Development of microfluidic devices to model Heart-Brain interactions in Torsades de Pointes – NeuroCard

Here, we postulate that a decrease in IKr could result in lengthened cardiac ventricular repolarisation and hyperexcitable AMNs. These combined alterations could synergistically trigger TdP, justifying the study of neurocardiac interactions in this particular arrhythmogenic process using a suited approach. This bring a new paradigm to decipher fundamental mechanisms of hERG-related SD induced by TdP.

The neurocardiac axis is usually studied in whole animals, but these models often produce confusing results due to complex neuron–cardiomyocyte (CM) interactions and species differences in ion channels and arrhythmia risk. Human induced pluripotent stem cell (hiPSC) technology offers a better alternative, allowing researchers to derive patient-specific neurons and CMs for controlled studies. However, hiPSC-derived cells are still immature, limiting their use in pharmacology and proarrhythmia testing. Co-cultures combining autonomic motor neurons (AMNs) and CMs have been explored to improve cell maturity and function through direct contact and secreted factors in a unique dish. Yet, these traditional co-cultures are hard to manage because the cell types need different media and conditions, causing inconsistent phenotypes and experimental bias. This makes it difficult to design reliable pharmacological studies using mixed-cell systems. So, we aim to generate both autonomic motor neurons and ventricular cardiomyocytes from healthy and LQT2 patients hiPSC lines.

Then, we propose to develop a new tool to investigate neurocardiac axis based on compartmentalized co-culture of hiPSC-AMNs and hiPSC-CMs to model the neurocardiac network. This has become recently possible by the introduction of microfluidic chamber (µFC) technology where 2 or more individualized compartments are connected through microchannels.

With this tool, we propose to characterize, individually, the healthy and LQT2 hiPSC-AMNs and hiPSC-CMs from structural to electrophysiological and calcium levels. We test the feasibility to establish functional neurocardiac coupling in this device and test the efficacy of our in vitro model to identify the torsadogenicity of drugs.

As part of this study, three types of compartmentalized microfluidic devices were developed and optimized: the MBBT chips called Brainies™, designed for “classical” neuron and cardiomyocyte cocultures; the MBBT chips called Mindies™, specifically designed for direct coupling with microelectrode arrays (MEAs); and finally, a third type of chip referred to as “3C,” incorporating a central compartment intended for complementary experiments, enabling the selective severing of axons using Triton, that serve as control experiment. Each chip consists of two culture chambers, each connected to two reservoirs (the left one being the neuronal compartment and the right one the cardiac compartment) and separated by asymmetric microchannels to prevent cells from crossing between compartments. We generate ventricular cardiomyocytes, as well autonomic nervous motorneurone (sympathetic) from healthy and LQT2 hiPSC lines and structurally and fonctionnaly characterized them. We succeeded to develop a organ on chip resulting in hIPSC-cardiomyocytes innervated by sympathetic motoneurons. Motorneurones are able to modulate the excitation-contraction coupling of the cardiomyocytes through neurotransmiters. We used two co-culture conditions of the two cell types, either derived from the same hiPSC line, i.e., healthy or mutated, which we compared. We also used a ‘mixed’ condition combining healthy neurons with LQT2 cardiomyocytes, and vice versa. The results obtained with cell types derived from the same genotype indicate that once cardiomyocytes are innervated by sympathetic neurons, the beating frequency of healthy line increased whereas using the LQT2 genotype, it significantly reduced, suggesting a profund remodeling of neucocardiac junction in LQT2. The decay time of calcium transients in LQT2 co-cultures is markedly increased compared to the healthy condition, leading to a much higher incidence of cellular arrhythmias compared with the healthy condition. Interestingly, in the presence of healthy neurons, the proportion of LQT2 cardiomyocytes exhibiting arrhythmias tends to decrease.

In parallel, using healthy cells (neurons innervated cardiomyocytes) in the microfluidic device, we tested the efficiently of this devices to highlight a iatrogenic risk. This system has proven highly effective in identifying the arrhythmogenic potential of risky compounds, without producing false negatives, and in detecting compounds with intermediate risk—capabilities that the threshold hIPSC cardiomyocyte model did not demonstrate.

We succeeded in generating human cardiomyocytes which express hERG protein and present adult ventricular profile with respect to both structure (sarcomeric arrangements) and function (excitation-contraction coupling). These cardiomyocytes were derived from an healthy and an LQT2 line. In monoculture and with the pharmacology, we fully characterized both healthy and LQT2 type cardiomyocytes from structure, hERG expression and localization, action potential, rhythmic profiles, to excitation-contraction coupling. We also characterize the sympathetic neurons derived from healthy and LQT2 hIPSCs. Academic partners and Microbrain Biotech develop microfluidic devices and we successed to generate cardiomyocytes innervated remotely by sympathetic neurons in two compartments separated by microchannels. Thanks to these devices, we thus proposed a mechanism by which our mutation is responsible for a arrhythmogenic phenotype. Thanks to fluidic control, we also realized compartmentalized pharmacology on the cardiomyocytes or neurons, or the combined outcome of both of several compounds. We tested the technical and scientific capability of the OOC, using healthy hIPSC derived neurons and cardiomyocytes to predict an arrhythmic risk. For that, reference compounds were used and we shown that our methodology is able to predict high risk compound (no false negative), without no false positive. In addition, it sort NS1643 as risk class compound whereas this compound was not highlight on isolated cardiomyocytes.

As perspective, we are differentiating parasympathetic neurons and characterize them in mono- and co-cultue in microfluidic devices. By integrating both parasympathetic and sympathetic neurons, the new configuration of our model will provide a more predictive platform for assessing the pathogenicity of specific gene variants and for the early detection of the torsadogenic New chemical entities.

Torsades de Pointes (TdPs) are malignant ventricular arrhythmia leading to sudden death. TdPs are observed in active patients without any structural cardiac defect, thus emphasizing the need of an early detection of “at-risk” patients. TdP is manifested by an abnormal lengthening of the ventricular repolarization (VR) on the electrocardiogram (long QT interval). To date, the pathophysiological mechanisms of TdPs are poorly understood. Of note, we have shown that activation of the autonomic nervous system is involved in TdPs, as an abnormal VR only is not sufficient to trigger this fatal event. This is true for the hereditary long QT syndromes (LQTS) and drugs blocking the potassium HERG channel (acquired LQTS).
The excitability of cardiomyocytes (CMs) and autonomic motor neurons (AMNs) depends on ion channels, including the hERG channel. HERG mutations lead to TdPs in LQTS type 2 (LQT2). LQT2 patients commonly present neurological comorbidities, especially epilepsy, identifying LQT2 as a neurocardiac disease. To date, studies exploring AMN modulation of CMs are sparse and it remains unknown whether the hERG depression in AMNs plays a role in the cardiac LQT2 symptoms. Besides, to estimate TdPs risk of drug candidates before entering human clinical trials, regulatory guidelines request to identify their hERG-blocking properties. However, it appears that the current safety HERG assays, using CMs or HERG-expressing cell lines, are not fully satisfactory to detect compounds exhibiting TdPs risk, because of false-positive and false-negative results reported. This pitfall may be due to AMN involvement, not considered in the current assays.
The intrinsic properties of AMNs and their ability to modulate CMs are poorly known, essentially because of the difficulty to model the neurocardiac interactions. To address this issue, the NEUROCARD multidisciplinary consortium merges experts in electrophysiology and Ca2+ signaling, in human-induced pluripotent stem cells (hiPSCs) both in cardiology and neuroscience, associated with a Neurotech Company specialized in the development and commercialization of ready-to-use microfluidics devices. NEUROCARD aims at building a biological on-chip AMN-CM network to study the neurocardiac communications in normal and diseased states (LQT2 and acquired TdPs). NEUROCARD has 3 main objectives. (1) To develop a new device using the microfluidics technology to co-culture hiPSC-derived CMs innervated by hiPSC-derived AMNs (sympathetic or/and cholinergic neurons) in microfluidic-connected chambers (µFCs). Such a device is critically needed for tailored pharmacology and toxicology investigations, as it allows compartmentalized application of drugs, as well as molecular, cellular, and detailed functional analysis of the electrophysiological and Ca2+ signaling responses of each cell type. The AMN-CM µFC co-cultures will be here evaluated as a new HERG-safety assay towards identifying better torsadogenicity biomarkers. (2) To establish LQT2 (patient vs gene-corrected isogenic control) hiPSCs and study the intrinsic properties of hiPSC-derived AMNs and CMs. We will here provide new knowledge about the molecular and functional alterations in AMNs derived from one LQT2 patient with both neurological and cardiac symptoms. (3) To decipher the neurocardiac signature in LQT2 using µFC co-cultures of hiPSCs-CMs and AMNs. The hiPSC-derived AMNs and CMs (LQT2 versus isogenic-corrected) will be ‘mixed and matched’ in the µFCs to unravel the precise molecular and functional interactions between AMNs and CMs and identify the AMN-CM dysfunction in LQT2.
Overall, the technological developments promoted by this project (µFCs, LQT2 hiPSC lines) will offer a unique opportunity to decipher the neurocardiac dysfunction underlying TdP-causing conditions. This project should also help evaluate new drugs for cardiac diseases with the final aim to pave the way for precision medicine-based prevention of TdPs.