Abstract
PODIUM PRESENTATION ABSTRACTS
Invited Abstracts were published in September 2016 SBI2 Special Issue and are available online at
Molecular Devices, LLC, Cellular Dynamics International
Cell models are becoming more complex in order to better mimic the in vivo environment and provide greater predictivity for compound efficacy and toxicity. There is increasing interest in exploring the use of three‐dimensional (3D) spheroids for modeling developmental and tissue biology with the goal of accelerating translational research in these areas. Accordingly, the development of higher throughput quantitative assays using 3D cultures is an active area of investigation. In this study, we have developed and optimized methods for the formation of 3D liver spheroids derived from human iPS cells and using those for toxicity assessment. We used confocal imaging and 3D image analysis to characterize cellular information from a 3D matrix to enable multi‐parametric comparison of different spheroid phenotypes. The assay enables characterization of compound toxicities by spheroid size (volume) and shape, cell number and spatial distribution, nuclear characterization, number and distribution of cells expressing viability, apoptosis, mitochondrial potential, and viability marker intensities. In addition, changes in the content of live, dead, and apoptotic cells as a consequence of compound exposure were characterized. We tested 50 compounds, including a number of known hepatotoxic compounds (e.g. anti‐cancer and antipsychotic drugs, pesticides, and others). Comparing iPSC‐derived hepatocytes and HepG2 cells in both two‐dimensional (2D) and 3D cultures, we observed significant differences in the pharmacological effects of compounds across the two cell types and between the different culture conditions. Our results indicate that a phenotypic assay using 3D model systems formed with human iPSC‐derived hepatocytes is suitable for high throughput screening (HTS) and can be used for hepatotoxicity assessment in vitro.
Hubert Tseng, William Haisler, Pujan Desai, Jacob Gage,
Nano3D Biosciences, Inc., University of Texas Health Science Center
Biomedical research has gravitated towards 3D cell culture as scientists seek cellular models and assays that represent in vivo tissue more accurately than traditional 2D monolayers. Challenges in 3D cell culture both technical challenges in formation and handling, but also analytical challenges, as the density of 3D cell cultures can make imaging difficult. Towards that end, we discuss a recently developed platform for 3D cell culture, magnetic 3D bioprinting, that addresses the technical issues of 3D cell culture. The principle behind magnetic 3D bioprinting is the magnetization of cells using a biocompatible nanoparticle assembly (NanoShuttle), which can then be rapidly aggregated into spheroids using mild magnetic forces. Magnetization occurs at the cellular level by the binding of nanoparticles to cell membranes, and spheroids can be scaled down to small sizes (<1,000 cells) for high‐throughput formats (384‐, 1536‐ well). The magnetization of the spheroid also allows for spheroids to be held in place with magnetic forces during liquid transfer, thereby improving sample retention. With this system, we can easily create representative models in vitro for research. To overcome analytical challenges that all 3D cell culture platforms encounter, particularly light and reagent penetration in 3D cell cultures, we designed and created unique assays that allow for image‐based endpoints that escape these limitations with the use of magnetic forces. This includes: wound healing in 3D rings; contraction in spheroids; beating in cardiomyocyte spheroids; and CYP induction/inhibition in hepatocyte spheroids. These endpoints are easy to image and analyze, and while they do not depend solely on fluorescent imaging for results, their label‐free and lytic nature allow for multiplexing with other endpoints. Our results demonstrate that magnetic 3D bioprinting is singularly built to engineer assays that can model cellular and tissue function in vitro for high‐ throughput and high‐content screening.
Vivek C. Abraham
AbbVie Inc.
The podocyte is a functionally and morphologically unique cell type in the mammalian kidney, where it plays a fundamental role in maintaining glomerular ultrafiltration. Perturbation of these highly specialized cells has been extensively documented in multiple chronic renal diseases, such as diabetic nephropathy, focal segmental glomerulosclerosis, minimal change disease and collapsing glomerulopathy. Therefore, cellular models of podocyte perturbation that measure disease‐relevant parameters such as cytoskeletal integrity, cell adhesion and increased oxidative stress are expected to be valuable to drive further understanding of podocyte biology and interrogate targets whose function may be manipulated to treat disease. We report the development and application of such a cellular model using a subclone of previously generated human conditionally immortalized podocytes (Saleem, et al., 2002 J Am Soc Nephrol 13: 630–638). The subcloned cells were extensively characterized by benchmarking them against freshly isolated primary human podocytes with regard to expression of podocyte‐specific proteins, phenotypic characterization of cytoskeletal and cell cycle status, and multiparametric responses to stressors such as puromycin aminonucleoside and TGF‐b, that are thought to be relevant to chronic kidney disease. We have applied this cellular model to profile the protective effects of compounds that inhibit targets such as adenosine kinase, setting the stage for further investigation in translational studies. Options for further enhancement of cellular models of podocyte injury will also be discussed.
DZNE Bonn
Neuroinflammation is an important hallmark in many neurodegenerative disorders. In the CNS innate immunity plays an important role and in particular the activation of NLRP3 inflammasome is an essential event observed under neuroinflammatory conditions. Indeed, it was demonstrated that activation of the NLRP3 inflammasome contributes to the pathology in a mouse model of Alzheimer's Disease (Heneka et. al. 2012). Recently it was shown that NLRP3 plays also a role in psychological stress where its activation is regulated via extracellular ATP and purinergic Type 2X7 receptor (Iwata et. al. 2016). Hence, evidences indicate that NLRP3 pathway might represent a new therapeutic target for pharmacological intervention in CNS and in particular in neurodegenerative disorders. However, the pathophysiological mechanisms and their associated factors are only partially understood. We have established a fully automated phenotypic screening assay using primary mouse microglia cells, which we have validated on the basis of an industrial standard. We are using our model system to monitor NLRP3 inflammasome activation and correlate this with basic cellular functions (cytokine production / receptor trafficking etc.). In our cell‐based assays, we combine the acquisition of 4 image‐based, flanked by 2 homogenous readouts. We are profiling all of our phenotypes by multivariate feature extraction and clustering methods using our high content data analysis pipeline. By using a reverse chemogenomics approach based on reference compounds we aim to identify biological activities of unknown compounds. Additionally, we developed a custom data integration / data fusion pipeline using data from diverse sources (public databases, gene expression data, screening data, clinical data) that allows us to identify associated pathways, molecular mechanisms of actions and new targets for disease modification.
Bonnie F. Sloane
Wayne State University
To define protease‐related druggable pathways that are involved in malignant progression of cancer, we have pioneered novel techniques for functional live‐cell imaging of protease activity, initially concentrating on pathomimetic avatars for breast cancer. We analyze proteolysis in the context of proliferation and formation of structures by benign and cancerous cells in 3‐D cultures over time (4D). In order to recapitulate the cellular composition and architecture of tissue, we include other tumor‐associated cells (e.g., fibroblasts, lymphatic and blood vessel microvascular endothelial cells). We also model non‐cellular aspects of the tumor microenvironment, e.g., an acidic pericellular pH. Use of these pathomimetic avatars in concert with imaging probes has allowed us to image, quantify and follow the dynamics of proteolysis in the tumor microenvironment and to test interventions that impact directly or indirectly on proteolytic pathways. To facilitate use of the pathomimetic avatars for drug screening, we have designed culture chambers with multiple wells that are either individual or connected by a bridge to allow cells to migrate between wells. Optical glass microscope slides underneath an acrylic plate allow the cultures to be imaged with an inverted microscope. Fluid ports in the acrylic plate are at a level above the 3D cultures to allow introduction of culture media and test agents such as drugs into the wells and the harvesting of media conditioned by the cultures for immunochemical and biochemical analyses. Covers contain integrated gas exchange ports and sensors to monitor oxygen levels, pH and temperature over the extended time periods in culture and to insure maintenance of such experimental conditions as hypoxia and/or a pericellular acidic pH. We predict that the pathomimetic avatars will accelerate identification of druggable pathways, screening of drug and natural product libraries and the entry of validated drugs or natural products into clinical trials.
Amee George1, Mei Wong1, Adam Stephenson1, Sheren Al‐Obaidi1, S. Peter Klinken2, Louise Winteringham2,
1Australian National University, Canberra, Australia, 2Harry Perkins Institute of Medical Research, Perth, Australia, 3Peter MacCallum Cancer Centre, Melbourne, Australia, 4ARVEC, University of Queensland, Brisbane, Australia
Apart from its role in ribosome biogenesis, the nucleolus also monitors the cell for changes, and coordinates a response if issues are detected. This ‘nucleolar stress response’ can rapidly mediate cell cycle arrest, apoptosis, or senescence depending on the cell type and severity of the insult to the cell. Often, this response requires the accumulation of p53, which occurs due to the sequestration of MDM2 by free ribosomal proteins. Recent work by us has demonstrated that this process is responsible for the apoptotic cell death of cancer cells treated with CX‐5461, a Pol I transcription inhibitor. Intriguingly, evidence also exists for the aberrant activation of this response in ribosomopathies (such as Diamond‐Blackfan Anaemia, DBA), for which treatment options are limited. Yet the precise molecular mechanism underlying the p53‐dependent mechanism remains elusive, and much detail is lacking. To address this, we have taken a novel approach by using high content imaging technology to perform genome‐wide high content loss‐of‐function (RNAi) and gain of function (ORF overexpression) screens, and have also screened compound libraries of clinically approved therapeutics, to unbiasedly identify the critical genes and pathways implicated in the nucleolar stress response due to ribosomal protein haploinsufficiency (as observed in DBA). We have uncovered a suite of novel genes and biological processes involved in this process, which we are currently validating using in vitro and in vivo models. These studies will enable the development of (i) treatments for ribosomopathies and (ii) second‐generation drugs to activate the nucleolar stress response as cancer therapeutics.
Chris Bakal
Institute of Cancer Research, London
Over the last ten years, high‐content screening has grown from an unwieldy, and sometimes error prone methodology, to a robust tool for discovery and systems‐level studies. But there have been growing pains along the way. The field of high‐content screening has had to deal with problems of data quality, analysis, and storage; as well as unexpected issues of large variability, poor reproducibility, and off‐target effects.
I will present our successes, and some failures, over the last ten years in using high‐content RNAi screening to describe the dynamics of signaling networks regulating cell morphogenesis and proliferation. Specifically, I will highlight how we have used novel statistical and computational methods to fully leverage the rich data high‐content screens can provide. Moreover, I will discuss our recent studies where we are performing high‐throughput single cell screens in 3D and/or over time. While new hardware means performing 3D and live cells is now more feasible than in the past, analyzing these screens presents new challenges. Finally, I will outline how we are integrating large image‐based data sets with genomic and proteomic data to make clinically relevant predictions.
Holger Hennig
Dept. of Systems Biology and Bioinformatics, University of Rostock, Rostock, Germany & Imaging Platform, Broad Institute, Cambridge, MA, USA
Imaging flow cytometry combines the high‐throughput capabilities of conventional flow cytometry with single‐cell imaging. Here we demonstrate label‐free prediction of DNA content and quantification of the mitotic cell cycle phases by applying supervised machine learning to morphological features extracted from brightfield and the typically ignored darkfield images of cells from an imaging flow cytometer [1]. This method facilitates nondestructive monitoring of cells avoiding potentially confounding effects of fluorescent stains while maximizing available fluorescence channels. The method is effective in cell cycle analysis for mammalian cells, both fixed and live, and accurately assesses the impact of a cell cycle mitotic phase blocking agent. As the same method is effective in predicting the DNA content of fission yeast, it is likely to have a broad application to other cell types. We are exploring other applications in immune system cells, and are developing a user‐friendly open‐source software workflow.
[1] T. Blasi, H. Hennig, H.D. Summers, F.J. Theis, D. Davies, A. Filby, A.E. Carpenter, P. Rees. Labelfree cell cycle analysis for high‐throughput imaging flow cytometry. Nature Communications 7, 10256 (2016)
Dept. of Cell Biology, University Medical Center Utrecht, Dept. of Information and Computing Sciences, Utrecht University
The mining of high content, (HC), data sets, continues to be a bottleneck in phenotypic screening projects. We have previously reported a web‐based data analytics tool called HC StratoMineR. With minimal training, biologists can investigate their HC data and generate phenotypic hit lists using unsupervised methods. For large HC datasets however, especially those comprised of data at cellular resolution, the use of unsupervised methods can be computationally intensive. For this reason we have developed a web‐based tool called HC ClassifieR. For hit detection we use a Support Vector Machine (SVM) approach using a one‐class classifier model. The model is built based on the phenotype generated by a particular reagent such as a negative control. This can be used to rapidly identify similar or dissimilar wells in a large dataset allowing for the detection of all outliers. Multiple combined classifiers can be used to characterize multiple phenotypes. We have used this method to analyze a genome‐wide siRNA screen and we can show that the one‐class classification model shows an accuracy of 94.46%. The supervised method also gives exactly the same number of desired hits, (69) with a smaller p‐value when compared to the unsupervised results. Supervised methods can also be useful for the detection of false positive hits. For this we use a Random Forest, (RF), classifier that can be used to rapidly identify extreme outliers. In HCS these are frequently caused by fluorescent compounds. A binary classification method is used which requires high and low intensity controls. A model is generated based on these controls that can identify wells containing large numbers of highly fluorescent cells. These can be flagged for elimination from hit lists or for the inspection of images. In a compound screen the method correctly identified 84% of fluorescent compounds.
Feng Shan, Daid A. Close, Daniel P. Camarco,
Department of Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, University of Pittsburgh Cancer Institute
Head and neck cancer (HNC) results in ∼600,000 new cases and ∼300,000 deaths per annum. Surgical resection, radiotherapy and chemotherapy have not significantly improved HNC prognosis with cures and 5‐year survival rates of 50%, and median survival of 6–12 months in patients with recurrent or metastatic HNC. There is therefore a need for new effective therapies. Although 2D tumor cell growth inhibition HTS assays predominate in cancer drug discovery, cells in solid tumors exist in a highly interactive 3D microenvironment where cell‐cell interactions, cell‐ECM interactions and local gradients of nutrients, growth factors, secreted factors and oxygen regulate cell function and behavior. Cumulative evidence indicates the 3D microenvironment profoundly alters many biological responses including those to cancer drugs. To determine if drug penetration might be a contributing factor, we utilized ultra‐low attachment microtiter plates to generate HNC spheroids and used HCS imaging and analysis methods to measure drug accumulation in 2D and 3D cell cultures. Although the amount of drug absorbed was higher in spheroids, in part due to higher cell numbers, ellipticine, idarubicin, daunorubicin and doxorubicin accumulation exhibited similar concentration and time dependent behaviors in 2D and 3D. While drugs were uniformly distributed in 2D monolayer cells, cells in spheroids exhibited a distinct concentration gradient transitioning from high in the surface layers to low in the inner core. We speculate that the cell‐cell contacts and multilayer structure of spheroids form a permeability barrier that restricts drug penetration thereby reducing the concentrations experienced by cells in the interior. The reduced drug concentration and diminished proliferative capacity of cells in the interior of spheroids might both contribute to the apparent resistance to some drugs. We believe that the use of 3D HNC models for cancer drug discovery has the potential to significantly improve the success rate of HNC drug development.
Patrick McDonough, Ranor Basa, Wiem Lassoued,
Scintillon Institute, Vala Sciences, Inc.
Parkinson's Disease (PD) occurs due to loss of dopaminergic neurons in the Substantia Nigra, an area of the brain involved in motor control. Certain environmental toxins (e.g., paraquat, rotenone) increase the risk of PD. Additionally, evidence links the protein alpha‐synuclein (α‐Syn) to PD, including the presence of α‐Syn in Lewy bodies, the neurotoxicity of “prion‐like” α‐Syn aggregates, and the predisposition to early‐onset PD in carriers of the α‐Syn‐A53T mutation. We present progress on a Kinetic Image Cytometry (dynamic HCS) assay to test chemicals for PD‐relevant neurotoxicity. Accordingly, we utilized an isogenic pair of human induced pluripotent stem cell (hiPSC) lines, featuring wild‐type cells and cells with the A53T mutation of α‐Syn, respectively, which were differentiated to dopaminergic neurons (from Cellular Dynamics International). The activity of the neurons was investigated in 96‐well plates using fluorescent indicators for intracellular calcium (Fluo‐4) and membrane voltage (FluoVolt) by recording digital movies on living neurons in each well, and performing single cell (cell‐by‐cell) analyses of characteristics of the Ca++ transients and action potentials. The synchronization of spontaneous calcium transients and action potentials between multiple neurons varies with conditions and cell type. For example, rotenone diminishes the synchronized activity in a time (e.g., effect at 2 days> over night >2 hr) and dose‐dependent (effect at 1 μM > 0.1 μM) fashion, particularly in the A53T‐α‐Syn neurons, suggesting this mutation increases the sensitivity of the neurons to rotenone toxicity. The assay enables high‐throughput testing of compounds for toxicity and efficacy relevant to PD.
University of Toronto
Short half‐life proteins regulate a wide variety of essential cell processes, from cell cycle control to transcription. For many of these critical short half‐life proteins, only a handful of well‐characterized regulatory pathways have been identified. We have developed a new tool to identify novel regulators of short half‐life proteins in a high‐throughput manner. We have piloted this technique with the highly regulated c‐Myc protein (t½ = 30 min), fused to Venus, and expressed in a non‐transformed cell line. Treatment with an inhibitor of Myc turnover doubles the percentage of cells with Venus signal above an established threshold (% Venus‐positive cells), providing a read‐out of increased stability. Surprisingly, the nuclear intensity of the cells did not markedly change with treatment. We applied this probe to screen 320 kinase inhibitors using the Opera confocal high‐content imaging platform. By prioritizing compounds based on z‐score cutoffs calculated from the % Venus‐positive cells in each treatment, we subsequently identified 21 and 9 compounds that increased or decreased the % Venus‐positive cells, respectively. False positives were eliminated through secondary screening at a range of doses, followed by western blotting for Myc levels in cells treated with the identified inhibitors. Compounds that increased Myc levels by western blotting were then assessed for their effect on Myc half‐life. To identify false negatives, we extracted morphological and intensity features for all cells. We performed classification and unsupervised clustering of compounds that fell above a lowered z‐score threshold. Two compounds were identified that grouped with top priority compounds, but fell below the original z‐score cutoffs. The fusion of short half‐life proteins to Venus has enabled us to conduct high‐throughput screening for novel regulators of short half‐life proteins. We look forward to applying this new screening technique with more comprehensive libraries to better characterize the regulation of understudied short half‐life proteins.
POSTER ABSTRACTS
Laboratory of Systems Pharmacology, Harvard Medical School, HMS LINCS Center, Harvard Medical School; Broad Institute of MIT and Harvard, Cambridge, Massachusetts, USA; RareCyte Inc. Seattle, WA
Tumor microenvironment plays an important role in disease progression and therapy resistance. Increasing understanding of the heterogeneity in both tumor and its microenvironment will be crucial to development of more effective therapies. Recently, several studies employing single‐cell sequencing methods reveal enormous complexity in tumor microenvironment. However, the spatial information and cell‐to‐cell interaction could not be preserved in these dissociated samples. Immunofluorescence has been widely used in different fields of biological and medical research for decades. The ability to obtain in situ and single‐cell information makes this technique particularly important in tumor biology. However, biochemical and optical constraints limit the number of signals that could be captured simultaneously. We have developed the CycIF (Cyclic Immunofluorescence), an easy and robust method to increase the multiplexity of conventional immunofluorescence. We implemented the CycIF on RareCyte Accucyte‐CyteFinder platform, allowing us to simultaneously monitor multiple antigens in the same samples. Furthermore, this workflow can apply to both liquid samples (blood & dissociated tissues) and tissue slides (FFPE or frozen sections). We used this platform to probe tumor heterogeneity, microenvironment and immune infiltration in various types of tumors. Up to 30 different markers could be simultaneously detected, and these markers represent a wide range of biological processes, including the key molecules for lymphocyte surface makers (CD45, CD4, CD8 etc.), stromal/EMT proteins (E‐Cadherin, Vimentin), cell cycle regulators (CycD1, PCNA, Ki67), signaling proteins (EGFR, pERK, pS6) and apoptosis mediators (p53, Bax, Bcl‐2). Our study not only provides the first detailed map of tumor and its immune microenvironment, but also illustrates a robust multiplexed imaging platform for probing tumor heterogeneity. Furthermore, the capacity of CycIF/RareCyte platform to probe various types of samples allows us to use it not only in the basic research but also in the pre‐clinical or clinical trial stages of drug discovery.
INO
Image based High Content Screening (HCS) is a growing field of drug research. The combination of Fluorescence Lifetime Imaging Microscopy (FLIM) and hyperspectral imaging modes in drug screening can provide information on the molecular specificity as well as the mechanism of action of candidate molecules. Unfortunately, these techniques are typically time consuming, and are limited by the number of photons coming from the sample and the required spectral resolution. Conventional FLIM systems can record images of mCerulean3‐Venus pairings in live cells at a maximum rate of 1 image per 80 seconds for a 400x400 pixel image. Conventional hyperspectral systems typically have 32 channels which limit the spectral resolution or bandwidth of the system. With those limitations in mind, we have developed a new HCS FLIM and Hyperspectral system that images at a speed of greater than 0.1 fps. The FLIM detection module employs 8 detectors to overcome the pile‐up limitation of conventional Time Correlated Single Photons counting (TCSPC) systems, while the hyperspectral detection module has been designed to provide 64 spectral channels resulting in a spectral resolution of 8nm over 450nm to 850nm. In this Poster we present the design of our FLIM and hyperspectral submodules, and demonstrate their performances. We also present the overall capabilities of our Confocal/FLIM/Hyperspectral microscope for HCS.
Nexcelom Bioscience, Immunocore
Cell‐mediated cytotoxicity assays have been frequently performed to characterize cancer cytotoxic potential of immune cells, antibodies, and drug compounds. Traditionally, these assays are performed using release assays such as Cr51 (radioactivity), Calcein (fluorescence), or LDH (enzymatic). However, release assays have limitations such as the handling of hazardous material, the indirect measurement of cell death leading to an under estimation of cytotoxicity, the requirement for a large volume of cell sample, and the inability to visually confirm, image and track the assay kinetically. In the recent years, Celigo image cytometry has been used to perform high‐throughput cell‐mediated cytotoxicity assays using a direct cell counting method where cancer cells (Target) are stained with Calcein AM. Upon the addition of effector immune cells with the fluorescently stained Target cells in the presence or absence of antibody or drug compounds, the Celigo image cytometer is used to analyze the change in Target cell count over time to determine the cytotoxicity percentage. In this work, we demonstrate a novel method of analyzing T cell‐mediated cytotoxicity on 3D tumor spheroids in the presence of absence of ImmTAC molecules, which can promote higher T cell killing. In this experiment, MDA‐MB‐453 GFP expressing breast cancer cells are used to form tumor spheroids in an ultra‐low attachment plate. The spheroids are then treated with primary T cells at 1:10 and 1:50 E:T ratios, as well as 10, 1, 0.1, 0.01, and 0.001 nM ImmTAC. The results showed a dose response effect of T cell killing with the addition of ImmTAC molecules by measuring spheroid size in GFP fluorescence and viability using propidium iodide. The ability to screen cytotoxic effects of immune cells, antibody, and drug compounds on 3D tumor spheroids can provide an alternative tumor model for identifying more qualified cancer drug candidates for drug discovery campaigns.
Miami Project to Cure Paralysis, Univ. of Miami
One of the challenges for any high content screening campaign is the consistency and reproducibility within the assays that are used. Scientists typically focus on assay development and performance metrics but overlook system variations that can also contribute to inconsistency. Instrument light sources, filters, cameras, and stage performance can change over time. It is accepted that Standard Operating Procedures (SOPs) should be established for cell culture and immunofluorescent staining when performing high content screening experiments. SOPs are also needed to detect variations of instrument performance. We contacted 19 HCA facilities around the world and received responses from 10. Interestingly, there is only one facility with a SOP (for form factor calibration). The other facilities do not have documented SOPs. However, many perform testing procedures from time to time to ensure consistency between experiments. These procedures include checking the light source intensity, checking the environmental control module, testing Laser autofocus optimization for new plate types, and running a fluorescent bead plate before image acquisition. Most of the facilities rely on the instrument vendor for annual testing. Some facilities refrained from doing system testing to prevent being disqualified from their Preventive Maintenance agreement.
The use of routine instrument testing in other core facilities environments, such as FACS core facilities and clinical laboratories, is not just common but often required by various regulatory agencies. If HCA approaches are going to enter the field of personalized medicine, then systems for routine HCA machine testing will be required. While it would be ideal if the instrument vendors provided SOPs and testing frameworks, the HCA community will need to take the lead in developing these practices.
This project was supported by the Miami Project to Cure Paralysis and the Walter G Ross Foundation.
Istanbul Kemerburgaz University, School of Medicine, State University of New York at Binghamton, Department of Chemistry, Binghamton, NY USA
Site‐specific bioorthogonal imaging of proteins in vitro/vivo is one of the most powerful and conventional strategies in chemical biology. It is also essential to point out that the use of chemoenzymatic tools for site‐specific targeting of proteins in their native conditions is critical inside living cells. Organic reactive pairs frequently used for chemical conjugation are aldehydes/ketones with hydrazines/hydrazides/hydroxylamines. Although the reaction is generally specific for the two components, even in a cellular environment, the reaction is very slow under physiological conditions. Addition of a phosphate group at the ortho position of an aromatic aldehyde drastically increases the reaction rate by an order of magnitude and enhances the aqueous solubility of the reagent and the product. We have synthesized phosphate‐substituted aldehyde synthetic models to study kinetics of their reactions with hydrazines and hydrazides that contain a fluorophore. This rapid bioorthogonal reaction should therefore not only be potentially a very useful reaction for site‐specific chemical ligations to study and image complex cellular processes, but also improve the reaction biocompatibility, specificity, and reaction efficiency rates inside living systems.
Gigmade ltd., Isomics Inc.
Biologists pursuing commercial research need to quantify cell microscopy images. Tools such as Cell Profiler are usually used for such purposes. Unfortunately, biologists tend to lack the luxury of time or informatics backgrounds to learn using complex software. Additionally, users in commercial settings often have to overcome other hurdles: installation, reliable running and scaling performance of such software for large datasets can be difficult issues to manage. The disconnected nature of outsourcing projects prevents the biologist from later changing minor aspects such as detection thresholds, or re‐running the analysis on a new plate without the involvement of the external vendor. CellExpress is a web‐based platform that allows biologists with no informatics knowledge to quantify cell microscopy images, by providing an easy way of re‐using old projects, outsourcing new analysis needs, and adjusting biologically adequate parameters in a user friendly manner. Cell‐Express installs and runs Cell Profiler at multiple frozen software versions, and automatically runs many copies in parallel in a cloud environment for larger datasets. Users access a simple web page that can retrieve their images from connected data storage. Users can then make image samples available to internal or external experts via the web page. When the expert has finalized an analysis configuration, they can register it with CellExpress. Users can then independently run the analysis on the original dataset or any other dataset, at their own pace and without requiring the involvement of the expert unless they are looking for substantial changes. Additionally, users can adjust parameters specifically enabled by the experts as adequate for the project.
Visual Poster:
Faculty of Pharmacy, Hasanuddin University, Makassar 90245, Indonesia
Marine resources produce secondary metabolites with a huge array of pharmacological activities. We describe development of a fast screening method, based on the principle of nano‐magnetic dynabeads, for a quick screening of Glioma (GLI) inhibitor compounds from marine resources. GLIGST were cultured, immobilized on carboxylic acid magnetic dynabeads, and mixed with extract of sponges. This method is more effective than bioassay‐guided screening of natural products because it diminishes numerous stages of fractination and chromatography. Of 20 sponge extracts, Xestospongia testudinaria, Clathria sp. and Callyspongia sp. were selected on the basis of GLI inhibitory activity screening. Theonelapeptolide and derivative compounds were isolated. The structure elucidation was on the basis of spectral data of 1D NMR compared to literatures.
Department of Safety and Exploratory Pharmacology, SALAR, Merck Research Laboratories, West Point, PA
The early and accurate prediction of cardiotoxicity in preclinical drug development is required to decrease the costs during the late stage and minimize the risk of post market drug withdrawal. Human induced pluripotent stem cells‐derived cardiomyocytes (hiPSC‐CM) have been widely utilized as a relevant in vitro model for cardiotoxicity. hiPSC‐CM cultured in 2D environment express an immature phenotype compared to native cardiac cells, and choosing a suitable 3D matrix may enhance the cell maturity both structurally and functionally. To test that, we cultured hiPSC‐CM on novel 3D aligned nano‐fiber plates (Mimetix, The Electrospinning Company, UK), as well as on regular 2D culture plates for comparison, and stained them with antibodies for f‐actin. Results showed that, unlike cells grown on 2D plates, iCell cardiomyocytes grown on Mimetix plates aligned in the same direction as the fiber orientation, resembling a native cardiac tissue, and the inner cell structure shared closer similarity to the native CM. In addition, an effect of two compounds (sunitinib and haloperidol, 0.03–30 uM) at 2, 4, 24 and 48 hr on nuclei, mitochondria, and cell survival was monitored in hiPSC‐CM cultured on 3D aligned fiber plates (both straight and wavy configuration) and stained with Hoechst/TMRM/ Red Dot dyes. A panel of cellular features indicating nuclear and mitochondrial intensity, morphology and texture, was analyzed using Columbus software. Both compounds displayed concentration and time dependent effects on cellular features, with sunitinib being much more potent than haloperidol at similar concentrations. The scale of changes between the compounds and the negative control (DMSO) was more pronounced for the wavy nano‐fibers for the majority of features, as compared to the straight fibers, indicating that wavy fibers provide more favorable conditions for detecting compound‐induced changes in hiPSCCM. More compounds with different mechanisms of action need to be tested to confirm this finding.
Lee B. Barrett
Boston Children's Hospital
The F.M. Kirby Neurobiology Center at Children's Hospital Boston houses the Assay Development and Screening Facility (ADSF) which is located on the 12th floor of the CLS building, 3 Blackfan Circle. The core enables researchers to develop assays of their targets suitable for subsequent industrial‐scale, high throughput and high content screening, thereby filling this critical gap in the drug development pathway. Additional funding by Mass Life Sciences (MLSC) enables external and industrial researchers access to the core facilities. The facility is built around two instruments; Hamamatsu functional drug screening plate reader (FDSS 7000) and a ThermoFisher Arrayscan XTI High content screening platform, with live cell chamber. The core provides access to annotated NIH and FDA approved screening libraries, as well as chemically diverse sets of small drug‐like compound libraries. The ADSF has, in addition to HTS and HCS instrumentation, state of the art liquid handling resources from Agilent and Wellmate, as well as dedicated BL2 rated tissue culture facilities. Stop by to talk to the Core Manager, Lee Barrett, to arrange a tour of our facility.
PerkinElmer, Inc.; F. Hoffmann‐La Roche Ltd.
High content screening (HCS) is a well‐established technique for biological research and drug discovery to identify substances that alter the phenotype of a cell in a desired manner. With the recent substantial technological innovations in optics and software, modern imaging instrumentation allows to generate unprecedented amounts of high‐resolution image data in short time frames, thus enabling the investigation of multiple cell lines, cell kinetics or 3D images in a single screening campaign. At the same time, complex image analysis algorithms such as texture analysis, morphology analysis or 3D segmentation have been introduced that precisely characterize phenotypes of individual cells or even cells in tissue samples. Consequently, the performance of IT infrastructure used by scientists in phenotypical screening has become one of the limiting factors. Here we report on a collaborative proof‐of‐concept study between Hoffmann‐La Roche in Basel and Perkin Elmer, Inc. where cluster‐based high‐performance computing has been employed to significantly reduce the computing time for analyzing large amounts of high content screening data. Images captured by a Perkin Elmer Opera QEHS system (x 384‐well plates, y channels, z number of fields per well) were extensively analyzed by an Acapella script that calculates approx. 230 descriptors per cell (nuclei, cytoplasm, membrane and spot segmentation; morphological, intensity and classification parameters). A high‐performance cluster based on DELL hardware, interconnected by InfiniBand for storage access and by 10G Ethernet was set up to be controlled by a SLURM job scheduler. Each scheduled job analyzed a single well of a single microtiter plate. We achieved an average analysis time of approx. 2 minutes per 384‐well Microtiter plate; the whole set of images of an entire high content screening campaign can be analyzed within just 40 minutes instead of several of weeks.
Allen Goodman and
Broad Institute of MIT and Harvard, Cambridge, Massachusetts, USA
Increasingly, researchers create complex biological model systems involving three‐dimensional structures, including organoids, tumor spheroids, and even whole organisms. High‐content screening systems enable acquisition of 3D images in high‐throughput yet the software to analyze crucial metrics from these images has lagged behind. We have recently added 3D segmentation, registration, and measurement capabilities to CellProfiler, open‐source software for high‐throughput image analysis. As a result, essentially any operation possible for 2D images now also works for 3D images. We have begun testing these new features in several experimental situations including gene‐edited colonies of human induced pluripotent stem cells. This work was funded in part by the Allen Institute for Cell Science and the National Institutes of Health.
Michael Halter
National Institute for Standards and Technology
Induced pluripotent stem cell populations have enormous scientific and therapeutic potential. These cells are frequently cultured as colonies that may contain thousands of cells which span multiple field of views on an optical microscope. In order to monitor and accurately characterize the dynamic and heterogeneous character of stem cell populations, many individual colonies must be imaged over time. We present imaging and analysis methods that make it possible to quantify the dynamic and spatial behavior of a reporter of stem cell pluripotency in 100s of hESC colonies over extended periods of time. To demonstrate these methods, we observe and quantify Oct4 expression under culture conditions that are expected to retain pluripotency. The image analysis, visualization software, and analytical pipeline developed for this study allow spatial and dynamic characterization of the growth and expression of Oct4 in a large number of hESC colonies. These methods for acquisition, image analysis, and visualization allow us to quantify differences in 3 preparations of pluripotent colonies, and to identify rare behaviors in colonies.
Griffith University, Nathan, Queensland, Australia
The tumor microenvironment is unique for each type of cancer. To discover new therapeutic options or profile small molecules identified for the potential treatment of cancer, advanced in vitro cell culture models have been implemented. These in vitro culture models include three‐dimensional (3D) cell cultures, which are considered more physiologically relevant when compared to monolayer cell cultures. Investigations undertaken utilizing 3D cell cultures have demonstrated outcomes with greater similarities to the in vivo tumor microenvironment than those obtained with cells grown in a monolayer. Drug discovery requires robust assays that are miniaturized for compatibility with liquid handling robotics, high‐content imaging equipment and multi‐label plate readers. With this in mind, 3D cell culture assays have been developed in both 384‐ and 1536‐well microplates for a range of cancers, including breast and pancreatic cancer. These advanced assays were developed utilizing a MatrigelTM‐based biological scaffold and have been assessed for parameters such as reproducibility, cellular proliferation and assay end point measurements (total well viability and live cell imaging). The 3D cell culture models have been utilized in research focusing on small molecule profiling, including standard‐of‐care chemotherapeutics. Results show that the efficacy of anti‐cancer agents is modulated in advanced 3D model systems from those observed with monolayer assay formats. Furthermore, cancer cell resistance to selected therapies, such as doxorubicin, have been linked to cell‐to‐extracellular matrix signaling. Advanced cell culture techniques have the potential to deliver information‐rich data of considerable value for drug discovery efforts. It is anticipated that these more biologically relevant 3D cell culture models will significantly impact on our understanding of the microenvironment and compound efficacy, facilitating widespread incorporation of such models into cancer drug discovery programs.
Michael Halter
National Institute for Standards and Technology
High content imaging facilities are a core tool for discovery biology and translational research. Despite the abundance of critical image data acquired by high content imaging instruments, methods for assuring the consistency and comparability of the image data are not routinely implemented. At the NIH National Center for Advancing Translational Sciences (NCATS), high content analysis is used as a core technology to screen for new molecular probes and to uncover biological mechanisms. We implemented an easy to implement procedure (Halter et al, Cytometry A, 2014) at NCATS so that instrument performance could be measured routinely and compared over time. The procedure benchmarked the limit of detection, intensity response function, and field uniformity of the instrument to a stably fluorescent Schott 475 GG glass reference material. The method was implemented weekly over the course of several months on highly used instruments. We found that the analytical performance of the instruments we benchmarked to be highly stable over this time. This initial study demonstrates the implementation of a routine benchmarking procedure that can be used to provide evidence that image data are comparable and identify when an instrument's performance has changed.
Daniel A. Gutierrez, Helena Bonin, Matthias Fassler, André Stephan,
Genedata
Adding to current image‐based screening activities such as Digital Pathology (DP) and High Content Screening (HCS), new biomedical imaging technologies such as Mass Cytometry Imaging (MCI) and various Single Plane Illumination Microscopy (SPIM) techniques are becoming commonplace in the field of preclinical drug discovery and development. These emerging technologies excel during the late Research & Development phases, allowing scientists to quantify drug efficacy, understand cellular mechanisms and drug mode of action, and study system‐level toxicology and ADME on model organisms. However, the multiple images resulting from preclinical experiments come at a very high cost to pharmaceuticals. Their sheer number, diverse meta‐data structure, and in the case of SPIM, terabyte‐sized single files, make it extremely difficult to establish consistent image analysis and management workflows that are simple for scientists to use, while being based on commonly available IT infrastructure. As a consequence, imaging workflows lack the integration, infrastructure, and automation needed to grant fast, systematic, and timely access to data. This poster presents a conceptual image management workflow that attempts to overcome such limitations. As a use case, we study one of the most challenging image‐generating instruments in the field: the Lattice Light Sheet Microscope. We discuss aspects of central image storage and software design strategies, new concepts such as Deep Learning and parallelized analysis in local or remote cloud systems, and how to incorporate best practices from science and modern IT into an integrated solution. Our goal is to arrive at an end‐to‐end implementation that significantly optimizes workflows in preclinical imaging, so that pharmaceutical companies make better decisions taking full advantage of the imaging technologies they have invested in.
BRITE, North Carolina Central University
Hit selection in imaging‐based drug discovery has often relied on univariate response data based on a single parametric endpoint that may overlook significant synergistic biological descriptors that could be useful for improving signal‐to‐noise ratio, identifying and characterizing modes of action or alternative target pathways. High content assays show significant potential for discovering hits by utilizing quantitative image microscopy data patterns to extract indirect effects by analyzing phenotypic responses to compound treatments. We developed phenotypic profiles based on a spectrum of analytical techniques to detect biologically relevant synergistic feature correlation on compound induced differentiation of pancreatic Mesenchymal Stem Cells toward an insulin producing beta‐cell type. In the absence of a positive control, we were able to deduce interesting phenotypes by using exploratory analysis of cellular morphology attributes. We exploited multivariate outliers for hit profiles, and generated idiosyncratic profiles by summarizing feature distributions and differences between treatment and control distributions. Separately, data driven feature reduction was employed by using principal component analysis to infer latent features to circumvent redundancy prior to clustering. This enabled the classification of phenotypes by compounds in a highly annotated library while reducing the dimensionality of the data. Herein we describe a method to implement a multivariate scoring system between subtle phenotypic profile variations without a positive control.
Michael J. Van Kanegan
ZenBio
The activation and recruitment of brown adipose tissue has become an exciting target in the fight against obesity and its related metabolic diseases. Brown adipose tissue differs from white adipose tissue in its critical ability to burn energy as heat through UCP‐1 driven adaptive thermogenesis. Expression of mitochondrial UCP‐1 uncouples oxidative phosphorylation, resulting in accelerated substrate oxidation but low ATP production. Activation of brown adipose tissue and stimulation of the browning of white adipose tissue in rodent models results in increased energy expenditure, plasma triglyceride clearance and decreased adiposity. To date, screening efforts have been limited to rodent models leaving a void in the therapeutic development pipeline. ZenBio has addressed this need by providing the first characterized human brown adipocyte cell culture system to the research community. Additionally, we are developing high content/high throughput screening platforms capable of multiplexing molecular and functional assays to screen potential candidates for stimulation of human brown adipogenesis. We have established protocols for individual assays that measure UCP‐1 protein expression, mitochondrial mass, lipid content, and glucose uptake and are currently validating multiplexed analysis according to standard high content screening practices. A second image‐based, phenotypic screening platform to measure gene expression is being developed using ViewRNA® in situ hybridization (ISH) technology to perform single cell analysis of multiple genes of interest. Development of these fully customizable screening platforms provide a high demand service to identify agents that can increase differentiation and enhance activity of human brown adipocytes as potential treatments for obesity, T2D and metabolic disease.
Pfizer, Cambridge, Massachusetts
Culture of neurons for in vitro studies such as compound screening is conventionally performed with cells isolated from prenatal rodent sources. While these neurons are a staple of neuroscience research, there is great interest in studying humanized systems as models of neurodegenerative disease. While a highly scalable alternative, immortalized human neuroblastoma cell lines lack many morphological, electrophysiological, and phenotypic characteristics of primary neurons. Embryonic (ES) and induced pluripotent stem cell (iPSC) research has progressed in recent years to offer a new path in generating neuronal cells from human sources. Some drawbacks of common human ES (hES) and iPSC‐based neuronal differentiation methods are: 1) the long differentiation cycles (weeks to months) involving labor intensive culture maintenance and, 2) yields of heterogeneous cell populations.
To address the need for neuronal cell platforms amenable to microplate‐based applications such as high content analysis, engineering hES cells to express neuronal fate‐determining transcription factors offers a promising avenue. Using genome editing, we have developed an inducible hES cell system that can rapidly produce functional neurons in under 2 weeks. Upon induction, these cells begin exhibiting neuronal‐like phenotypes in just 2–3 hours and are morphologically similar to rodent primary neurons after 5 days. The induced cultures exhibit a high degree of homogeneity and can be frozen as differentiated neurons which allows for rapid re‐plating into screening microplates. With this new cell system, we have been able to develop high content assays compatible with small molecule and functional genomic screening workflows. Characterization of the neurons produced by inducible transcription factor expression is ongoing but may offer a promising tool for modeling human phenotypes in HTS applications.