Use of Human-induced Pluripotent Stem Cell-derived Cardiomyocytes As a Screen for Drug-induced Cardiotoxicity
Kyle Kolaja, Cellular Dynamics International
Abstract: Advances in pluripotent stem cell derived cardiomyocytes have enabled the use of human cell based systems in early discovery stages of research. Stem cell derived cardiomyocytes display similar cardiac action potentials and drug induced responses such that they can be used to study drug induced cardiac injury. A better understanding of cardiovascular drug induced injury is critical for developing effective counter screens, biomarkers, and gleaning true mechanistic understanding. It is particularly important to determine if the triggering event is due to electrical disruption such as QT prolongation or overt cardiac damage. This presentation will explore recent advances in human induced pluripotent cell derived cardiomyocytes and how they are being applied to mitigate cardiovascular drug risks.
Cellular Dynamics and ACEA Biosciences
Cardiotoxicity manifests through several distinct mechanisms. This workshop will present (1) novel and peer-reviewed data using human cardiomyocytes with impedance and electrophysiological testing that validates a simplified and highly predictive workflow for electrical, biochemical, and contractile based cardiotoxicity detection and (2) integration with the FDA’s Comprehensive In Vitro Proarrhythmia Assay (CIPA).
- Biomedical Engineering, University of Wisconsin-Madison
- Department of Orthopedics & Rehabilitation, University of Wisconsin-Madison
- Materials Science Program, University of Wisconsin-Madison
Abstract: The large quantity of known environmental and synthetic compounds calls for a robust screening assay to identify potentially toxic compounds. Many of these compounds have been characterized as putative vascular disruption compounds (pVDCs) using assays of human vasculature in a dish. However, current assay systems typically measure two-dimensional (2D) endothelial tubule formation as a measure of vascular disruption. These assays are often performed on ECM-mimicking materials such as Matrigel, which is composed of hundreds of unique proteins and exhibits lot-to-lot variability that may reduce reproducibility. Recently, we have observed endothelial tubulogenesis and sprouting in assay platforms comprising well-defined, synthetic hydrogels. Here we report the analysis of de novotubule network formation and EC sprouting using a high throughput, multivariate approach. Human induced pluripotent stem cell-derived endothelial cells (iPSC-ECs) were encapsulated in a synthetic hydrogel composed of poly(ethylene glycol) (PEG) and tethered cell attachment and cell-degradable crosslinking peptides. iPSC-ECs exhibited invasion that was correlated with the concentration of cell adhesion peptides and degradable peptide crosslinks. Immunocytochemistry of encapsulated cells demonstrated that iPSC-ECs expressed CD31 and exhibited morphologically distinct sprouting behavior consistent with endothelial cell sprouting. iPSC-EC invasion was potently inhibited by VEGF signaling inhibitors while survival was unaffected by inhibition, which suggests that VEGF is required for maximal iPSC-EC invasion but not for cell survival in hydrogel arrays. The screening approaches described here are capable of quantitating single-cell invasion, sprout length, survival, and tubule network formation simultaneously and thus can efficiently distinguish the particular influence of pVDCs on multiple EC functions.
J Naciff1, Y Shan1, X Wang1, K De Abrew1, B Wetmore2, R Thomas3, R Settivari4, E Carney4, R Adams1, J Tiesman1, and G Daston1
- Procter and Gamble, Cincinnati, OH
- The Hamner Institutes for Health Sciences, Research Triangle Park, NC
- US Environmental Protection Agency, Research Triangle Park, NC
- Dow Chemical Company, Midland, MI
Abstract: Continuing with our efforts to develop toxicogenomics-based toxicity testing methods using in vitro human cell models, we have characterized the usefulness of three cellular models of human hepatocytes by evaluating the transcriptional response of HepG2, HepaRG and iCell Hepatocytes (Cellular Dynamics International) to 30 chemicals, including some known hepatotoxicants of varying potencies and modes of action (clofibrate, troglitazone, diethylhexyl phthalate, phenobarbital, tetrachlorodibenzo p-dioxin, imidacloprid, etc.) as well as chemicals not known to be hepatotoxicants. Gene expression profiles were determined using a custom approach which measures the expression levels of approximately 1000 landmark genes, and uses a computational model trained to infer the response of all other genes (L1000; Genometry). Each of the 30 chemicals was tested at 3 concentrations, and the transcript profile was determined after 6h of exposure. All three models had significant gene expression changes to most of the chemicals tested. The responses were specific enough to identify matches in the cMAP database (Broad Institute) for the same chemical or comparable agents acting via the same mode of action. For example, the transcriptional signature of ketoconazole (a broad spectrum imidazole antifungal agent), from any of the three hepatocyte models matched the one elicited by other imidazole derivatives in the database, most prominently with miconazole, econazole and sertaconazole. These results indicate that any of the cell types may be useful as an in vitro alternative to determine transcript profile data usable to define chemical mode of action. HepG2 cells appeared to be better at detecting peroxisome proliferators. Disclaimer: This abstract does not necessarily reflect U.S. EPA policy.
S Marukian1, TK Baker2, B Blackman1, R Feaver1, L Mark1, A Mackey1, D Manka1, B Wamhoff1, and A Dash1
- HemoShear, Charlottesville, VA
- Eli Lilly, Indianapolis, IN
Abstract: Induced Pluripotent Stem Cell (iPSC) derived hepatocytes (iHeps) offer the promise of an unlimited supply of cells from single donors and specific disease populations with genetic polymorphisms that may contribute to idiosyncratic drug induced liver injury. However, though these cells express various liver specific features and functions, it is well documented that they are immature compared to human primary hepatocytes, expressing fetal-like proteins, and possess lower levels of activity for certain CYP450 metabolic enzymes, that play a crucial role in drug metabolism and toxicity. In this study, we exposed fresh iHeps to hemodynamic blood flow and transport parameters that recreate the physiological microenvironment of the sinusoidal space. Under these same conditions, it has been previously shown that cryopreserved human adult primary hepatocytes restore morphology, biology, and demonstrate efficacy or toxicity responses to drugs at concentrations that approximate in vivo exposure levels. Microarray analysis of RNA was performed comparing iHeps in a static cell culture collagen gel sandwich and iHeps exposed to liver physiological parameters. We show that the following pathways are enhanced when iHEPs are exposed to physiological parameters relative to static conditions: Key nuclear factors driving hepatocyte differentiation like HNF4α, NF1I2 (PXR) and NR1I3 (CAR), and their downstream pathways – Phase 1 (cytochrome p450), Phase II, and transporters. Peroxisome proliferator-activated receptor alpha (PPAR-α) was activated with enhanced mitochondrial beta oxidation, with consequent suppression of LXR related pathways such as cholesterol biosynthesis and basal inflammatory state. Although promising, current studies are ongoing to further confirm these results while comparing iHep molecular and functional phenotype directly to human primary hepatocytes in a more physiological context.
- NIH, NIEHS/DNTP, Research Triangle Park, NC
- Molecular Devices, Sunnyvale, CA
- Cellular Dynamics International, Madison, WI
Abstract: The developing brain is vulnerable to chemical-induced injury yet many chemicals remain inadequately tested for developmental neurotoxicity (DNT). In an effort to develop and characterize an in vitro model system for DNT screening, we exposed human iPSC-derived neurons to a diverse set of 80 chemicals (e.g., neurotoxicants, drugs, pesticides, flame retardants (FRs), polycyclic aromatic hydrocarbons (PAHs)) across a 6-point concentration range (~0.3 to 100 μM) in 384-well plates. Using HCS imaging, effects on neurite outgrowth parameters (total outgrowth, processes, branching) and cell viability were monitored after 72 h of exposure. Also, mitochondrial membrane potential (MMP) was evaluated at 1 h to assess the potential contribution of MMP to altered neurite outgrowth. The assay-specific noise threshold was calculated based on DMSO control variability and concentration-response profiles were evaluated using a Hill model to derive benchmark concentration (BMC) point-of-departure values. Following assay validation with controls and test replicates, chemicals were ranked by toxicity and selectivity (i.e., effects on neurite outgrowth parameters independent of cytotoxicity). Neurite total outgrowth and branching were the most sensitive endpoints; 15 chemicals (19%) had an effect on neurite outgrowth independent of cytotoxicity. The pesticide rotenone was the most selective, while1 FR (triphenyl phosphate) and 3 PAHs (chrysene, dibenz(a,h)anthracene, acenaphthylene) inhibited neurite outgrowth. Of the 80 compounds, 41 (51%) decreased MMP and 9 were active in both assays, which might indicate that alterations in MMP are linked with neurite outgrowth inhibition for these compounds. These studies have important implications for moving the DNT field forward from a more traditional assessment in animals to implementing novel and improved methodologies and should allow for a more rapid screening and prioritization of potential neurotoxicants.
C Snyder1,2, T Ngo1,2, D Misner1, N Stagg1, and K Staflin1
- Safety Assessment/Investigative Toxicology, Genentech, South San Francisco, CA
- Pharmacology and Toxicology, UC-Davis, Davis, CA
Abstract: A major cause of drug attrition from clinical trials is neurotoxicity, a problem likely due to current preclinical models not adequately predicting human safety. One of the most common drug-induced neurotoxicities observed clinically is peripheral neuropathy. Existing in vivo animal models are time-consuming, expensive, and lack a direct species relationship to humans. Current in vitro models consist of immortalized cell lines and primary rodent cultures, which show variable correlation to human physiology. There is an unmet need for high throughput, in vitro, human-based assays that can identify compounds with neurotoxic and neuropathic potential. These in vitro assays can precede in vivo animal studies, and augment those studies with data in the most relevant species before clinical trials begin. We chose to establish such assays with human induced pluripotent stem cells (iPSCs) that have been differentiated into mature neurons. These iPSC neurons offer a renewable source of mature cells that may be used to investigate human drug safety. Our characterization of these neurons shows that they express mature markers at both the transcript and protein levels. Using kinetic and high-throughput imaging we evaluated a large panel of test compounds known to cause clinical neuropathy for their ability to affect neurite outgrowth and viability in these neurons. Results from our assays show that these test compounds exert neurotoxicity in a class specific manner and that our assessment correlates with clinical findings. We propose the use of these in vitro human assays to evaluate the potential for neurotoxicity and to provide utility in drug safety applications.
A Essex, B Cai, J Sharp, E Batchelder, S Feng, N Prigozhina, J Evans, P McDonough, and J Price, Vala Sciences Inc, San Diego, CA
Abstract: Low level exposure to some toxicants can affect neuronal structure leading to developmental and cognitive impairment. Synaptic transmission is the fundamental unit of communication between neurons. Here, we used human induced-pluripotent stem cell (hiPSC) neurons grown in 384 well dishes for up to 15 days to examine the effects of toxicant exposure on synapses in a high content screen (HCS). These cells demonstrate action potentials as early as 9 days and sensitivity to chemical inhibitors as early as 12 days. To detect synapses, we used antibodies to the pre- and post-synaptic proteins, Synapsin -1 and post-synaptic density protein 95 (PSD95), respectively. To analyze synapse response to toxicant exposure, we analyzed fixed images using algorithms to mask neurite regions demarked by expression of β-III tubulin and measured colocalized signals for the pre- and post-synaptic markers only in these functionally-relevant regions. Our HCS approach has enabled us to screen a library of EPA ToxCAST compounds, FDA-approved drugs (the NCC2 library), as well as other agents, such as glutamate and glycine, and Brefeldin A, known to modulate neuronal function. Additional information that may be related to the mechanism of action for chemicals that showed an effect in the assay included the amount of neurite outgrowth and alterations in nuclear texture. Using this approach we have developed a robust platform for large-scale screening of chemicals that affect synapse formation, the basic unit of neuronal function in humans.
R Calvert1, S Vohra1, M Ferguson2, and P Wiesenfeld1
- Division of Toxicology, US Food and Drug Administration, Laurel, MD
- Division of Mathematics, US Food and Drug Administration, Laurel, MD
Abstract: A variety of dietary supplements contain ingredients that are cardiac stimulants and therefore potential cardiotoxins. In vitro studies can facilitate identification of ingredients of potential concern. iCell cardiomyocytes, a beating human heart cell line, were used to evaluate the effects of phenylethylamine (PEA), higenamine, ephedrine and caffeine on heart cell function. For PEA and higenamine studies, the exposure levels in cell media were based on published blood levels in humans or animals after intravenous administration. Ephedrine and caffeine levels were based on published blood levels following various oral doses in humans. At low to midrange levels, each chemical was examined either with or without added caffeine. Caffeine levels (50 uM) approximated human blood levels reported after consumption of caffeine-enriched dietary supplements. To obtain values for beats per minute (BPM), peak width, etc., rhythmic rise and fall in intracellular calcium levels following 30 min of treatment were measured using a FLIPR Tetra instrument. All experiments were conducted in triplicate. Higenamine 31.3, 31.3 + caffeine and 313 ng/ml significantly increased BPM in an escalating manner. Peak width narrowed with increasing BPM. PEA caused increases in BPM at 0.8 and 8 ug/ ml, along with narrower peak widths. In contrast, 80 ug/ml PEA greatly reduced BPM and widened peak width. This may indicate a toxic effect of PEA at 80 ug/ml. Adding caffeine to PEA 8 ug/ml or higenamine 31.3 ng/ml further increased BPM. Ephedrine produced a significant increase in BPM dose response at 0.5 to 5.0 uM. Caffeine alone resulted in an increased BPM only at a toxic level of 250 uM, but had no effect at 10 and 50 uM. There was no significant enhancing effect of caffeine when combined with ephedrine versus ephedrine alone. Generally, our iCell results correlated with expected effects. We suggest that additional testing may be warranted in vivo to further evaluate these cardiovascular effects.
- Applied/Developmental Research Directorate, Frederick National Laboratory for Cancer Research, Frederick, MD
- Division of Cancer Treatment and Diagnosis, National Cancer Institute, Bethesda, MD
Abstract: Chemotherapeutics targeting the bcl-2 family member, Myeloid cell leukemia sequence 1 (Mcl-1) are being evaluated as potential anticancer therapies due to the role of Mcl-1 in promoting survival of Myc-induced cancer. It has become apparent that both the anti-apoptotic and mitochondrial functions of Mcl-1 play SOT 2015 Annual Meeting 247 important roles in promoting the survival of cancer cells, hematopoietic stem cells, mitochondrial homeostasis and induction of autophagy in the heart. We characterized the role of Mcl-1 in maintaining contractile function and structural integrity of human iPSC-derived cardiomyocytes in vitro and examined whether the pathology observed in Mcl-1 knockout mice can be recapitulated in vitro. Using Mcl-1 siRNA, we knocked down (KD) Mcl-1Long (Mcl-1L) protein expression by over 80%, without a consistent effect on Mcl-1Short (Mcl-1S). Mcl-1L KD was typically associated with a decrease in beat amplitude and cellular impedance by 20%, and a reduction in cellular ATP by 10%. We also observed increased LDH activity in the cellular medium of ~70% compared to non-targeted siRNA controls. There were no measurable changes in mitochondrial membrane potential, but observable changes in mitochondrial morphology presented in electron microscopy images. Interestingly, the KD of Mcl-1L was associated with an increase in activated caspase-3/7. Surprisingly, a pan caspase inhibitor Z-VAD, failed to prevent alteration in beat amplitude and cellular impedance, reduced ATP or LDH release. Finally, Mcl-1L KD potentiated doxorubicin-induced elevations in beat rate and further reduced beat amplitude. Our results imply that more detailed investigations of the role of MCL-1 in regulation of mitochondrial function in human iPSC-derived cardiomyocytes are warranted before these cells can be used to screen compounds for cardiotoxic risk via this mechanism. Funded by NCI Contract No. HHSN261200800001E.
- Texas A&M University, College Station, TX
- Molecular Devices, LLC., Sunnyvale, CA
- University of South Carolina, Beaufort, SC
- Shell International BV, The Hague, Netherlands
- ExxonMobil Petroleum and Chemicals, Machelen, Belgium
- CONCAWE, Brussels, Belgium
Abstract: Hazard assessment of data-limited chemicals by chemical structure-based category read across is widely used for regulatory submissions such as the European initiative, REACH. While such a strategy is sensible for chemically-characterized compounds, it cannot always be used to assess the hazard of complex substances (e.g., petroleum substances UVCBs). Thus, we hypothesized that a biological data-based read across, i.e. safety evaluation centered on categorizing substances according to similarities in their biological response, may represent a feasible alternative. To test this, we selected 26 petroleum substances from six distinct categories: SRGO (Straight Run Gas Oils), OGO (Other Gas Oils), VHGO (Vacuum & Hydrotreated Gas Oils), Bitumens, RAE (Residual Aromatic Extracts), and HFO (Heavy Fuel Oils). For high-content, multi-parametric toxicity assessment in a concentration- and time-response design we exposed induced pluripotent stem cell (iPSC)-derived cardiomyocytes and hepatocytes to a DMSO-based extract dilution series in logarithmical order over five logs. The cardiomyocyte-derived phenotypes included the measurement of effects on contractility, beating pattern and amplitude, as well as cell viability, morphology and mitochondria integrity measured through high-content live cell imaging. For hepatocytes we determined cell viability by live and fixed cell imaging. Cell-specific effects were observed and used as high-dimensional “biological” data inputs for evaluation of the similarities and differences both within and across different substance categories. Collectively, our results indicate that high-content in vitro screening of iPSC cardiomyocytes and hepatocytes is useful for categorization and biologically-based read across of complex substances.
K Dreher1, J Strickland2, W Polk3, and T Shafer1
- National Health and Environmental Effects Research Laboratory, US EPA, Research Triangle Park, NC
- Contractor, US EPA, Research Triangle Park, NC
- University of North Carolina at Chapel Hill, Chapel Hill, NC
Abstract: Exposure risk to silver and metal oxide nanoparticles (NPs) continues to increase due to their widespread use in products and applications. In vivo studies have shown Ag, TiO2 and CeO2 NPs translocate to the heart following various routes of exposure. Thus, it is critical to assess NP systemic toxicity including their effects on the heart and identify properties regulating NP cardiotoxicity. This study examined the cardiotoxicity of 4 Ag (10 or 110 nm,) citrate (cit) or polyvinylpyrrolidone (PVP) coated, 3 CeO2 (<7 – 105 nm), 3 TiO2 (10 – 40 nm) NPs using human induced pluripotent stem cell-derived cardiomyocytes (CM). Metal oxide NPs were sonicated in medium containing 20% fetal bovine serum while Ag NPs were resuspended without sonication. Cytotoxicity was determined using Cell Titer Blue, MitoTracker Deep Red, and nuclear staining assays at 48 h post-exposure to 3 – 50 μg/ml of NP or AgNO3. CM function was monitored using microelectrode array technology measuring field potential duration, beat period, sodium spike amplitude, beat rate prior to exposure and at 1, 24, and 48 h post-exposure to each NP at 3 or 25 μg/ml, or AgNO3 at 3 μg/ml. CM isoproterenol (ISO) (25 and 50 nM) responses were assessed at 48 h post-exposure. Ten nm Ag cit or PVP NPs were cytotoxic to CM at 50 μg/ml and AgNO3 was cytotoxic at 6.3 μg/ml. At 3 mg/ml, only the <7nm CeO2, 10 nm Ag cit or PVP coated NPs decreased all CM functional endpoints and ISO responses at 24 and 48 h post-exposure, while AgNO3 effects were not identical to Ag NPs. Our results demonstrate: i) altered electrophysiology is a sensitive endpoint to assess human NP CM toxicity; ii) size, composition and coating regulate human NP cardiotoxicity; and iii) establishment of an alternative model for human CM NP toxicity testing. (This abstract does not reflect Agency Policy.)
L Pang1 , B Word1 , B Lyn-Cook1 , X Yang2 , N Stockbridge3
- Biochemical Toxicology, NCTR/FDA, Jefferson, Arkansas
- Systems Biology, NCTR/FDA, Jefferson, Arkansas
- Cardiovascular and Renal Products, CDER/FDA, Silver Spring, Maryland
Abstract: Drug-induced proarrhythmia is a major safety issue in drug development. Sensitive in vitro assays that can predict drug-induced cardiotoxicity have been the focus of toxicology research for several decades. Recently, human induced pluripotent stem cell derived-cardiomyocytes (iPSC-CMs) have become a popular model because they largely resemble the electrophysiological behaviors of human ventricular myocytes. However, human iPSC-CMs are derived from individuals with diverse genetic backgrounds, and different laboratories/suppliers may use different differentiation processes and various conditions to culture these cells. Therefore, the responses of different iPSC-CMs to cardiotoxic drugs may vary. In this study, we compared human iPSC-CMs from two major suppliers: Cellular Dynamics International (CDI) and Axiogenesis. We found that the two lines of cells had different sensitivities to the hERG channel blocker dofetilide: 3 nM of dofetilide was sufficient to induce irregular beats in iCells from CDI, but a dose of 30 nM was required to induce arrhythmia in Cor.4U cells manufactured by Axiogenesis. Moreover, the expression levels of cardiac ion channel genes were different between the two lines of cells: iCells had higher expression levels of SCN5A, CACNA1C, and KCNJ2, while Cor.4U cells have more transcripts of KCNE1 and KCNIP2 genes. The expression profiles of cardiac-specific genes and cardiac differentiation genes were also different. The difference in cardiac ion channel gene expression profiles and sensitivity to proarrhythmic drugs between iCells and Cor.4U cells can significantly affect the data interpretation for utilizing these cells for drug safety assessment; a comprehensive characterization of iPSC-CMs models is needed.
J Cohen1,2, K Miyamoto2, Y Tanaka2, M Ishii2, K Nagatome2, H Fuse2
- Drug Safety Res. & Eval., Takeda California, Inc, San Diego, California
- Integrated Technology Laboratories, Takeda Pharmaceutical Company Limited, Fujisawa, Japan
Abstract: This study evaluated the feasibility of using automated high content analysis (HCA) to detect chemical inhibition of neurite outgrowth on human iPS cell (hiPSC) derived neurons, alone and in co-culture with hiPSC-derived astrocytes. To date, limited data has been published on the sensitivity of hiPSC-derived neurons to neurotoxicants and nothing has been published describing co-culture with astrocytes, which might more closely mimic the complexity of human brain. After 2 h post plating neurons in single culture or in co-culture with astrocytes plated 48 h prior, neurons were exposed to acetaminophen (neg. control), Bis-1, rotenone, diazinon, and U0126 (pos. controls) at 0.1 to 100 μM, and stained for beta-III-tubulin 24 h post chemical exposure for morphological (neurite length, cell body area, number of neurites per cell body) and cytotoxicity (neurons per field, area nuclei) endpoints. The number of neurites per cell body was the most sensitive toxicity endpoint in both single and co-culture. Bis-1 exposure to neurons resulted in decreases in neurites at ≥10 μM, area of cell body at ≥30 μM, neurite length at 100 μM, and cell viability at 100 μM; whereas co-culture attenuated both the decreases in neurites at ≥10 μM and in cell body area at ≥30 μM. Rotenone exposure to neurons alone resulted in decreased neurite length at≥1 μM and neurites at ≥1 μM, with no change in cell viability at ≤100 μM; whereas co-culture had increased cytotoxicity at ≥10 μM. Diazinon and U0126 toxicity were similar between single and co-cultures. Acetaminophen had no neurotoxicity in single or co-culture. With the exception of Bis-1, co-culture with astrocytes had minimal impact on the responses and/or sensitivity to the chemicals. These results show that HCA can be used to quantify neurotoxicity parameters in single or co-culture with astrocytes. Future studies may help demonstrate the benefit of a co-culture system in predicting neurotoxicity.
- Cardiovascular Science, University of Glasgow, Glasgow, United Kingdom
- Clyde Biosciences Ltd., Glasgow, United Kingdom
- InSphero AG, Schlieren, Switzerland
Abstract: The electrical characteristics of mammalian adult myocardium depend critically on the geometry of the tissue. Production of large quantities of human cardiac myocytes differentiated from induced pluripotent stem cells (hiPSC-CMs) is now available commercially. These cells, normally cultured on 2D surfaces, are increasingly used to assess the electrophysiological effects of drugs. This study examined the electrical activity of a commercially available hiPSC-CM cell line in both 2D culture and in 3D culture of microtissues.