2019 POSTDOC TALKS
Postdoc talk 1
Integrating clinical and –omics data to identify predictive markers of treatment response
Francesco Brizzi & Antoine Soubret
Roche, Basel, Switzerland
In clinical trials, patients are randomized to a treatment (or to a placebo) and their clinical status (often measured by a score) is monitored over time to understand the effect of treatment. Baseline patient characteristics (such as sex and age) are usually correlated with the clinical score to understand whether treatment is most effective within certain subpopulations. In recent years, genetic data (e.g. transcriptomic, proteomic) at baseline have been increasingly collected as part of clinical trials; such data are extremely large and noisy. Hence it is not clear how the correlation between this large number of markers and the longitudinal clinical score can be established. We have thus developed an innovative multi-omics approach that combines statistical (longitudinal) modelling of clinical data with machine learning variable selection methods to identify a small number of markers that are most predictive of both baseline disease activity and treatment response. Such markers could potentially allow us to identify responder subpopulations and to devise personalised treatment strategies. We illustrate how our methodology can be applied on both simulated data and data from a clinical trial on early rheumatoid arthritis.
Postdoc talk 2
Bridging the protein cavity and chemical spaces
Jean-Rémy Marchand
Novartis, Basel, Switzerland
Protein cavities are often associated with protein function, and are the target of most of the low molecular weight drug-like molecules. We are designing a novel algorithm to identify, characterize, and compare cavities, relying on a new fingerprint description of the pockets. This fingerprint retains three- dimensional data, allows for cavity partial matching and is very efficient, making quick comparisons of large number of binding pockets computationally amenable. In addition, this pharmacophore-based fingerprint is transferable to the ligand chemical space, which fills the gap between the cavity and ligand spaces. The possible applications include, for example, the prediction of potential off target activities and polypharmacology, compound target identification, structural pocket-based classification of proteins, or ligand screening.
Postdoc talk 3
Cores and loss of muscle strength in mice carrying recessive RYR1 mutations: an animal model for RyR1 linked congenital myopathies
M. Elbaz
Departments of Anesthesia and Biomedicine, Basel University
Mutations in RYR1, the gene encoding the RyR1, are associated with several neuromuscular disorders including core myopathies, congenital fiber type disproportion, centronuclear myopathy and the pharmacogenetic disorder malignant hyperthermia. Although the functional effect of dominant RYR1 mutations have been extensively studied, the consequences of recessive mutation have not been investigated. Using the CRISPR/CAS9 technology we created a mouse model knocked-in for two mutations identified in a severely affected multiminicore patient, namely the RYR1 RyR1Q1970fsX16+A4329D mutations. The mutant mice are viable, but during the first 20 weeks after birth their body weight was on average 20% lower than that of WT or single RYR1 mutant heterozygous littermates. In addition, muscles from mutant mice contained “cores” characterized by severe myofibrillar disorganization associated with misplacement of mitochondria. Spontaneous running distance of the compound heterozygous mice was 50% lower than that of WT littermates. Furthermore, their muscles developed less force, and had smaller electrically evoked calcium transients. Mutant RyR1 channels incorporated into lipid bilayers exhibited lower calcium sensitivity and no changes of the calcium conductance. In conclusion, the results of our study show that the phenotype of the RyR1RyR1Q1970fsX16+A4329D compound heterozygous mice largely recapitulates the clinical picture of patients harboring recessive RYR1 mutations. The compound heterozygous RYR1 mutant mouse can be exploited for preclinical studies aimed at the treatment of recessive RYR1-linked congenital myopathies.
Postdoc talk 4
Filling the gaps: Roles of glyceraldehyde 3-phosphate dehydrogenase homologs in Pseudomonas aeruginosa metabolic fluxes
Klotz Alexander
Biozentrum, University of Basel, Basel, Switzerland
A highly versatile and tightly regulated metabolism allows Pseudomonas aeruginosa to adapt to various environmental conditions, infect a wide range of hosts and successfully dwell on different host tissues. However, little is known about the role of the central carbon metabolism and its complex regulation within the host. Here, we analysed distinct physiological functions of three paralogs of glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a key enzyme of the central carbon metabolism. Preliminary experiments indicated that while GapA is bi-functional, GapB and GapC primarily catalyse glycolytic and gluconeogenic reactions, respectively. To dissect the role of the single GAPDH, transcriptional reporters were constructed allowing us to examine the carbon fluxes in vivo and at single cell level, notifying expression differences during growth phases and medium, used for growth. In contrast to gapA and gapB mutants, a strain lacking GapC completely lost its virulence in a Galleria mellonella animal model, indicating that in this host gluconeogenesis is of key importance for the pathogen. This study contributes to a better understanding of how carbon metabolism contributes to growth of this important human pathogen providing detailed insights into the control of central metabolic processes. The long-term goal is to use this information to assess important bacterial metabolic processes in the human patient.
Postdoc talk 5
Using Whole-Genome Deep Sequencing to Understand the Evolution and Diversity of European Enterovirus D68 Outbreaks in 2014, 2016, and 2018
Emma B. Hodcroft
Biozentrum, University of Basel, Basel, Switzerland
Enterovirus D68 (EV-D68) has attracted recent media attention due to its link with acute flaccid myelitis, a serious paralyzing condition. Using near full- length genome (NFLG) sequences from 54 samples obtained in Sweden during EV-D68 outbreaks in 2014 and 2016, we investigated the introduction, diversity, and evolution of the virus during these outbreaks. We estimate a strong inter-patient temporal signal (3.8×10-3 subs/site/yr), and found evidence that the EV-D68 was introduced to Sweden multiple times during the 2016 outbreak. The 2016 B3-subclade does not appear to have arisen directly from the 2014 B1-subclade, however, and their most recent common ancestor dates to 2009. Following on from this, we have reached out to European countries, offering to do near full-length genome (NFLG) sequencing on EV- D68 samples from 2018 to better understand the current outbreak. Three sequences from Stockholm from September/October 2018 have successfully already been generated, 20 further samples from Belgium and the Netherlands are being sequenced and samples from Spain and Switzerland are being collected. Unpublished reports on partial sequences (VP1) from 2018 elsewhere in Europe suggest the samples fall within the B3-subclade, and phylogenetic analysis of the three NFLG Swedish 2018 sequences confirms this (Fig 1). Interestingly, they do not root among the most recent 2016 outbreak samples, but deeper in the B3 subclade, and while two of the sequences are very similar genetically, the third is distinct, suggesting at least two separate B3 lineages in Stockholm. Here, we will perform a comprehensive analysis of all available NFLG European 2018 EV-D68 sequences, as well as further analysis using only the VP1 segment, allowing inclusion of 1000 more publicly available background sequences.
Postdoc talk 6
Teaching old drugs new tricks: Repurposing Acute Medications for Enhanced Recovery after Spinal Cord Injury
Catherine Jutzeler
Department of Biosystems Science and Engineering, ETH Zurich, Basel, Switzerland
Currently, there is no cure for people with a spinal cord injury (SCI). This means that lost sensation and muscle strength rarely fully return, and people with SCI are left with severe, life-long disabilities. Trauma to the spinal cord is immediately coupled with various secondary health conditions (infections, spasticity, pain) necessitating the administration of a plethora of medications. Surprisingly, little is known about the impact – so-called disease-modifying effects- of these medications on functional recovery, e.g., walking function. Concomitant medication log records of failed or otherwise terminated clinical trials are an excellent data source to answer this question. Typically, concomitant medication log records are collected during clinical trials to assess pharmacological interactions with the investigational medication. Leveraging data from the largest failed clinical trial in the field of SCI, we identified 500 unique medications that were administered to the 797 enrolled patients within the first month after injury. On average patients received 20 unique medications – often in a combinatorial or overlapping fashion. Moreover, we discovered that patients with SCI, who receive anticonvulsants, achieve greater returns of muscle strength compared to patients who receive other pain medications. The notion that some medications give rise to unintended benefits introduces the exciting concept of drug repurposing, whereby concomitant medications are repositioned as therapies to enhance functional recovery following SCI. Conclusively, our study demonstrates that in the interim of developing more effective interventions, evaluating current standards of acute care represents an intriguing alternative strategy to improve the lives of people with SCI.
Postdoc talk 7
TNF-α signaling employs STING to promote interferon and interferon-stimulated gene induction
Joschka Willemsen
Autoimmunity, Transplantation, and Inflammation Disease Area, Novartis Institutes for BioMedical Research, Basel, Switzerland
Autoimmune diseases such as rheumatoid arthritis (RA), Crohn’s disease and psoriasis are strongly associated with TNF-α signaling. The importance of TNF-α in these diseases is demonstrated by the successful treatment with TNF-α antagonists. In addition, these diseases are usually associated with a type I interferon signature. However, the source of the interferon is not entirely clear. In fact, prolonged TNF-α treatment in vitro was demonstrated to induce interferon and lead to sustained expression of interferon related chemokines such as CXCL10. In order to understand how TNF-α leads to interferon production we analyzed cells that are deficient for key members of the interferon induction pathways (MAVS, cGAS, STING, MyD88, TBK1 and IRF3). Using this approach, we have identified Stimulator of Interferon Genes (STING) and Cyclic GMP-AMP synthase (cGAS) as a critical factor in the TNF-α mediated interferon-stimulated gene induction. The involvement of STING downstream of TNF-α signaling was further validated in primary human dermal fibroblasts and primary bone marrow derived macrophages. In addition, knockout of several components of the TNF-α signaling cascade, demonstrated the requirement of TNFR1 and TNFR2 signaling for interferon induction. In conclusion, our results identified a new link between TNF-α and the cGAS/STING signaling pathway, which will help to understand the induction of interferon in TNF-α driven diseases.
Postdoc talk 8
Single-cell-RNA-sequencing-coupled lineage tracing in cerebral organoids
Rebecca Petri
Lab of Quantitative Developmental Biology, Department of Biosystems Science and Engineering, ETH Zurich
3D cell culture systems derived from human embryonic stem cells or induced pluripotent stem cells (iPSCs)– so-called organoids- mimic the in vivo architecture and the multilineage differentiation of actual human tissues. These 3D structures allow the manipulation and analysis of human tissue development and the generation of patient-derived cell culture models in order to study complex human diseases. Single-cell RNA-sequencing (scRNA-seq) has been an important tool to analyse cell types and developmental trajectories in organoid systems and to investigate their ability to mimic their in vivo counterpart. scRNA-seq however only gives a snap-shot of the cell state and it remains challenging to reconstitute cell lineages and to analyse when and how lineages are established during development. We therefore established a technique that allows us to capture lineage information and cell identity using single-cell RNA sequencing in cerebral organoids. We label iPSCs with a highly complex barcode library to distinguish different clones and further culture them to generate cerebral organoids. At different stages of organoid development, we induce CrispR/Cas9 genetic scarring to label progenitor cells that will give rise to distinct mature cell types with additional lineage barcodes. Together with the transcriptome profiles of single cells, we can now use the barcodes and genetic scars to identify distinct cell types and reconstruct developmental lineages in cerebral organoids at high resolution. This technique therefore represents an opportunity to trace cell lineages in various 3D cell culture systems and opens up the possibility of dissecting changes in cell lineages in patient-specific models of disease.
Postdoc talk 9
Antibody bivalency can substitute for Fc functions in antiviral protection
Mehmet Sahin
Department of Biomedicine – Haus Petersplatz, Division of Experimental Virology, University of Basel, Basel, Switzerland
Across the animal kingdom, multivalency discriminates antibodies from all other immunoglobulin superfamily members. The evolutionary forces conserving multivalency above any other structural hallmark of antibodies remain, however, ill-defined. Here we have used antibody engineering techniques to investigate mechanisms of protection of neutralizing and non- neutralizing antibodies (nAbs, nnAbs) in a viral infection model in mice. Antibody bivalency enabled the tethering of virions to the infected cell surface, thereby inhibiting the release of infectious virions in cell culture and suppressing viral loads in vivo independently of Fc. Conversely, virion release inhibition and in vivo control of infection by monovalent antibody formats were strictly Fc-dependent. Comparable amounts of nAb and nnAb were required for in vivo protection and correlated with similar virion release-inhibiting activity. These observations provide a mechanistic understanding of the evolutionary conservation of antibody bivalency and may help establishing better correlates on nnAb protection for vaccine development.
Postdoc talk 10
THE HUNT FOR HIGH AFFINITIES – BUT CAN WE EVEN MEASURE THEM?
Christina Lamers
Molecular Pharmacy University of Basel, Basel, Switzerland
In drug development, the first aim in structure-activity-relationship studies is to identify the most potent derivative of a certain compound scaffold. Especially in the case of peptides, high affinities in the picomolar range can be achieved. But do we have suitable methods to determine affinity values accurately in that range to distinguish the activity of highly potent compounds? The pharmaceutical development of Cp40, a picomolar inhibitor of the complement cascade, has confronted us with this question. The complement system serves in blood circulation as “first line of defense” against injurious stimuli and invaders. Upon activation, a series of cascading enzymatic reactions lead to amplification of the signal and to pathogen clearance and opsonic cell killing. Yet complement has also gained increasing interest as a potential drug target, since it may be inadvertently triggered and contribute to clinical complications in the pathogenesis of various autoimmune, inflammatory and age-related diseases as well as transplant rejection. While the involvement of dysregulated complement activation in inflammatory and autoimmune diseases is now widely recognize, so far only one complement- specific drug has reached the market. In this presentation we reflect on the development of the picomolar complement inhibitor Cp40, the optimization of its pharmacokinetic profile and the journey to an accurate affinity characterization of its derivatives, applying several biophysical methods.
Postdoc talk 11
Scaling up image analysis for microfluidics experiments: take the human out of the loop
Michael Mell
Biozentrum, University of Basel, Basel, Switzerland
To study gene regulation in single cells, the Mother Machine (MM) is a well- established microfluidic design where bacteria are trapped and grown in line in single-ended channels. Their progeny will continuously be pushed out at the open end of the channel, which enables studying cell growth and gene expression in the same single-cell using quantitative timelapse fluorescence microscopy. Our group has recently open-sourced [1] a design for a dual-input Mother Machine (DIMM) chip that enables controlled variation of external conditions and direct observation of gene regulatory responses to changing conditions in single cells. The accompanying software (MoMA for the Mother Machine Analyzer) enables segmenting and tracking cells with high accuracy, while requiring minimal curation. However, continuous improvement of our microfluidic designs has opened the door to studying several strains and conditions in parallel, but also makes human curation increasingly intractable. Here we will present ongoing work on MoMA to improve its segmentation and tracking using deep learning with the goal of taking humans out-of-the-loop of MM data processing.