. (A) Layers of cells surrounding a blood vessel in a xenograft of a cervix cancer. (B) Cords of cells surrounding a blood vessel in a xenograft of colon cancer. More cells are proliferating (bromodeoxyuridine-labelled black cells) close to the blood vessel. Green staining indicates hypoxic regions identified by pimonidazole staining. Proliferation occurs primarily close to blood vessels. Endothelial cells are colored blue.
(C) A diagrammatic representation illustrating the gradients in oxygen, nutrients and energy, and drug concentration. Figure and legend adapted from Minchinton et al. (41).
Despite the potential role of dormant cells in limiting the effectiveness of cancer treatments, few efforts have been made to specifically target this tumor cell population (43,55-57). In part, this reflects the lack of appropriate, screening-compatible in vitro models that are able to simulate the metabolic microenvironment in tumors (54).
3D cancer cell culture models have the potential to mimic the complex three-dimensional organization of tumor tissue in vivo. Similar to native tumor tissue, cells cultured as multicellular tumor spheroids (MCTS) show strong proliferation gradients that reflect distribution gradients of oxygen, nutrients and energy, as well as the accumulation of metabolites from outer to inner spheroid regions (Figure 4) (32). It also has been shown that the 3D organization of cells shows significant changes in gene expression, morphology and response to anti-cancer drugs (33,58-60).
illustration not visible in this excerpt
Figure 4: Various gradients exist in multicellular tumor spheroids (MCTS). Combination of analytical images of multicellular tumor spheroids (MCTS) median sections studied with different technologies: autoradiography, the tunnel assay, bioluminescence imaging, and probing with oxygen microelectrodes. Together these measurements demonstrate the gradients of cell proliferation, viability and the microenvironment within large MCTS that resemble parameters of in vivo tumors. Figure and legend adapted from Hirschhaeuser et al. (21).
Consequently, the MCTS is a physiologically-relevant cell culture model to mimic tumor dormancy in vitro and could aid in the discovery of new drug candidates that target dormant cancer cells in solid tumors.
A drawback in using 3D cell cultures is the limited possibilities to identify localized phenotypes yet. To monitor compound effects on MCTS, simple endpoint assays are frequently used (e.g. viability readouts) (61,62). These assays are also often conducted on whole MCTS or even on pooled samples. Thus no spatial or temporal information can be collected, making it in turn infeasible to discriminate between different populations or properties of the respective subpopulations including collecting single cell information (63).
New microplate-compatible integrated microscopy/image processing systems with low phototoxicity (e.g. light-sheet illumination microscopy (64) or confocal spinning disc microscopes) improve penetration depth and resolution of 3D samples. The possibility of live cell imaging in combination with non-toxic fluorescent cell labeling allows new insights into drug mechanism with high spatial resolution (54,63,65). Therefore a high throughput, high-content microscopy compatible 3D cell culture assay to identify substances that specifically target dormant cells in MCTS core regions would be a valuable tool in oncological drug discovery.
Cell migration and invasion are processes present in key physiologic and pathologic phenomena such as wound healing and cancer metastasis and affect many different cell types (66,67). Populations of cancer cells in tumors display variability in their ability to metastasize (40). Early cancer cell invasion is considered to be an important step in tumor metastasis (67,68). Involved signaling pathways control cytoskeletal dynamics in invading cells and the turnover of extracellular matrix and cell-cell contacts, enabling cell migration into adjacent tissue (27). However, malignant invasion processes are not limited to cancer but also occur in fibrotic diseases and contribute to different kinds of malignant processes, for example in pulmonary fibrosis (66,69). Here, invaded fibroblasts are the primary cell type that abnormally synthesizes extracellular matrix (ECM) components such as fibronectin, proteoglycans, hyaluronic acid, and interstitial collagens (70). Various invasion assays for in vitro studies have been successfully established, like the Matrigel invasion assay (based on a gelatinous protein mixture), 3-dimensional collagen assay, and the scratch invasion assay to support drug discovery (66,71). Although very useful tools in metastasis research, these in vitro models do not necessarily reflect the physiological events that facilitate the dissemination of cancer cells or fibroblasts. Moreover, these assays have not yet been adapted for screening purposes and therefore their throughput is restricted (72). Although a few screening compatible assays like the transwell migration/invasion assay or the platypus invasion assay exist to monitor fibrotic processes, no or only artificial matrices (e.g. Matrigel or collagen gels) are used to mimic physiological ECM composition of tissue. Therefore, they neglect crucial physiological parameters of invasion into ECM of tissue (73).
As molecular mechanisms underlying migration and invasion are complex and depend on direct cell- cell and cell-matrix interactions, in vitro assays that better resemble physiological conditions are needed.
3D cell cultures are able to reflect physiological conditions in regard to ECM composition (57), stiffness (74) and establishment of cell-cell junctions (75,76) in 3-dimensions. Thus, 3D cell cocultures could provide a relevant microenvironment for cellular studies of invasion. The determination of the migratory and invasive potential of tumor cells and activated fibroblasts is fundamental for drug discovery and in proposing novel clinical strategies.
A lack of physiologically-relevant phenotypic screening assays hampers the discovery of novel drug candidates for one of the most threatening fibrotic diseases, idiopathic pulmonary fibrosis (IPF), which currently has no treatment options available (77). Monitoring invasion processes in a physiological 3D in vitro co-culture model could therefore support the discovery of novel inhibitors/modulators of invasion processes in IPF.
In idiopathic pulmonary fibrosis (IPF) aberrant migration and proliferation of fibroblasts into normal lung tissue as well as the accumulation of extracellular matrix (ECM) proteins such as collagen (78,79) leads to a dramatic decrease of oxygen capacity and eventually death in IPF patients (Figure 5). The rapid progression of this interstitial lung disease is associated with a median survival rate of 2-3 years from initial diagnosis (80).
Figure 5: Schematic comparison of normal lung architecture and impairment by idiopathic pulmonary fibrosis (IPF).
(A) Normal lung function is characterized by intact gas exchange and homeostasis of fibrotic and epithelial tissue. (B) In IPF fibrotic scarring drastically decreases gas exchange and therefore oxygen consumption into the blood stream. Fibrosis is thought to be a main cause of impaired respiratory function. Image copyright by National Institutes of Health / Department of Health and Human Services.
The histopathological criteria for IPF are the same as for usual interstitial pneumonia with a temporally and locally heterogeneous combination of fibrosis, scarring, and honeycombing (fibrotic cystic changes) within the lung (79). There has been a shift in the understanding of the pathophysiology of IPF from one of a chronic inflammatory disease state to one of abnormal wound healing, fibroblast recruitment, and migration process. Indeed, this could be a reason why broad anti-inflammatory and immunosuppressive therapies have so far not been able to alter fatal outcomes (79,81). However, targeting cytokine networks involved in immune and cell activation is still under investigation and crosslinks them with pro-proliferative and pro-invasive factors like transforming growth factor-beta (TGF- he vascular endothelial growth factor receptor (VEGFR).
Recently discovered pathogenic pathways and mediators in IPF focus on aberrant fibroblast activation processes, abnormal tissue remodeling, and ECM accumulation that is linked to myofibroblast differentiation (82,83).
Induced myofibroblast differentiation in IPF | Fibroblasts are largely responsible for synthesizing the ECM components in fibrotic lesions inside the lung of IPF patients. The main components are collagens type 1 and 3 and fibronectin, the secretion of which is directly influenced by different growth and differentiation factors. These factors, among them TGF- alpha- -SMA) expression and therefore differentiation of fibroblasts to ! " -SMA expression alters the shape and morphology of fibroblasts and induces stress fiber formation. This type of fibroblast with contractile properties is thought to be instrumental for lung contraction, alveolar collapse, and ECM deposition (77,84,85) present in IPF. Therefore, two main treatment strategies seem promising: First, inhibition and therefore prevention of invasion of fibroblasts in lung tissue and second, identification of compounds that limit TGFinduced myofibroblast differentiation.
Models for IPF | The use of fibroblast cultures has allowed investigation of the characteristics of fibroblasts or myofibroblasts in IPF, of their resistance to apoptosis, and of the synthesis of profibrotic factors and ECM components (72,73). However, there are still some obstacles in adapting cellular in vitro assays mimicking the fibrotic process for screening purposes prior to animal testing. Specifically 2D cell culture settings lack crucial factors of ECM composition, stiffness, and cell-cell or cell-matrix interactions (72,86).
Therefore, the main method for preclinical evaluation of new treatment options in IPF is the bleomycin-induced IPF in vivo mouse models (87). As these models have some limitations in regard to screening capability, such as high costs, low throughput, and ethical considerations, novel predictive in vitro assays for lung fibrosis are required.
Heterotypic 3D cell co-culture models have the potential to mimic the complex three-dimensional organization of tissue in vitro. Cell-cell contacts present in 3D cell co-cultures can reproduce physiological conditions that induce signaling events that regulate cell invasion and are crucial for the fibrotic process (63,77,88).
Novel 3D co-culture invasion assay | A 3D co-culture assay for tissue interaction that uses multicellular tumor spheroids (MCTS) to provide a 3D tissue-like surrogate (TLS) in combination with lung fibroblasts could mimic the invasion processes underlying IPF in vitro.
In this context, HCA technology could allow spatial and temporal monitoring of invasion and migration of lung fibroblasts into 3D tissue-like surrogates. Moreover, subsequent automated image analysis would enable compound characterization in a high-throughput, High-Content screening (HCS) campaign. Such microplate-based co-culture assays could therefore have the potential to reveal new drug candidates in a phenotypic screening approach with higher probability of translation into in vivo activity.
Dormant tumor cells represent a non-proliferating tumor cell fraction that can resist commonly used cytostatic-based chemotherapies. Tumor dormancy is therefore thought to be a major source of tumor relapse and recurrence in solid tumors. Only a limited number of assays are available that would allow for the reproduction of tumor dormancy in a physiologically-relevant in vitro model with high-throughput capability. The aim of this work was to develop and evaluate a high-throughput, high-content microscopy compatible 3D multicellular tumor spheroids (MCTS) assay in microplates to identify substances that specifically target chemotherapy-resistant dormant cells in MCTS core regions. This included the validation of techniques to establish high-throughput compatible methods to produce compact and reproducible MCTS. Furthermore, adequate staining procedures to identify localized cell death in inner core region were developed. To fulfill the requirements of HTS, the setup was (semi-) automated and image acquisition as well as image analysis was automated. As a proof of principle, a screen of two compound libraries was conducted. Resulting hits that specifically target cells in inner MCTS regions, and leave cells in outer regions or cultured under 2D cell culture conditions unaffected, were further validated and classified as respiratory chain inhibitors. In addition, secondary assays were established to identify drug targets of hit compounds and decipher their molecular pathways. Finally, benefits of combinatorial therapies of identified hits showed increased efficacy when combined with commonly used chemotherapeutic drugs in vitro.
As targeting malignant invasion processes is of great medical need in several indications, a novel 3D cell co-culture assay to mimic complex interaction of different cell types was established. In idiopathic pulmonary fibrosis (IPF), aberrant fibroblast invasion into normal lung tissue dramatically reduces oxygen uptake capacity and leads to high mortality rates. So far, no effective treatment strategies are available. Thus, suitable cell types to mimic invasion and migration were investigated to model fibrotic processes in vitro. In addition, suitable staining strategies to discriminate different subpopulations in heterotypic 3D cell co-cultures and adaptation of image analysis routines for automated analysis were evaluated. With this setup modulators/inhibitors of fibrotic processes in a pilot screen were identified.
In summary, this work could form the basis for later studies in this research area leading to novel, physiologically-relevant 3D assays that could accelerate the drug discovery process and help to reduce high attrition rates.
illustration not visible in this excerpt
illustration not visible in this excerpt
illustration not visible in this excerpt
illustration not visible in this excerpt
Unless otherwise mentioned, kits and enzymes were used according to manufacturers’ instructions.
Normal goat serum blocking (NGSB) | 2% goat serum, 1% BSA, 0.1% Triton X-100, 0.05% Tween 20, 0.01 M PBS, pH 7.2, stored frozen at - 20 °C.
Spheroid permeabilization buffer | 1% BSA, 2% Triton X-100, 0.05% Tween 20, 0.01 M PBS, pH 7.2, stored frozen at - 20 °C.
Complete Lysis Buffer for Mesoscale kit| see MSD instructions
illustration not visible in this excerpt
illustration not visible in this excerpt
ACTB (encoding Beta-actin), Hs01060665_g1
ACYL (ATP citrate lyase), Hs00982738_m1
ANGPTL4 (angiopoietin-like 4), Hs01101127_m1
BNIP3 (BCL2/adenovirus E1B 19kDa interacting protein 3), Hs01087963_m1 CA9 (carbonic anhydrase IX), Hs00154208_m1
GLUT1 (solute carrier family 2, member 1), Hs00892681_m1
RPL13A (encoding 60S ribosomal protein L13a), Hs04194366_g1 VEGFA (vascular endothelial growth factor A), Hs00900055_m1
Genedata Condoseo for IC50 Determination, Genedata
Genedata Screener for High-Content Screening, Genedata In-house HCA Data Management software
Illustrator CS5 Graphical Layout software, Adobe
Imaris 7.6, Scientific 3D/4D Image Processing & Analysis, Bitplane
MetaXpress High-Content Image Acquisition & Analysis Software, Molecular Devices Office 2010 Package, Microsoft
Prism Scientific Presentation Software, GraphPad Software
For a 1,5 % agarose solution (wt/vol), Agarose NA (GE Healthcare) was added to 100 ml phenol redfree DMEM (w/o FCS) in an appropriate beaker. The solution was heated three times to boiling in a microwave and kept at 60-70°C on a stirring heating plate. The agarose solution was kept near a Multidrop Combi (Thermo Fischer), which was equipped with a 2500 μl 8-channel dispensing cassette. In order to prevent premature solidifying of agarose solution, tubes and the dispensing cassette were kept at 40-50°C with red light lamps.
With this setup up to twenty 384-well μClear imaging microplates were coated with 10 μl agarose solution per well in one run (Figure 6). For temperature and humidity equilibration the plates were kept open in a clean bench for 20 minutes at room temperature until further usage. Plates wrapped in moistened aluminum foil were stored for up to one week. The described procedure is transferable to 96-well plates by increasing the dispensing volume to 40 μl per well.
illustration not visible in this excerpt
Figure 6: Workflow of semi-automated agarose coating for microplates. Production of agarose coated imaging microplates is feasible for 96, 384, and 1536-well microplates with heated dispensing systems and automated Multidrop dispenser. Several plates in one run can be coated and used for initiation of spheroids from single cell suspensions of widely used cancer cell lines (see also Table 1). After initiation (for 3-4 days) round and compact spheroids can be used for assessment of compound effects. Imaging and readout is scalable to automatable staining protocols and microscopy systems (no need to remove spheroids from the imaging plates) or 3D data processing with manual high resolution microscopy systems or immunostaining of whole spheroids (see section 3.2.8) (54).
All cell lines used were cultured in RPMI 1640 (Gibco) supplemented with 10% FCS (Gibco-BRL), 2 mM L-Glutamine (Gibco-BRL), and 1% Penicillin/Streptomycin (Sigma Aldrich). For T47D cells 0.01 μg/ml Insulin was included. Fibroblast cell lines (MRC-5, Normal dermal fibroblast (NDF) or human pulmonary fibroblasts (HPF)) were cultured in DMEM/HamsF12 Nutrient Mix with 15% FCS and 1% Penicillin/Streptomycin. Monolayer cells were detached with trypsin and resuspended in culture media. 100 μl of the suspension was added in 10 ml CASY-TON and used for automated cell counting in a CASY cell counter.
Cell line seeding number was optimized to obtain multicellular tumor spheroids (MCTS) with an approximate diameter of 400 μm on day 4. 2000 cells per well (c/w) for T47D, 5000 c/w for DLD1, 2000 c/w for DU145, 2000 c/w for MRC5 (fibroblast), 2000 c/w for NDF (human dermal fibroblast), and 1000 c/w for primary colon cancer cells. For 384-well plates 40 μl of the cell suspensions and for 96-well 50 μl respectively were seed with a Multidrop Combi.
The plates were incubated under standard cell culture conditions at 37°C and 5 % CO2 in humidified incubators for 4 days to allow formation of reproducible spheroids of defined size and morphology. Compounds were added to 20 μl culture medium for an additional 72 h. For the invasion assay MCTS grown for 4 days was simultaneously incubated with 20 μl compound and 20 μl of the invading cell type, e.g. MRC-5 fibroblasts (2000 c/w) for 72 h.
Prior to imaging, spheroids were stained for 24 h by adding Hoechst 33342 DNA stain(1 mg/ml, Life Technologies) to visualize all cell nuclei. SytoxGreen (2 mM, Life Technologies) was used as a cell death marker. It is a membrane-impermeable DNA dye that diffuses only into dead cells as a result of impaired membrane permeability. Both dyes were added at a final dilution of 1:10.000.
illustration not visible in this excerpt
Figure 7: Integration of 3D cell cultures in a High-Content Screen (HCS) setup. After spheroid generation, compound incubation and staining procedure imaging was done using automated microscopy systems (Perkin Elmer confocal Opera system or Molecular Devices Micro XL widefield system). Subsequent data transfer and migration was done with custom build in-house software. Image analysis was done with custom analysis routines in MetaXpress. Hit evaluation and interpretation of data was done in GeneData software suites.
The non-toxic live cell stain Vybrant DiD (Vybrant® DiD Cell-Labeling Solution, Life Technologies) was used to stain fibroblasts for visualization of invasion. Briefly, cells were collected in 2 - 4 ml serumfree media (~ 1 × 106 /mL) and incubated with Vybrant DiD (f.c. 1:200) for 15 min at 37°C. Cells were washed once with 10 ml of FCS-containing media, centrifuged (5 min, 800 rpm) and resuspended in 5 ml warm culture medium for cell counting and further use.
Images of MCTS or 2D cell cultures were either captured by Molecular Devices Micro widefield system with a 2X objective or in a confocal Perkin Elmer Opera microscopy system. Data handling was accomplished with in-house data migration software. Automated image analysis and quantification of inner core cell death was done with MetaXpress software (Molecular Devices) using custom written image analysis routines. Briefly, spheroid borders were detected on Hoechst-stained fluorescent image and masks were generated, scaled down and transferred on SytoxGreen-stained fluorescent image to quantify cell death in inner spheroid regions. To observe invasion of fibroblasts the area of the invaded fibroblast was additionally quantified and used as readout.
illustration not visible in this excerpt
Figure 8: Custom-build mask generation and artifact removal allows identification of localized phenotypes. Image analysis in MetaXpress can discriminate different kinds of hit classes: By measuring spheroid area growth, arresting compounds can be excluded. In addition general cytotoxic compounds can be excluded and, by building a ratio of the dead cell intensity channel, localized cell death in inner core regions of MCTS can be quantified and analyzed. Generated data files are then further analyzed with GeneData Screener software.
Brightfield images were acquired on Celigo 200 automated light microscopy system (3.5X objective) and further processed with built-in application modules. Briefly, spheroids were detected in brightfield images and area and/or diameter was calculated by the Celigo software and further processed in Excel (Microsoft).
All images were captured as 12-bit tiff files and no non-linear corrections were being applied.
Light sheet-based microscopy is a new method for imaging large fluorescent samples. Optical sectioning is achieved by illuminating the biological 3D sample with a sheet of light, then observing fluorescence with a wide-field microscope. A standard charge-coupled device (CCD) camera is used to rapidly acquire images with a high dynamic range. Because only the plane that is in focus is illuminated, phototoxicity is limited and many planes can be acquired. The sample moves along the optical axis of the detection system while immersed in a glass capillary containing chemical clearing solution (64).
illustration not visible in this excerpt
Figure 9: Pivotable incubation chamber of a custom-build light sheet microscope. Sample can be moved on the XYZ stage through a focused sheet of light. Emitted light is then detected in 90 degree with a detection objective coupled to a charge coupled device (CCD) camera. Low phototoxicity and the possibility to image the specimen in different angles are some advantages over standard fluorescent microscopes. By this, optical sectioning through large specimen is feasible for the generation of 3D data sets. Moreover, live cell observations are possible with a heatable and perfusable incubation chamber (64,65).
The effects of chemical clearing rely on immersing samples in a medium having a refractive index similar to biological structures. Light scattering is thus substantially reduced and the sample becomes transparent (89).
3D image acquisition was performed on a custom build monolithic digital light sheet based fluorescence (mDSLM) microscopy system as described earlier (64). Briefly, samples were dehydrated in an ascending ethanol series (50%, 70%, 85%, 99%) for 5 min each. Then spheroids were transferred to benzyl alcohol/benzyl benzoate (1:1, v/v), transferred to a glass capillary (Borosilicate glass capillaries) and imaged using 2,5X illumination objective and 10X detection objective. About 100-140 z-planes with 2,58 μm spacing were imaged for T47D spheroids if not otherwise specified. 3D reconstruction and movie generation was done with Imaris software (Bitplane).
All images were captured as 16-bit tiff files and no non-linear corrections were being applied.
For 2D toxicity assessment T47D cells were seeded at 2250 cells per well in 40 μl on 384-well plates and were allowed to attach for 24 h. After 3 days drug incubation, cell viability was determined. Again Hoechst and SytoxGreen were used for staining (with the same concentration as used in 3D toxicity assessment). This time cell death was calculated by counting all cells (as detected by Hoechst staining) divided by SytoxGreen positive cells with build-in applications of the MetaXpress software.
Viability for in vitro combination studies in MCTS was measured with Cell Titer Glo Assay (Promega). The assay measures ATP content after cell lysis using ATP-dependent bioluminescence of luciferase enzyme. To support reagent penetration, lysis, and ATP recovery from MCTS, an equal volume of reagent was added to the sample and shaken for 15 min at 450 rpm. Luminescence readout was performed after 30 min incubation at room temperature on a Tecan Infinite M1000 luminometer.
Antibody penetration throughout the whole 3D sample is one of the main problems arising in confocal imaging of thick biological samples. Therefore, I adapted a protocol from Louis-Bastien Weiswald et al. (90) with long antibody incubation times (48 h) and extended use of permeabilization reagent (1% BSA, 2% Triton X-100, 0.05% Tween 20). Briefly, prior to harvest, spheroids were fixed for 1 h in 4% PFA. After 1 h blocking in mild permeabilization solution (1% BSA, 0.1%Triton, 0.1% TWEEN-20) spheroids were exposed to spheroid permeabilization buffer (1% BSA, 2% Triton X-100,
0.05% Tween 20) and subsequently incubated with primary antibody (fluorescein isothiocyanateconjugated antibody against phosphorylated histone H3 Ser-10) for 48 h. After staining, whole spheroids were transferred to 384-well plates and imaged on a Perkin Elmer confocal microscopy system with a 20X objective. 28 z-planes were taken with 4 μm spacing.
Prior to harvest, spheroids were fixed for 1 h in 4% PFA. Then spheroids were transferred to 50 ml tubes (Falcon), washed twice in ice-cold DPBS and equilibrated in 30% sucrose (w/v) DPBS solution for 1 h. Then spheroids were transferred to cryomolds and covered in Tissue-Tek OCT compound. After 30 min of equilibration cryomolds were frozen by incubation in a mixture of dry ice and 2- Methylbutane (Sigma Aldrich). Prepared samples were cut into 5 μm sections by cryostat, mounted on SuperFrost Plus slides (Menzel-Glaser) and then rehydrated in DPBS for 20 min. After 1 h in blocking and permeabilization solution (1% BSA, 0.1%Triton, 0.1% TWEEN-20) primary antibody was incubated over night at 4°C. Staining with Hydroxyprobe-1 mouse monoclonal IgG1 (Hypoxyprobe-1, Chemicon) labeled with FITC and cell labeling with Click-iT EdU imaging kit (Alexa Fluor 555 azide, Life Technologies) was performed according to the respective manufacturers instructions. Finally, slides were incubated with Hoechst 33342 (0.1 μg/mL) for 5 min before an additional wash step. After staining, slides were mounted in Slowfade Gold (Life Technologies) and imaged on AxioInvert 500 (CarlZeiss) with 10X air objective and attached camera.
For comparison of hypoxia in eight different cancer cell lines, spheroids were grown for seven days. Prior to harvest EdU (10 μM) incubation for 6 h and pimonidazole (100 μM) incubation for 2 h was conducted. The staining procedure was carried out as described above. Cell numbers were used as follows (cells/well = c/w): Mouse cell lines: Lewis Lung 300 c/w, F9 100 c/w, MM5MTC 4000 c/w, EMT6 500 c/w. Human cell lines T47D 2000 c/w, DLD1 3000 c/w, H460 200 c/w, H1299 5000 c/w.
AMP-activated protein kinase (AMPK) activation was determined by MSD Prototype assay for human duplex (phospho) alpha1-AMPK. Electrochemiluminescence detection uses labels that emit light when electrochemically stimulated. Background signals are minimal because the stimulation mechanism (electricity) is decoupled from the signal (light). Labels are stable, non-radioactive and offer a choice of convenient coupling chemistries. They emit light at ~620 nm, eliminating problems with color quenching. Multiple excitation cycles of each label amplify the signal to enhance light levels and improve sensitivity (Information adapted from MSD application homepage (91)). The first capture antibody is directed against catalytic subunit alpha1 of AMPK and the second detects endogenous AMPK-alpha1 only when phosphorylated at threonine 172. Samples were collected following manufacturer’s instructions and adjusted to same protein concentrations (30 μg/well). 96 spheroids per condition were collected in 50 ml tubes (BD Falcon), washed with 5mL PBS, incubated in 200 μl MSD complete lysis buffer for 30 min, vortexed and vigorously pipetted up and down every 10 minutes. Phosphorylated AMPK was normalized to total AMPK amount and used as readout for AMPK activation.
RNA was extracted using the 6100 NucleicPrepStation (Applied Biosystems) according to manufacturer's instructions and quantified on the 8-channel Nanodrop spectrophotometer (ND8000, Thermo Scientific). cDNA was produced using the GeneAmp ® RNA PCR Kit (Life Technologies) and quantitative reverse transcription PCR (RT-qPCR) was conducted using Applied Biosystems TaqMan® Gene Expression Assays before analysis on the 7900 PCR machine (Applied Biosystems). Relative mRNA levels were calculated to the geometric mean of reference genes ACTB (encoding Beta-actin) and RPL13A (encoding 60S ribosomal protein L13a). Full gene names: ACYL (ATP citrate lyase), GLUT1 (solute carrier family 2 (facilitated glucose transporter), member 1), ANGPTL4 (angiopoietin-like 4), BNIP3 (BCL2/adenovirus E1B 19kDa interacting protein 3), CA9 (carbonic anhydrase IX), VEGFA (vascular endothelial growth factor A).
Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured with an XFe Extracellular Flux Analyzer (Seahorse Bioscience) according to manufactures instructions. Briefly, cells were plated at 25.000 cells/well in standard cell culture medium. After 24 h media was exchanged to non-buffered RPMI1640 containing 11 mM glucose and 2 mM glutamine, and equilibrated in a CO2-free incubator for 1 h. The XF assay consisted of sequential mix, pause and measurement steps, allowing determination of OCR and ECAR every 10 min for up to 180 min. The concentrations used in this assay are provided in Supplementary Table 9.3.
For the extraction of metabolites and sample preparation for1 H NMR spectroscopy, the protocol from Ellis et al (92) was adapted to T47D breast cancer cells. Briefly, the cell culture media of each well of the 12-well plate was removed quickly and the cells were washed twice with ice-cold PBS. To quench, 1 ml ice-cold methanol was added to each well, scraped and transferred to an Eppendorf tube. Each well was washed again with 1 ml ice-cold methanol and the two samples pooled. The pooled quenched methanol sample was dried in a Speedvac R concentrator (SPD111V, Thermo Scientific) at 50°C for 1 hour and stored at -80°C until subsequent extraction.
Water-soluble metabolites were extracted from the T47D breast cancer cells by adding 300 μl chloroform/methanol (2:1) solution to each of the dried quenched methanol samples. The sample was mixed with a vortex for 30 seconds, 300 μl of ultrapure water was added and vortexed again. The sample was centrifuged at 13.000g for 5 min and the aqueous and the organic layer removed to separate sample tubes. The organic phase was not further analyzed in this study. The aqueous phase was dried in a Speedvac R concentrator (SPD111V, Thermo Scientific) at 50°C for 1 hour and reconstituted in 300 μl of ice-cold phosphate buffer (0.2 M Na2HPO4, 0.043 M NaH2PO4, 25 μM TSP, 3 mM NaN3 in 99.8 % D2O) and centrifuged at 16.000xg for 5 min at 4°C. From the supernatant 290 μl were transferred to a well of a 96-deep well plate and placed on ice. The samples were transferred with a Gilson Liquid Handler System (Bruker BioSpin) from the positions in the deep-well plate (kept at 4 °C) to 3 mm NMR tubes.
Spectra were acquired in 3 mm NMR tubes at 600.13 MHz and 300 K using a Bruker AVANCE III spectrometer equipped with a TCI-Cryo-Probe and a sample jet system (Bruker BioSpin). Until measurement, the samples were kept at 6°C by the sample jet system. The residual water signal was suppressed by a 1D-NOESY pre-saturation pulse sequence. Typically, a total of 512 transients each of 64 k data points was acquired with an acquisition time of 2.65 s, an interpulse delay of 4 s, a spectral width of 20 ppm and a pulse width of 8.2 μs at 5 dB (90°). The free induction decay (FID) was multiplied by a 0.3 Hz exponential line-broadening factor to improve the signal-to-noise ratio prior to Fourier transformation. Phase correction and referencing was performed using Topspin 2.1 (Bruker BioSpin). For the baseline correction ACD Software Suite 12 (ACD/Labs) was used. The TSP signal was set to 0.00 ppm.
Over the spectral region of 0 - 10 ppm, the spectra were segmented into buckets of an equal width of 0.04 ppm and the signal intensity in each region was integrated using AMIX (version 3.8.4, Bruker BioSpin). The region between 4.6 - 5.0 ppm was deleted to avoid artefacts from the water suppression. To account for potential inter-individual variations in the diuresis, the spectra were normalized to a constant integrated intensity of 1 unit.
Multivariate analysis of the integrated bucket data was performed using SIMCA-P software (version 13.0, Umetrics AB) applying pareto scaling. Unsupervised principal components analysis (PCA) and supervised models (PLS-DA, OPLS-DA) were used to extract the main drivers, or spectral regions of the spectra responsible for group separation.
Metabolites that are likely to account for integral changes of buckets selected by the procedure described above were manually annotated and quantified with the help of the Chenomx NMR Suite 7.5 (Chenomx Inc.). The metabolite concentrations were expressed in (mM) using the integrated TSP region (0.025 mM). Quantitative metabolite data were exported to MS Excel (version 2010), where the data were further processed as follows: To account for variations in cell number, the intracellular metabolite concentrations were normalized for each sample to 1 unit. The cellular metabolite concentrations, therefore, are shown as mM. Multivariate analysis of the quantitative metabolite data was performed according to the procedures described above. However, all metabolite concentrations were scaled to unite variance (UV).
The heatmap was built using the same called function of GNU R 3.0.1. Data was logarithmized in advance and clustered with hierarchical clustering using Euclidean distance measure and complete linkage agglomeration.
One-way ANOVA tests for combination therapy experiments were carried out using Prism software (GraphPad). Significance of differences between multiple groups was compared using a Bonferroni posttest analysis at the last time point of treatment. RT-qPCR gene expression levels and AMPK activation levels were compared by multiple t-test using Prism software (two-tailed t-test with Welch’s correction).
Normalization, quality control, hit list generation and fitting curves for AC50 determination of identified hit compounds was done with Genedata Screener® for High Content Screening and Genedata Condoseo modules (Genedata AG). AC50 represents the concentration of the compound eliciting 50% of the maximal cell death in inner core regions.
With three-dimensional growth conditions, 3D multicellular tumor spheroids (MCTS) or “spheroids” mimic several parameters of the in vivo tumor microenvironment. Therefore they have the potential to better represent the tumor environment as well as simulate physiological drug responses in vitro. In recent years MCTS have been recognized as valuable cell biology tool, which has led to an increase of new technologies that support the formation of MCTS.
Evaluation of 3D cell cultures for phenotypic high-throughput assays | Advantages and disadvantages of these techniques are strongly dependent on the context in which the multicellular tumor spheroids (MCTS) are used (see Introduction 1.5). The following requirements must be met for the integration of MCTS in a High-Content Screening (HCS) campaign:
-At least 96-well microplates are supported (ideally 384- or even 1536-well microplates)
-Automatable and reliable methodology exists for MCTS generation
-Uniform spheroids (spheroids with defined and reproducible size, composition and morphology) may be created
-Automated microscopy for High-Content Analysis (HCA) is available
-Cellular structures or markers for HCA may be visualized
-No transfer of spheroids is needed if possible
-Image acquisition and analysis routines are suitable for 3D data sets
As the goal was to establish an HCA-screening compatible 3D cell culture system, the focus of this work was to meet the special needs and requirements for the adaptation of 3D cell cultures in an automated high-throughput screening and High-Content screening (HTS/HCS) environment.
Most scaffold-based 3D cell culture systems are not available in microplate formats, which is a prerequisite for image-based High-Content screening applications (32,37). Additionally, almost all scaffold-based techniques interfere with microscopic readouts and the scaffolds have to be removed prior to imaging (75). Therefore only scaffold-free systems were evaluated.
Bachelorarbeit, 55 Seiten
Bachelorarbeit, 72 Seiten
Bachelorarbeit, 55 Seiten
Doktorarbeit / Dissertation, 166 Seiten
Doktorarbeit / Dissertation, 365 Seiten
Diplomarbeit, 100 Seiten
Masterarbeit, 148 Seiten
Bachelorarbeit, 37 Seiten
Bachelorarbeit, 37 Seiten
Der GRIN Verlag hat sich seit 1998 auf die Veröffentlichung akademischer eBooks und Bücher spezialisiert. Der GRIN Verlag steht damit als erstes Unternehmen für User Generated Quality Content. Die Verlagsseiten GRIN.com, Hausarbeiten.de und Diplomarbeiten24 bieten für Hochschullehrer, Absolventen und Studenten die ideale Plattform, wissenschaftliche Texte wie Hausarbeiten, Referate, Bachelorarbeiten, Masterarbeiten, Diplomarbeiten, Dissertationen und wissenschaftliche Aufsätze einem breiten Publikum zu präsentieren.
Kostenfreie Veröffentlichung: Hausarbeit, Bachelorarbeit, Diplomarbeit, Dissertation, Masterarbeit, Interpretation oder Referat jetzt veröffentlichen!