Report of the Second European Zebrafish Principal Investigator Meeting in Karlsruhe,Germany,March 21–24,2012

Introduction

The end of March, 2012, saw the second European Zebrafish Principal Investigator (PI) Meeting (EZPM) being held in Karlsruhe. This meeting followed the success of the inaugural EZPM meeting that took place in Padova, Italy in 2010, and which launched a biannual series of meetings. The primary goal was to present the latest results and also to discuss key strategic issues faced by European researchers who use fish genetic models such as zebrafish and medaka. Basing this particular meeting in Karlsruhe had additional significance this year because it coincided with the launch of the European Zebrafish Resource Centre (EZRC). The EZRC is located at Karlsruhe Institute of Technology (KIT) and will serve both as a European stock center for zebrafish and medaka lines, as well as providing infrastructure, expertise, and panels of mutant fish for support of major screening projects. Thus in more ways than one, the second EZPM represented an important step forward for this rapidly growing branch of European research. The meeting was organized by Jochen Wittbrodt (Heidelberg University, Germany), Nicholas S. Foulkes and Gudrun Knedlitschek (both of Karlsruhe Institute of Technology, Germany), and was held over 3 days. It hosted 120 attendees invited from all over Europe and representing a wide range of topics. Short presentations were divided into seven sessions: 1) Omics and networks; 2) Injury and repair; 3) Disease models; 4) Morphogenesis and organ formation; 5) Behavioral and toxicological analysis; 6) Cellular polarity and dynamics; and 7) Image processing in zebrafish. Thus, the second EZPM provided a valuable snapshot of the current state of the art in fish-based research in Europe.

Omics and Networks

This session illustrated the use of various technologies developed to systematically characterize the regulome, the transcriptome, and protein interactions of the zebrafish. Since the quality of the zebrafish genome sequence has now reached a high level, these techniques are becoming increasingly important in the zebrafish field. Thus, several talks provided fundamental insights into the zebrafish genome and its regulation. Ferenc Mueller (Birmingham, United Kingdom) reported a precise mapping of all transcription start sites in the zebrafish genome. Peter Aleström (Oslo, Norway) summarized his results on transcript isoform dynamics around the mid-blastula transition, and Silke Rinkwitz (Sydney, Australia) presented a detailed characterization of the cis-regulatory elements linked to the arx locus. Bernard Peers (Liège, Belgium) presented a new approach that will certainly open up possibilities for finally mapping and cloning many mutants that still float around in labs all over the world. This method is based on the use of next generation sequencing for mutation mapping. Since his approach relies on whole genome rather than exome sequencing, this technology offers also the very attractive potential to discover mutations in cis-elements and regulatory RNAs that were under-represented in previous screens.

While these DNA- and RNA-based omics technologies are in full swing, proteomics is still in its infancy in zebrafish research. Most impressive pioneering work was reported by Christian Soellner (Tübingen, Germany) who identified extracellular protein interaction networks important in patterning the ventral neural tube of the zebrafish. This clearly shows the potential of proteomics approaches to obtain new insights into the molecular mechanisms underlying development. However, concerns were raised that the sensitivity of current technology would not allow efficient detection of regulatory proteins normally present at low abundance in cells.

Injury and Repair

Unlike mammals, teleosts have an extraordinary capacity to regenerate damaged tissue. The Injury and Repair session presented the latest findings on neural, cardiac, and fin regeneration in the zebrafish. We also learned about the mechanisms of muscle cell repair and how regeneration is influenced by circadian rhythms. A key challenge in the field is to elucidate the origin of regenerated cells and tissues. Are they derived from resident stem cells, do terminally differentiated cells regain pluripotency, or do preexisting cells only partially dedifferentiate in order to proliferate and replenish the lost tissue? Two talks addressed this issue. Using genetic lineage tracing, Gilbert Weidinger's group (Dresden, Germany) found that dedifferentiation from mature osteoblasts drives bone regeneration upon fin amputation. Thus, contrary to previous models, this work shows that newly forming bone is derived from differentiated osteoblasts, which retain a memory of their origin. In a second talk, Nadia Mercader (Madrid, Spain) described the fate of epicardial-derived cells during heart regeneration. This heterogeneous cell population does not contain myocardial progenitors but instead contributes to the transient fibrotic scar and participates in the revascularization of the injured myocardium. Christopher Antos and co-workers (Dresden, Germany) described Simplet (smp) as a novel gene involved in β-catenin-dependent Wnt signaling that is central to regeneration and embryogenesis. Through gain and loss of function studies, they showed that smp controls the nuclear localization of β-catenin. The cellular response to tissue damage is also controlled by environmental signals, as shown by Daniela Vallone (Karlsruhe, Germany). During fin regeneration, proliferation of the epidermis covering the blastema is light sensitive, with highest proliferation rates occurring close to the light–dark transition, possibly reflecting an evolutionary mechanism to minimize UV-mediated DNA damage in the skin. The talk by Uwe Strähle (Karlsruhe, Germany) moved from tissue regeneration to cell repair. Muscle cells have a high repair capacity, which is lost in certain diseases such as muscle dystrophy. Using high-resolution in vivo imaging, the team showed that the sarcolemmal repair of muscle cells proceeds via a specific sequence of events whereby distinct annexin subtypes are recruited to the injury site. Finally, Francesca Peri (Heidelberg, Germany) illustrated what happens in situations and cell types where repair is not possible. Her group found that glutamate released from neurons establishes a Ca²⁺ gradient that rapidly attracts phagocytic microglia to brain injuries, thus describing a molecular and morphogenetic mechanism by which dying cells are eliminated.

Disease Models

The Disease Models Session included a series of talks highlighting the use of zebrafish to elucidate molecular mechanisms underlying human pathology. Selected presentations ranged from examples of infectious diseases to disorders with either known or unknown genetic origin. Corinne Houart (London, United Kingdom) opened the session by demonstrating how zebrafish could model distinct neurodegenerative diseases such as spastic paraplegia, amyotrophic lateral sclerosis, and fronto-temporal dementia, allowing detection of functional links among the causative proteins, with particular emphasis on their role in RNA regulation. Linked to the neurodegeneration research field, Roland Dosch (Göttingen, Germany) proposed an interesting use of zebrafish mutants with defects in yolk endocytosis, to elucidate the cellular basis of the human genetic disorder hereditary spastic paraplegia (HSP). For example, in souffle mutants he identified the defective gene as an endosomal protein corresponding to human SPASTIZIN. These mutants, with their early phenotypes, appear highly promising for drug screening, where currently no treatment exists, and to identify novel candidate genes for HSP. Alyson MacInnes (Utrecht, The Netherlands) focused instead on ribosomopathies, a rather unexplored class of human disorders based on abnormalities in ribosome structure and function. She showed how analysis of ribosomal protein loss in zebrafish could help to recapitulate some common features, such as cytopenia, bone marrow failure, and predisposition to cancer. These studies successfully identified a p53-independent cell death mechanism, linked to autophagy, which may play a major role in the pathogenesis of diseases affecting ribosome biogenesis. Manfred Schartl (Würzburg, Germany) took advantage of medaka genetics to model human melanoma. In his presentation, medaka lines with different backgrounds were shown to make easier the identification of tumor “up” and “down” modifiers. Moreover, transcriptome sequencing of medaka melanoma could be compared with zebrafish and human datasets, allowing more robust identification of new melanoma markers and molecular signatures. In the last talk of the session, Annemarie Meijer (Leiden, Netherlands) presented the “FishForPharma” Marie-Curie Initial Training Network, comprising research groups and Biotech/Pharma partners, aiming to exploit zebrafish for infection studies and anti-microbial drug discovery. To improve high-throughput screens, a novel automatic injection system was illustrated. Moreover, a series of recently available tools were described, for fully exploiting the potential of the zebrafish model, including antibodies, reporter lines, and transgenic strains for immune cell types.

Morphogenesis and Organ Formation

The talks in this session tackled morphogenesis from different perspectives, and studied very diverse processes, from organ formation during embryogenesis to adult tissue homeostasis. However, all of them highlighted the importance of generating appropriate tools to visualize, quantify, and model cellular behavior in vivo. Julien Vermot (Strasbourg, France) and Daniela Panakova (Berlin, Germany) both specialize on the impact of mechanical forces and electrochemical cues on morphogenesis. Combining high-resolution imaging, embryo manipulation, and mathematical modeling, they analyzed either the influence of hemodynamic forces or electrical conduction upon the development of the cardiovascular system. They especially focused on local differences within the developing vasculature or heart.

To investigate the role of cellular interactions in morphogenesis, transgenic lines allowing for the detailed observation, as well as manipulation of cellular or tissue behavior are key tools. Using such lines, Markus Affolter (Basel, Switzerland) presented the detailed cellular rearrangements necessary for vessel fusion and for the subsequent reorganization of the newly formed unicellular tube into a multicellular tube, while Elke Ober (London, United Kingdom) focused on how EphrinB/EphB signaling regulates endoderm–mesoderm interactions during liver development.

Matthias Carl (Heidelberg, Germany) used two-photon microscopy to image about 300 microns deep into the developing CNS and to observe the development of the habenular (neural) circuit: from neuron birth until the axons reached their target, impressively over the course of 5 consecutive days. They applied a new algorithm to color-code the dorso-ventral position of each neuron and its axon, allowing the visual integration of positional information with high-resolution long-term imaging. The group of Claudia Linker (London, United Kingdom), studying neural crest migration, has developed algorithms for automated 3D tracking of cells permitting quantitative analysis of cell migration parameters such as speed, persistence, directionality, and cell division. It is now paramount for the community to generate ways of sharing these data sets allowing their wide application, and importantly setting standards for future quantitative and qualitative analysis of in vivo imaging data.

Once organogenesis is complete, tissue homeostasis must be maintained throughout the life of the organism. Certain tissues in our body require the continuous addition of new cells throughout life, and new cells must also be generated upon regeneration after injury. Laure Bally-Cuif (Gif-sur-Yvette, France) and Catherina Becker (Edinburgh, United Kingdom) presented recent work in their laboratories that exploit the imaging and gene manipulation advantages of the zebrafish to learn about the delicate balance of signals required to modulate proliferation both in the stem cell niche, and during regeneration after tissue injury. Interestingly, these studies uncovered the essential role of the Notch pathway in both processes. They also highlighted our lack of appropriate genetic tools to abrogate gene function conditionally in the adult and the need to invest time and resources to generate such tools.

Behavioral and Toxicological Analysis

Research into the genes and neural circuits that underlie behavior represents one of the most exciting up-and-coming fields of fish-based research. The behavioral and toxicological analysis session showcased zebrafish as a model organism for behavioral genetics, including studies of behavior at different life stages, high-throughput screening, and the ability to measure both complex cognitive behavior and social interaction in a vertebrate.

Marc Muller (Liège, Belgium) demonstrated the power of using zebrafish larvae to analyze the impact of nonlethal concentrations of toxins on behavior. Locomotor-based read-outs of larvae were used to define behavioral parameters in a principal component analysis. This toxicological analysis allows the clustering of chemical compounds according to their behavioral profile, and may be used to study the environmental impact of common pollutants, including pesticides and heavy metals.

The development of zebrafish embryos outside of the mother allows the genetic pathways underlying drug abuse to be examined without any confounding interference from maternal influences. Raquel Rodriguez (Salamanca, Spain) demonstrated that morphine and cocaine, common drugs of abuse, act via a mechanism that includes the microRNA mir-133b in order to perturb early embryonic dopaminergic neuron development.

Caroline Brennan (London, United Kingdom) unveiled both a screen for reward pathway mutants (where the distribution of a fish in a population allows its genotype to be deduced) and a 3-choice serial reaction time task (3CSRTT) assay for adult zebrafish impulsivity. The 3CSRTT can be extended to include five choices, and will now be used to screen mutant lines that have been generated by the Sanger centre TILLING project. Finally, Gonzalo de Polavieja (Madrid, Spain) presented software that can accurately track groups of adult fish in a tank and so is ideal for studies of social behavior. The software is able to identify each fish definitively in a group, based on small morphological differences, allowing error-free tracking of up to 20 individuals. These results were used to mathematically model the rules that govern social interactions among individuals in a group of fish. Together, these talks demonstrate that the future of zebrafish as a model for behavioral genetics looks bright.

Cellular Polarity and Dynamics

The session on cellular polarity and dynamics highlighted current efforts to obtain an integrative view of signaling pathways extending from organ-specific tissue architecture to the ultimate visualization of downstream response elements.

Felix Loosli (Karlsruhe, Germany) illustrated how intracellular signaling pathways ensure the maintenance of tissue morphology and cell polarity in the medaka eye. Starting from an eye pigmentation mutant identified in a forward genetic screen, he identified a novel link between a regulator of the actin cytoskeleton and apico-basal polarity in the fish retina. Sylvie Schneider-Maunoury (Paris, France) addressed the question of how the organization of individual cells is integrated in the context of overall tissue polarity. Through a study of zebrafish cilia, her work sheds new light on the molecular mechanisms through which spatial cues provided by the planar cell polarity pathway are used to ensure the proper subcellular positioning of these important signaling organelles.

Two presentations documented strategies for using zebrafish to visualize endogenous signaling molecules in the context of an intact living organism. Graham Lieschke (Clayton, Australia) talked about the mechanisms used by leukocytes to regulate the gradient of hydrogen peroxide generated after wounding by using in vivo imaging approaches and an ENU-mutant generated in their forward genetic screen. Maximilian Fürthauer (Nice, France) is studying the importance of endocytic trafficking for the regulation of Delta/Notch signaling in the context of zebrafish neurogenesis. In order to visualize this process in vivo, he established a novel live imaging assay to track endogenous signaling molecules on their journey through the endo-lysosomal system.

The last two presentations concentrated on imaging signal transduction activity in vivo with the help of transgenic zebrafish reporter lines. Francesco Argenton (Padova, Italy) described the generation of fluorescent reporter lines for several major developmental signaling pathways that can be used to follow signaling activities in cell proliferation, stem cell maintenance and differentiation, tissue homeostasis, and cancer. These lines provide valuable tools for mutant phenotyping, drug screening, and monitoring the activities of specific pathways in tumors. Thomas Dickmeis (Karlsruhe, Germany) presented the development of transgenic zebrafish expressing a reporter gene under the control of a glucocorticoid-responsive promoter. The performance of the reporter line was shown to be comparable to that of conventional cell-based assays. Additionally, the detection of effects dependent on compound metabolism by the larval organs was demonstrated. The line shows great potential to be used in stress research, environmental monitoring, and pharmaceutical screens.

Image Processing in Zebrafish

This session including four talks and related posters provided an overview of recent activities in automated image processing and analysis. It encompassed the complete range from methodology to zebrafish-specific applications in high-throughput and high-content screening.

New methods for the labeling and automatic classification of images (Raphaël Marée, Liège, Belgium) were presented in addition to web-based platforms for the easy access to algorithms for external users (R. Marée and Wolfgang Driever, Freiburg, Germany). As a practical example of high-throughput screens, Herman Spaink (Leiden, The Netherlands) presented results for the analysis of fluorescence-labeled reporter genes in disease models using robotic microinjection and a commercial biosorter. The potential of high-resolution screens was shown for the landmark-based registration of experimental confocal data stacks and integration into a three-dimensional standard brain model using different fluorescence marked gene or mRNA expression patterns (W. Driever, Freiburg, Germany). The largest datasets were generated by 3D+time microscopy imaging with two-photon laser scanning or selective plane illumination microscopy and processed for reconstructing cell dynamics in embryonic development with single cell resolution (Nadine Peyriéras, Gif-sur-Yvette, France). All speakers emphasized the necessity of employing automatic image analysis to organize the data of complex screens and quantitatively analyze biological features with datasets up to the terabyte range.

The reuse of existing software tools by other groups to avoid redundancy in methods and tool development were among central issues of common interest in the following discussion. Specifically, this included a consideration of aspects of future web platforms, data storage, license policies, and application-specific adaptation of existing software.

Strategic Issues

The importance of the EZPM meeting was not only to present the latest experimental results but also to discuss strategic issues that were raised in the various sessions. Furthermore, two keynote lectures stimulated much discussion. The first by Jan Lohmann (Heidelberg, Germany) addressed a very relevant issue: the regulation of stem cell behavior during development as studied in the plant model, Arabidopsis. Although clearly very different from fish, plants provide fascinating insight into the general principles that govern stem cell behavior. Second, Ewan Birney (Cambridge, United Kingdom) reported on the latest progress of the human genome ENCODE project. This lecture also highlighted the great potential for using inbred lines of medaka derived from a natural population as a model for studying the genetic basis of complex traits (F. Loosli, Karlsruhe, Germany, and J. Wittbrodt, Heidelberg, Germany).

One important issue identified during the meeting was a general lack of bioinformatics expertise in zebrafish labs that represents a major bottleneck in the application of high-throughput sequencing and other omics technologies to the zebrafish field. It was suggested to apply for Marie Curie training network funding to bring together zebrafish researchers and bioinformaticians with the aim of establishing standardized techniques and expertise in zebrafish labs.

A vital issue that was raised was the need to improve the various tools available for studying zebrafish. There was discussion whether, similar to the modEncode project, functional genomics relevant data for zebrafish could be assembled in a database for public use. Furthermore, it was also recognized that there was an urgent need for a workshop for discussing how best to expand and implement the genomics resources in zebrafish and also to train scientists in their use. Such a workshop was planned for December 2012 and held in Cambridge, United Kingdom.

There was also a strong recommendation to lobby for the promotion of proteomic approaches, in particular to systematically document post-translational modifications during zebrafish development. Furthermore, there was a general agreement on the need for improved antibody tools, inviting the scientific community to share working protocols in the “Wiki” page of the ZFIN database. One important project that was launched in the meeting that may help to support these various initiatives was the ZeLP (Zebrafish E-Learning Project). This E-learning project aims to bring zebrafish researchers and bioinformatics experts together to create a digital resource encompassing all aspects of research using zebrafish. This would be accessible at various levels, from school children up to experienced researchers and would correspondingly provide different depths of information about zebrafish research that could be regularly updated.

Finally, after the meeting, a satellite workshop “Future image processing platforms” attracted more than 40 people. Here, more specific themes such as requirements for special screens, the role of metadata for information extraction, and the need for standardized data formats were addressed. To continue this work further, a new task force initiative called the Zebrafish Imaging Processing (ZIP) group has been launched. This initiative aims to provide an overview of existing tools, to develop a roadmap for future activities, and to initiate a standardization process (e.g., a module concept of single algorithms for a better operability and a workflow from the microscope to the database).

Looking forward, the newly established EZRC, together with the EZPM meeting series, will be vital for identifying and tackling many of these strategic issues. They promise to represent an important central reference point for European fish researchers. In this regard, plans for the third EZPM to be held in 2014 are already underway (for further details please consult the EuFishBioMed website (http://eufishbiomed.kit.edu/)).