The first versatile human iPSC-based model of ectopic virus induction allows new insights in RNA-virus disease

A detailed description of pathophysiological effects that viruses exert on their host is still challenging. For the first time, we report a highly controllable viral expression model based on an iPS-cell line from a healthy human donor. The established viral model system enables a dose-dependent and highly localized RNA-virus expression in a fully controllable environment, giving rise for new applications for the scientific community.


Scientific RepoRtS
| (2020) 10:16804 | https://doi.org/10.1038/s41598-020-72966-9 www.nature.com/scientificreports/ the virus genome encoding the viral capsid proteins (Fig. 1a). The CVB3-genome, including the two mutations henceforth called CVB3ΔVP0, prevents competent viral capsid formation, rendering the system non-infectious and thus makes its downgrade from biosafety level 2 to 1 possible 7 . This approach represents a huge advantage for experimenters and allows most labs highly controlled work on RNA-viruses. In addition, the construct has a doxycycline-dependent Tet-on promoter and a fluorescent Venus reporter, which enable the duration and degree of viral infection to be controlled and monitored by adjusting the time of application and the concentration of doxycycline administration (Fig. 1b). For local and selective viral induction in some selected cells to mimic natural infection patterns, we aimed using caged-doxycycline. Cambridge et al. have principally demonstrated that localized gene expression using caged-doxycycline was possible, however in a very simple cell system 8 . Locally controlled de-caging leads to confined regions in which the virus is expressed. This procedure can be used to mimic typical infection patterns, which are usually observed in e.g. enteroviral myocarditis of humans and mice 9 .

Results
The generation of a CVB3DVP0-expressing hiPS-cell line. For the generation of a fully controllable, CVB3-expressing iPS-cell line, a triple-vector system was stably transfected into wildtype human iPS-cells named SFS.1. An overview of the functional triple-vector system for the generation of an inducible, CVB3ΔVP0 www.nature.com/scientificreports/ expressing iPSC-line is shown in Fig. 1. Further details of the cell-line generation are given in the online methods. The main vector in this transfection system, KAO717-pPB-hCMV1-IRES-Venus, is a well-established stem-cell Tet-On transfection vector [10][11][12] . The DNA of CVB3ΔVP0 was cloned into KAO717-pPB-hCMV1-IRES-Venus by extracting CVB3∆VP0 from a synthetic CVB3∆VP0-PUC53 using the restriction enzyme EcoR1 and then reinserting it into KAO717-pPB-hCMV1-IRES-Venus within its EcoR1 cutting site. Expression is controlled by a CMV min promoter and the co-expression of the Venus marker is ensured by a linked IRES2-sequence. The newly generated construct was named KAO717-pPB-hCMV1-CVB3ΔVP0-IRES-Venus (Fig. 1a). The vector KA0637-pPgCAG-rtTAM2-IN encodes the necessary rTA-trans-activator to enable doxycycline-inducible CVB3ΔVP0 expression. Furthermore, KA0637-pPgCAG-rtTAM2-IN encodes a G418 resistance allowing for positive selection after lipofection. In order to generate an effective cell-line it is important that KA0637-pPgCAG-rtTAM2-IN and KAO717-pPB-hCMV1-CVB3ΔVP0-IRES-Venus are inserted together into the genome of an iPS-cell. The co-transfected vector PB200PA-1 encodes the PiggyBac transposase that is essential for the stable random insertion of KA0637-pPgCAG-rtTAM2-IN and KAO717-pPB-hCMV1-CVB3ΔVP0-IRES-Venus into the iPScell genome. This vector requires only transient expression in the early phase after lipofection. In this study, lipofection was optimized for huge vector constructs allowing for transfer of even larger RNA-virus genomes like CVB3ΔVP0 (Fig. 1c, d). Positively transfected clones expressing CVB3ΔVP0 were picked and purified to generate a homogenously CVB3ΔVP0 expressing cell-line (Fig. 1e, f). The clones #8 and #9 were picked for further investigation. Karyotyping of SFS.1-CVB3ΔVP0-IRES-Venus#9 verified the male genotype of the generated cell line and the integrity of the cell´s chromosomes (Fig. 1g). Karyotyping was performed using standard GTG banding procedures. A central factor for an effective viral model system is a tight control over viral gene expression. The amount of produced virus proteins is directly related to the virus load and allows to model different scenarios. In order to test our system for that applicability, different doxycycline concentrations were applied to SFS.1-CVB3ΔVP0-IRES-Venus#9 cells for three days. An immune staining that made the viral capsid protein CVB3-VP1 visible showed that the fluorescence intensity increased as a linear function of the doxycycline concentrations between 0.25 and 2.0 µg/ml. This is a clear indication of a controllable dose-dependent CVB3ΔVP0 expression.
The dose dependence of CVB3ΔVP0 expression on doxycycline offers the possibility to model different virus loads reflecting different phases of infection (Fig. 1h). This is important as in persistent CVB3 infections ongoing low levels of viral RNA are observed whereas in acute CVB3 infections we note high amounts of viral RNA especially in cardiomyocytes 9 . As CVB3 infections often target the human heart with acute and chronic outcomes, SFS.1-CVB3ΔVP0-IRES-Venus was used to differentiate human cardiomyocytes to study the effects of CVB3 expression in more detail.
Expression of CVB3ΔVP0 in differentiated cardiomyocytes. Next, we tested, whether the expression of CVB3ΔVP0 is still possible after differentiation of hiPSC. Cellular differentiation processes potentially can silence or shut off chromosomal regions and inactivate the stable transfected construct unexpectedly. We chose to differentiate hiPSC into cardiomyocytes, since cardiac cells are a main target for RNA viruses like CVB3 and SARS-CoV-2 viruses in human patients leading to severe outcomes as cardiac death. For differentiation, we used a recently established protocol in which the Wnt and BMP signaling are activated on day 0 of differentiation followed by a selective Wnt shutdown at day 2 and day 3 of differentiation leading to a high amount of ventricular-like cells (~ 90% efficiency) 13 .
The generated cardiomyocytes grow as monolayer and are spontaneously active (Supplementary Video 1), facilitating physiological measurements and allowing comparisons between virus-expressing and non-virusexpressing cells. Differentiated cardiomyocytes are capable of expressing the transfected CVB3ΔVP0-IRES-Venus construct after 5 days doxycycline treatment as live cell fluorescence imaging of Venus in differentiated cardiomyocytes clearly demonstrates (Fig. 2a). Immunostaining visualizing typical cardiac markers cardiac troponin I (TNI), cardiac troponin T (TNT) and α-actinin, prove the cardiac phenotype of the differentiated cells (Fig. 2b). Furthermore, analysis of large sarcomeric proteins with loose gel electrophoresis showed the presence of fetal titin N2BA and the sarcomeric signaling protein obscurin, demonstrating the exceptional good quality of the iPS-cell derived cardiac myocytes (Fig. 2c). In order to test for possible effects of the expression of CVB3ΔVP0-IRES-Venus on the human iPS-cell derived cardiomyocytes, the expression of the viral genes was induced with 2 μg/ml doxycycline for 5 days and for 21 days. As a readout for putative effects, the beating frequencies of the induced cells were compared to non-induced control cells. We found significant changes in the beating rates: Non-induced, four weeks maturated, cardiomyocytes showed a regular beating at a rate of approximately 35 ± 1.56 (mean ± SEM) beats per minute at 25 °C. iPS-cell derived SFS.1-CVB3ΔVP0-IRES-Venus#9 cardiomyocytes of the same age, which were treated with doxycycline for 5 days and 21 days showed significantly less contractions per minute. Cells induced for 5 days contracted 26 ± 1.12 (mean ± SEM) times per minute, while cells induced for 21 days contracted 14 ± 2.04 (mean ± SEM) times per minute on average (Fig. 2d). Furthermore, beating of cells induced for 21 days becomes irregular and uncontrolled (Supplementary video 2). These observations thereby cannot be explained with induced apoptosis by doxycycline as the performed life/dead staining of SFS.1CVB3DVP0-IRES-Venus#9 and SFS.1wt indicates (supplementary Fig. 1c). The expression of CVB3ΔVP0 thus exerts adverse effects on cell physiology, as expected for a well-defined disease model. Together with the previously described dose-dependent induction, the produced CVB3ΔVP0-IRES-Venus cell-line is functional and inherits great potential to study pathophysiological effects of CVB3 in a highly controlled, scalable human cell model system. Moreover, the expression of CVB3 can be controlled and fine-tuned, resulting in a population of cells, which are exposed to very similar pathophysiological conditions. This will enhance homogeneity/reduce variability and thus will yield clearer results. www.nature.com/scientificreports/ To further confirm the functionality of CVB3ΔVP0 induction in the generated cell system, a physiological characterization of differentiated cardiomyocytes concerning mitochondrial ROS production and alterations in electrochemical signal generation was performed. Previous studies of CVB3 infections already demonstrated significant impact of CVB3 on ROS production and changes in signal transduction in cardiac cells 14,15 . The expression of CVB3ΔVP0 was induced in differentiated cardiomyocytes for 21 days. The cells were stained with MitoTrackerCMXRos and analyzed for enhanced ROS production, as indicated by elevated fluorescence intensities in fluorescence microscopy. The fluorescence intensities in the cytosol and in mitochondria of 21 days induced cells and non-induced control cells were compared and statistically analyzed.    (Fig. 3d,f), whereas induced cells do not show any reaction to β-adrenergic stimulation (Fig. 3e,f). Equivalent QT-intervals stay continuously elongated in presence of isoprenaline indicating a desensitization of induced CVB3ΔVP0 cardiomyocytes against β-adrenergic agonists. The controlled expression of CVB3ΔVP0 does not only change the mitochondrial structure and electrochemical activity of cells, but also disrupts membrane integrity and leads to alterations in the cardiomyocyte phenotype. To image the three-dimensional structure of iPS-cell derived cardiomyocytes, 3-D holographic imaging (Nanolive) was performed visualizing cell structures.
The 3D scan of living iPS-cell derived cardiomyocytes expressing CVB3ΔVP0 for 21 days revealed drastic changes in overall cell morphology and internal structures while cardiomyocytes without CVB3ΔVP0 expression appear healthy and vital with a high density of intracellular membrane structures as mitochondria and SR and a well-defined cell membrane border. Intracellular membrane structures are partially degraded and the cell membrane borders seem frayed. In addition, CVB3ΔVP0-expressing cardiomyocytes show a high density of intracellular vacuoles marked in blue, compared to the non-induced cardiomyocytes (Fig. 4a, supplementary  Fig. 2a). These findings are supported by membrane staining with CellMask deep red in living cardiomyocytes. Non-induced cardiomyocytes show well-defined and clear cell membrane borders as well as a structured intracellular organization of membranes of organelles. After 21 days of CVB3ΔVP0 induction the intracellular membrane organization seems disturbed and degraded. The staining of many spherical structures within the induced cells is reminiscent of vacuole-like structures, which corresponds to the observations made with the holographic 3D scan. Furthermore, the cell membrane borders are not well-defined anymore and appear frayed as in the holographic images (Fig. 4b, supplementary Fig. 2b).
In summary, here we showed the functionality and applicability of CVB3ΔVP0 to mimic RNA-viral infections under highly controlled and safe conditions. We demonstrated the CVB3ΔVP0 impact on cardiomyocyte function and revealed a disintegration of several cellular organelles as a result of virus infection. However, it has to be mentioned that native viral infection patterns appear focal. Therefore, in a further line of experiments, we tested the SFS.1-CVB3ΔVP0-IRES-Venus cell-model for its feasibility to simulate a patterned virus induction to simulate infection foci.
Photoactive caged-doxycycline enables patterned CVB3ΔVP0 expression. As RNA-virus infections in the human patient do not occur homogenously throughout an infected tissue, but appear in so called plaques, we asked whether a localized induction of CVB3ΔVP0 in SFS.1-CVB3ΔVP0-IRES-Venus#9 would allow to simulate CVB3 plaques in a petri dish. For this purpose, a new derivative of doxycycline, caged-doxycycline (c-Dox), was synthesized and applied to the cell system. Caged-doxycycline is inactive until UV-light breaks it down into active cyano-doxycycline and a non-toxic residue 4,5-dimethoxy-2-nitrosoacetophenone (Fig. 5a). UV-irradiation and c-dox concentrations below 200 μg/ml thereby have no significant effect on cell survival (supplementary Fig. 1b). www.nature.com/scientificreports/ SFS.1-CVB3ΔVP0-IRES-Venus#9 cells were plated on 12 mm glass coverslips, inserted into newly designed and 3D printed wells with defined cavities on the bottom (Fig. 5b). C-Dox was applied to the cells for a short incubation time, then washed away. Via UV-irradiation, coming from the well bottom from a newly designed LED-Capsule (Fig. 5c), c-dox is broken down into active cyanodoxycycline, initiating the CVB3ΔVP0 expression in the illuminated cell layer. This approach creates the possibility to induce a very localized and controlled expression of the virus in cells from the SFS.1-CVB3VP0-IRES-Venus cell line in order to simulate the effects of CVB3 on local cell clusters (supplementary Fig. 3). With this tool in hand, it is now possible to observe changes in excitation spreading and contraction propagation if a signal passes CVB3 expressing cell clusters embedded in a non-induced control cell layer. As CVB3 is known to modulate cardiac ion-channels in membrane localization and function, the effects of CVB3-infected cell clusters in a non-infected monolayer are a valuable object for quantitative and qualitative analysis 16,17 . Moreover, completely new experimental setups can be designed and the pathophysiology of RNA-virus infections can be studied in more detail than ever before, especially when it comes to high-speed life cell imaging or in the use of voltage-dependent dyes.

Discussion
The newly produced human cell-line SFS.1-CVB3ΔVP0-IRES-Venus enables the investigation of RNA-virusinduced-effects in an entirely controllable, experimental setup. The complete genetic information of a noninfectious, but replicative RNA-virus strain (CVB3ΔVP0), linked to a Tet-On system for doxycycline dependent transcription, was stably transfected into the human iPS-cell line SFS.1. This setup provides the basis for noninfective and tightly controllable experiments with the RNA-virus CVB3. It is now possible to design experiments to determine the effects of different viral loads in various induction states under numerous aspects. As this RNA virus model system is human iPS-cell based, it is moreover possible to expand observations of RNA-virus effects to multiple other cell types in addition to cardiomyocytes. In the case of CVB3, the virus infects patient cells of the pancreas, the central nervous system and the respiratory tract. SFS.1-CVB3ΔVP0-IRES-Venus represents a basis to study viral infections in the desired cell types via transductions 6,18,19 .
Changes in the beating frequency of CVB3ΔVP0 induced, iPS-cell derived cardiomyocytes at different time points proved the principle and the functionality of the designed model. www.nature.com/scientificreports/ Physiological characterization of differentiated CVB3ΔVP0-expressing cardiomyocytes showed a significant increase of ROS production in the mitochondria. This observation correlates with findings of previous studies in infected mouse tissues in which elevated ROS levels are observed as well 14,15,20 . CVB3ΔVP0 expression also changed the electrical activity of the iPS-cell derived cardiomyocytes on the multi-electrode array. CVB3 induced cells showed an elongated QT-interval compared to the non-induced control cells. After supplementation of isoprenaline, non-induced cells reacted with a QT-shortening while induced cells showed no significant changes in the QT-interval. It might be, that the expression of CVB3ΔVP0 leads to a desensitization of cardiac cells towards β-adrenergic stimulation, either via proteolytic degradation of cAMP binding-sites of acceptor proteins or via continuously elevated cAMP-levels, which were found in infected mouse cardiomyocytes 15 . Continuously elevated cAMP-levels in mouse cardiomyocytes could explain the observed desensitization against β-adrenergic stimulation, which is a cAMP-activating process.
The expression of CVB3ΔVP0 in human iPS-cell derived cardiomyocytes also led to drastic changes in cell morphology and membrane structures. While non-induced cells showed a healthy phenotype with intact cell membrane and intracellular membrane structures, virus-induced cells appeared dysregulated in their membrane integrity. Intracellular membrane structures were degraded as well and many cells showed an increase of vacuolic structures. These observations are consistent with the results of previous studies, which postulated that the viroporin 2B of CVB3 is capable of disrupting membrane structures [21][22][23] . This disruption is accompanied by a degradation of mitochondria and the golgi apparatus and an increase in intracellular vesicles and vacuoles, as two independent imaging methods confirmed. Overall, the expression of CVB3ΔVP0 seems functional and well controllable in the presented system and gives the user the ability to study the effects of CVB3 in a controlled and time-dependent environment without the risk of infection. Moreover, the use of UV-activated doxycycline additionally offers for the first time the possibility of localized virus expression with a confinement of several hundred µm. This will lead to even more complex and detailed experimental designs in the future in which the influence of modeled viral plaques and infected compartments of the heart and other tissues can be assessed. The here presented system provides a novel approach to apply gene editing, in a highly controlled, versatile expression system of a pathogen in human iPS generated cell lines.