See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/6158218 Meeting report: Trends and challenges in high content analysis Article in Biotechnology Journal · August 2007 DOI: 10.1002/biot.200700101 · Source: PubMed CITATIONS READS 0 15 2 authors: Urban Liebel Wolfgang Link 56 PUBLICATIONS 1,763 CITATIONS 56 PUBLICATIONS 2,314 CITATIONS ACQUIFER AG SEE PROFILE Universidade do Algarve SEE PROFILE All content following this page was uploaded by Wolfgang Link on 18 December 2014. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. BTJ-FORUM Biotechnology Journal DOI 10.1002/biot.200700101 Meeting report Trends and challenges in high content analysis Image acquisition using robotic fluorescent microscopy and automated image analysis is now generally referred to as high content screening (HCS), a powerful tool to carry out cell biology on a large scale both in basic academic research and in early drug discovery. The recent Conference, Workshop & Exhibition “High Content Analysis, Spain” brought together scientists from the pharmaceutical, biotechnological and academic sectors interested in the trends, developments and recent advances in HCS technology. “High Content Analysis, Spain” was held in Madrid on the 26–27 March 2007, jointly hosted by the Spanish National Cancer Center (CNIO – Centro Nacional de Investigaciones Oncologicas) and Spanish National Cardiovascular Research Center (CNIC – Centro Nacional de Investigaciones Cardiovasculares). The meeting focused on a variety of HCS applications including compound and RNAi library screening, drug profiling and toxicity evaluation. In his welcome address, W. Link (CNIO, Madrid, Spain) described the technological and scientific context of HCS in the post-genomic era. The compound and RNAi libraries available provide opportunities to carry out large-scale inhibition of protein function or gene knockdown in mammalian cells. Indeed, a major breakthrough was achieved when HCS was used to study complex phenotypes in the context of the living cell. High content cellular imaging has met the challenge of introducing high-throughput procedures into the drug discovery process to identify high quality lead compounds [1]. As such, HCS has the potential to decrease the current drug attrition rate by means of large-scale analysis of disease relevant cellular events. 938 A cell-based model was presented by P. J. O’Brien (Veterinary Science Centre; University College Dublin, Ireland) that combined several critical features to assess human hepatotoxicity using high content technology. Drug-induced hepatotoxicity is one of the most clinically prevalent adverse effects and it represents a major cause of candidate drug attrition. Conventional assays have not been reliable in predicting cytotoxicity due to their low sensitivity in detecting potential human hepatotoxicity. Furthermore, there is little concordance between human hepatotoxicity and that observed in animal toxicity tests. However, the predictive values of several conventional cell-based assays is additive and, thus, predictive value can be dramatically increased in HCS assays that combine: multiple days of exposure of cells to drugs; the use of a metabolic competent human hepatocyte cell line; and the assessment of multiple pre-lethal effects in individual live cells, including mitochondrial toxicity, oxidative stress, deregulation of calcium homeostasis, phospholipidosis, apoptosis, and anti-proliferative effects [2]. Cell proliferation was the single-most sensitive parameter. The low proliferation rate of primary human hepatocytes that are more metabolically competent make them less suitable for cytotoxic assessment. Furthermore, primary human hepatocytes are phenotypically unstable under current cell culture conditions and dedifferentiate rapidly, resulting in decreased liver-specific activity. The use of high content technology for multivariate drug profiling was discussed by S. J. Altschuler (University of Texas Southwestern Medical Center, TX, USA). One of the key challenges for image-based analysis of cell © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Biotechnol. J. 2007, 2, 938–940 populations is to detect meaningful patterns among the countless events that occur in single cells. Mathematical modeling has been used to analyze multidimensional measurements of human cancer cells treated with different concentrations of an arrayed panel of 100 compounds at an individual cell level [3, 4]. This hypothesisfree approach to the phenotypic profiling of drug action is somewhat analogous to DNA microarrays. In fact, the results were visualized with heat map plots similar to the red and green gene expression fingerprints from DNA chips. The predictive power of this model is reflected by the fact that treatment with compounds known to act through the same mechanism but with very different chemical structures produced similar cytological profiles. Dr. Altschuler stressed the need for common standards in the HCS field to permit the interchange and comparison of data obtained using different HCS platforms and experimental procedures. Several talks focused on the use of high content approaches to study the impact of small molecule inhibitors or siRNA on the cell cycle. M. J. Lallena (Eli Lilly, Madrid, Spain) acknowledged the importance of finding selective cell cycle inhibitors for anticancer therapy. Using a high content fluorescence microplate cytometer to monitor the cell levels of pHis H3, CyclinB, p-Rb, and the DNA content in a multiplexed manner, proof of principle experiments were conducted using selective small molecules for various cell cycle kinases, arresting cells in G2/M or G1/S. A great deal of progress has been made in the last few years in developing new tools for fluorescent microscopy and increasing the versatility of imaging instruments. New developments in the field of commercially available automated imaging platforms and innovative fluorescence options for high content imaging were presented by N. Thomas (GE Healthcare), K. Herrenknecht (Evotec), T. Horn (BD Biosciences) and T. Bauer (Invitrogen). Biotechnol. J. 2007, 2, 938–940 online image compression, etc.). In combination with several of these open HCS microscopes, transfected cell arrays permit genome-wide inspection in live cells over 48 h, generating data sets with tens of millions of images. Recent developments, examples and bioinformatics data integration are accessible via http://liebel. fzk.de and http://harvester.fzk.de. M. de los Frailes (Glaxo Smithkline, Tres Cantos, Spain) discussed the characteristics of the high throughput assays suitable to screen the GSK corporate compound collection. The increased costs associated with drug discovery and development programs and the high attrition rate of drug candidates have changed the current drug discovery paradigm, shifting towards information-rich cellbased primary screening. More than 50% of high throughput screening (HTS) and compound profiling assays used in early drug discovery are cellbased assays, and this trend is clearly increasing. However, the challenge remains to implement screening procedures that can cope with increased throughput while retaining information-rich readouts. Fluorometric Imaging Plate Reader (FLIPR)-based quantitative optical screens for cell-based kinetic assays provide a good compromise between throughput and the capacity to generate complex and information-rich data sets. FLIPR permits simultaneous and real time measurement of agonist and antagonist activity in 60 000 samples per day. The identification of agonists and antagonists of the same receptor served as the basis to initiate parallel lead optimization programs for multiple therapeutic areas. Another FLIPR-based screening strategy permitted specific allosteric modulators to be identified for members of the class B and C seven transmembrane domain (7TM) receptor families conventionally considered as “non-druggable” via their orthosteric ligand binding site. Access to high quality cells has become an essential factor for the successful FLIPR HTS assays. The use of pre-prepared frozen cells uncouple cell production © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim from compound screening, increasing the flexibility in scheduling assays and thereby minimizing a major source of variability in the assay. A major focal point during the meeting was RNA interference (RNAi), which has driven the fast expansion of HCS technology especially in academic research environments [5]. In the words of Gwen Farwell (OpenBiosystems), HCS and RNAi is a marriage made in heaven. The development of large-scale RNAi application to functional genomic analysis has opened the way to silence gene expression without establishing preconceived ideas about the consequences. Image-based phenotypic assays that simultaneously analyze different parameters enable the effects of silencing on cell size, shape, cell proliferation, cell cycle, cell migration and even more subtle morphological features, to be assayed at a single-cell level. The use of several commercial vector-based shRNA and synthetic siRNA libraries in high content image-based screens was presented. As a part of the commercial agenda J. Alba (Bio-Rad), M. Brown (Thermo Fisher Scientific), H. J. George (Sigma-Aldrich) and G. Fewell (Open Biosystems) outlined the tremendous progress made with the efficacy and specificity of RNAi reagents. R. Beijersbergen (The Netherlands Cancer Institute, Amsterdam, The Netherlands) and colleagues have used vector-based shRNAs and synthetic siNAs for genotype-specific anti-cancer target and drug discovery. In the second generation of a retroviral NKI shRNA library, a selective marker was introduced in close proximity to the hairpin cassette to prevent recombination events [6]. The presence of a unique molecular bar code serves to identify the silenced gene by PCR amplification in a pooled, polyclonal screen. The relative abundance of each individual retroviral vector in a cell population can be measured using oligo-microarray hybridization. This experimental procedure provides the opportunity to passage and select cells and, hence, it is 939 BTJ-FORUM A. Niederlein (Max Plank Institute for Molecular Cell Biology and Genetics, Dresden, Germany) argued that the human brain is still the most efficient image analysis device we know, and she discussed what requirements are needed for automated high content image analysis solutions in a high throughput environment. The possibility to reduce images to numbers permits a quantitative evaluation of phenotypes. Therefore, data analysis of multiple parameters at a single-cell level opens new horizons for target discovery and our understanding of complex intracellular mechanisms. Commercial software applications have been used to analyze large image sets from experiments investigating lipid metabolism, monitoring the number, size, clustering and localization of lipid droplets. Endocytosis was studied using custom designed algorithms to track vesicle movement, enabling small circular or elliptic structures to be tracked. Urban Liebel (KIT – Karlsruhe Institute of Technology, Karlsruhe, Germany) discussed the importance of a multi-disciplinary background to efficiently run a HCS platform. Novel high throughput microscopes and tools have efficiently addressed challenging biological questions that require hundreds or even several thousands of experiments. The development of a modular and open HCS platform has been presented, which integrates liquid handling robotics, readout devices, data handling, image analysis and bioinformatics data integration in a LabView-based software environment. The open HCS platform technology allows rapid integration of novel developments (e.g., fish embryo screens, special readout devices, sophisticated specimen containers, advanced data processing, etc.). Hardware and software modules have been developed over recent years for several tricky assays, and several of these components have been implemented in a commercial software package (Scan® software, Olympus, 48-h livecell array imaging, large-scale data handling, cell-detecting auto-focus, www.biotechnology-journal.com BTJ-FORUM Biotechnology Journal not limited to short-term readouts. Conversely, arrayed assays allow straightforward identification of the gene that is knocked down. This RNA interference library has been used to identify genes whose inactivation affects the p53 tumor suppressor pathway [7]. Furthermore, R. Beijersbergen reported on a high content analysis approach to screen for molecular factors that were involved in the replication of salmonella. To discover targets associated with angiogenesis C, Weiss-Haljiti (Cenix BioScience, Dresden, Germany) reported a large-scale RNAi experiment using a custom-designed library of 17 181 synthetic siRNAs targeting 5664 human genes that were selected for their therapeutic relevance. Weiss-Haljiti pointed out that one of the major challenges that HCS technology faces is the adoption of more disease-relevant cell systems. Human primary cells resemble their in vivo counterparts better than immortalized cells. However, primary cells have proven notoriously difficult to procure, maintain and transfect, and they are prone to phenotypic drift. The large-scale loss-of-function angiogenesis screen at CENIX BioScience has been conducted using commercially available primary HUVEC cells that possess the ability to form vessel-like tubular structures in vitro. Primary HUVEC cells usually undergo senescence between 10 and 15 passages. Thus, the key to the success in this experimental system is the very careful quality control standardizing thawing procedures, amplification, passage number and cell seeding. The phenotypic changes cells went through after gene knockdown were tracked by combining four fluorescent markers to monitor cell proliferation, apoptosis, mitosis, cellular and nuclear size and shape. M. Stöter (Max Plank Institute for Molecular Cell Biology and Genetics, Dresden, Germany) introduced RNAibased screens aimed at the identification of regulators of the endocytic pathway. Viral infection screens were conducted using vesicular stomatitis 940 View publication stats Biotechnol. J. 2007, 2, 938–940 virus (VSV) and Simian virus 40 (SV40) that hijack clathrin- and caveolae/raftmediated endocytosis to infect host cells, respectively. Out of the 590 human kinases screened, 208 were identified as being involved in endocytosis, 92 were associated with the VSV pathway, 80 with the SV40 pathway and 32 with an overlapping profile [8]. Validation experiments were carried out tracking the uptake of fluorescent markers, and via morphological assessment of early and late endosomes. Recently, a similar strategy based on the endocytic uptake of fluorescently tagged epidermal growth factor (EGF) and transferrin (Tfn) was used to perform a genome-wide endocytosis RNA interference screen. Selective cargo sorting was analyzed in downstream validation experiments using multi-parameter subcellular endosome distribution assays. Acquiring 12 images from three channels per well produced almost 40 terabytes of data from 600 plates. Indeed, M. Stöter stressed the vital importance of efficient data management, and to store, analyze and visualize the data, a server-based software solution was implemented. The Laboratory Information Management Systems (LIMS) is a platform-independent system to integrate, browse and analyze large data sets and high content images. Thus, the meeting “High Content Analysis, Spain” highlighted the multidisciplinary and innovative approaches that tackle the major challenges faced by HCS technology, including data management and storage issues, the development of disease-relevant cellular models, and the building of common standards within the HCS field. Although not described in this overview, several practical sessions provided participants with the opportunity to familiarize themselves with different HCS platforms. Furthermore, there were exhibitions and poster presentations on several methodological aspects such as assay design, novel dyes and software solutions. We anticipate that HCS technology will have major im- © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim pact on future cell biology and drug discovery. By Urban Liebel1 and Wolfgang Link2 1Institute of Toxicology and Genetics, KIT, Karlsruhe Institute of Technology, Eggenstein-Leopoldshafen, Germany E-mail: ul@itg.fzk.de 2Experimental Therapeutics Program, Centro Nacional de Investigaciones Oncologicas (CNIO), Madrid, Spain E-mail: wlink@cnio.es References: [1] Lang, P., Yeow, K., Nichols, A., Scheer, A., Cellular imaging in drug discovery. Nat. Rev. Drug Discov. 2006, 5, 343–356. [2] O’Brien P, J., Irwin, W., Diaz, D., HowardCofield, E. et al., High concordance of druginduced human hepatotoxicity with in vitro cytotoxicity measured in a novel cellbased model using high content screening. Arch. Toxicol. 2006, 80, 580–604. [3] Perlman, Z. E., Slack, M. D., Feng, Y., Mitchison, T. J. et al., Multidimensional drug profiling by automated microscopy. Science 2004, 306, 1194–1198. [4] Loo, L. H., Wu, L. F., Altschuler, S. J., Imagebased multivariate profiling of drug responses from single cells. Nat. Methods 2007, 4, 445–453. [5] Haney, S. A., LaPan, P., Pan, J. Zhang, J., High-content screening moves to the front of the line. Drug Discov. Today 2006, 11, 889–894. [6] Bernards, R., Brummelkamp, T. R., Beijersbergen, R. L., shRNA libraries and their use in cancer genetics. Nat. Methods 2006, 3, 701–706. [7] Berns, K., Hijmans, E. M., Mullenders, J., Brummelkamp, T. R. et al., A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 2004, 428, 431–437. [8] Pelkmans, L., Fava, E., Grabner, H., Hannus, M. et al., Genome-wide analysis of human kinases in clathrin- and caveolae/raftmediated endocytosis Nature 2005, 436, 78–86. [9] Zanella, F., Rosado, A., Blanco, F., Henderson B. R. et al., A HTS approach to screen for antagonists of the nuclear export machinery using high content cell based assays. Assay Drug Dev. Technol. 2007, in press. This contribution was peer reviewed Received 31 May 2007 Revised 4 July 2007 Accepted 5 July 2007