key: cord-0256422-7ucd4enw
authors: Xia, Xing; Zhang, Jimmy; Vishwanath, Manoj; Sarafan, Sadaf; Torres, Ramses Seferino Trigo; Le, Tai; Lau, Michael P.H.; Nguyen, Anh H.; Cao, Hung
title: Simultaneous Cardiac and Neurological Monitoring to Assess Chemical Exposures and Drug Toxicity in Xenopus Laevis
date: 2021-09-08
journal: bioRxiv
DOI: 10.1101/2021.09.07.459337
sha: 2ffe0e3ebf17a5d978b6207c4293b8f7b88936fc
doc_id: 256422
cord_uid: 7ucd4enw

Simultaneous monitoring of electrocardiogram (ECG) and electroencephalogram (EEG) under chemical exposure requires innovative engineering techniques that can capture minute physiological changes in studied animal models. However, this is often administered with a bulky system that may cause signal distortions and discomfort for animals. We develop an integrated bioelectronic sensing system to provide simultaneous ECG and EEG assessment in real-time under chemical exposure for Xenopus laevis. The microelectrode array (MEA) membrane with integrated ECG and EEG sensing offers an opportunity to achieve multichannel noninvasive electrophysiological monitoring with favorable dimensions and spatial resolution. To validate the performance of our system, we assessed the ECG and EEG of Xenopus under exposure of Pentylenetetrazol (PTZ), an epilepsy-inducing drug. Effects of PTZ were detected with clear ECG and EEG alterations, including frequent ictal and interictal EEG events, 30 dB average EEG amplitude elevations, abnormal ECG morphology, and heart rate changes. Overall, our Xenopus-based real-time electrophysiology monitoring system holds high potential for many applications in drug screening and remote environmental toxicity monitoring.

Pharmaceutical development involves a complicated process of biochemical assays 21 and validation processes in animal models and humans before distribution to market 22 (Barros et al. 2008 ; MacRae and Peterson 2015; Zon and Peterson 2005) . Conventionally, 23 mammalian models for preclinical toxicity, safety assessment, and side-effect screening 24 needed to be determined first before experiments could be conducted. Large mammalian 25 models, such as monkeys and dogs, were used as they possess comparable size and 26 excellent gene similarity to humans. However, they were not always the ideal choice due 27 to high cost, time consumption, and complications in conducting the screens (Denayer et 28 al. 2014). Some mammalian animals, especially dogs and nonhuman primates, faced 29 strong ethical issues for such preclinical tests (Baumans 2004) . Hence, small mammalian 30 animal models, such as rodents and rabbits, had been widely used. Although they shared 31 a number of electrophysiological similarities to humans, there were some notable 32 disparities. For instance, their heart rate (HR) was several hundred beats per minute (bpm) 33 while the respective value for humans was ~60-70 bpm, questioning their suitability for 34 studying the changes in cardiac electrophysiology under the influence of drugs. Recently, 35 other models, such as zebrafish, had been explored, owing to their fecundity, 36 morphological and physiological similarity to mammals, and the complexity of the 37 circadian clock in relation to behavioral, sleep cycle, cellular and molecular responses. 38 For example, in (Cao et al. 2014 ) and (Lee et al. 2020 ), ECG and EEG alterations were 39 evaluated for cryo-injury assessment and epilepsy studies. Nonetheless, due to the small 40 size of zebrafish, developing high spatial resolution electrophysiology recording devices 41 for long term monitoring is still challenging. Furthermore, zebrafish do not possess lungs, body, simultaneous monitoring was often implemented separately using two different 48 systems (Seitsonen et al. 2000) . As a result, it was challenging to reduce the device size 49 and bring comfort to test subjects. Alternatively, Xenopus laevis (X. laevis) has a unique 50 body structure that could facilitate non-invasive recording of electrophysiological signals 51 from the brain and heart. First, X. laevis' hairless and highly conductive skin alleviates the 52 difficulties in obtaining high signal to noise ratio (SNR) and stable biopotentials. Second, 53 anatomical features such as the favorable location of the heart, the absence of ribs, and 54 the special structure of the transverse process on vertebrae allows easier acquisition of 55 the posterior ECG than the anterior ECG, which enables the recording of ECG and EEG 56 from the same side of Xenopus body. Third, the short distance between Xenopus heart 57 and brain permits the implementation of electrodes on a single piece of flexible membrane 58 to obtain ECG and EEG signals simultaneously. These advantages provide the 59 opportunity to minimize the device size without sacrificing comfort and signal quality, 60 promoting the use of Xenopus as an alternative model for drug screening studies. PTZ compound was used as a typical model for screening potential novel antiepileptic 64 drugs and has become the approved drug (Löscher 2011 size of zebrafish brain, the electrode size must be deliberated in order to guarantee high 72 quality and localized EEG signals. In addition, careful immobilization and perfusion were 73 needed during recordings as zebrafish and tadpoles cannot stay out of water for long time, 74 increasing the complexity of obtaining EEG signals.

In this work, we developed a single flexible MEA with gold electrodes on a flexible 76 polyimide film to simultaneously assess ECG and EEG in adult X. laevis. The acquired 77 multichannel electrophysiological signals were amplified by a differential amplifier and 78 digitized by an analog to digital converter for further data analysis. Conventional lithography and wet etching processes were used to pattern the four 50-102 μm-diameter circular electrodes. After the cleaning and post baking process, the traces 103 of the electrodes were encapsulated by a layer of hardened photoresist, with the electrode 104 areas and contact pads exposed.

In compliance with the brain structure of a Xenopus, we positioned the 4-channel (Fig 1A) . The reference electrode was located above the center of the cerebellum. Three total lengths (25, 35, 109 and 45 mm) of MEAs were designed to cater to different demands of flexibility. The ECG 110 working electrodes were integrated on the MEAs, which have a total length longer than 111 35 mm. The distance between the ECG working electrode and the reference electrode 112 was 20 mm, referring to the average heart-brain distance of Xenopus we measured. The 113 ECG and EEG recording electrodes share a common reference electrode. Since EEG 114 signals have lower amplitudes than ECG signals, the baseline noises caused by brain 115 signals from cerebellum area were filtered out by additional signal processing. The The obtained EEG signals were filtered using a 6 th order Butterworth bandpass filter 178 Five animals were used for device validation and then 24 X. laevis were divided into 179 4 groups for the drug-induced epilepsy study. Each Xenopus frog was treated with drugs, 180 and performed their 4-channel EEG and ECG recording were obtained 3 separate times.

After each experiment, Xenopus were returned to circulating fresh water and allowed to 182 recover from drug or anesthesia effects for at least 2 days. Xenopus in different groups to the skin of Xenopus; thus, the conformability and longevity were significantly stronger, 193 especially for long term recordings. In some of the recordings, we found that the EEG 194 signals were noisy if the Xenopus scalp had too much mucus. We used paper towels to 195 gently wipe the area before applying electrodes, greatly improving the signal quality. The 

[ Figure 2 ] 223 The simultaneous ECG signals were recorded by the working electrodes that were 224 dorsally placed close to the heart. We also attempted to obtain ECG signals from the front were good enough to be used for morphological analysis. Fig. 3A shows an example of 237 a portion of the raw signals recorded by the ECG electrodes. The R peaks had amplitudes 238 around 250 μV, which were much higher compared with the baseline noises (less than 239 20 μV). The R peaks were definable without additional filtering, so the heart rate and R-240 R intervals could be detected from the raw signals manually. We also tried to record the 241 ECG with an individual ECG reference electrode placed on the right leg of Xenopus. While 242 it provided better ECG signals, the trade-off of the total sizes and the signal qualities made 243 it less than ideal. Fig. 3B shows the ECG signals after noise cancellation and smoothing 244 filtering. After data processing, the PQRST waves were detected clearly. The normalized 245 13 ECG of Xenopus with PQRST marked is showed in Fig. 3C . We also obtained the human 246 ECG and zebrafish ECG signals using the same system. Compared to human and 247 zebrafish ECG, Xenopus ECG exhibited one extra wave between the S and T waves.

This wave could be from PTZ-induced seizure for ECG change particularly in high-249 susceptibility frog individuals. ECG changes due to PTZ-induced seizure had not been 250 well studied in Xenopus models; however, advantages in cardiac and brain systems of 255 The ECG and EEG signals were recorded and analyzed from 4 groups of Xenopus. 

In our experiments, the heart rates, R-R intervals, QTc intervals and morphological 301 characteristics of Xenopus ECG were analyzed and compared. For the 6 Xenopus frogs 302 in the PTZ group, we calculated the average heart rate every 2 minutes, from the first to 303 the 29 th minute mark in the recording. Since the heart rate of Xenopus differed widely in 304 certain individuals (from 11 bpm to 52 bpm in our dataset), the percentage change of 305 heart rate was chosen as the parameter to be analyzed. The average heart rate of the 306 whole 30 minutes recording was calculated individually. The relative heart rate (RHR), 307 which is the heart rate at one moment in time divided by the averaged heart rate of the 308 whole recording, was used to determine the relative variation of heart rate during the Fig. 5A . By investigating the treads of the Xenopus heart rate 316 variation in PTZ group, we found that the relative heart rate did not increase when 317 seizures occurred. On the contrary, the heart rate was even lower in the Ictal stage than the PTZ has caused increases in heart rate, but the heart rate gradually returned to 328 normal. Additionally, the ictal events did not have direct relations to the heart rate. depicts, the morphology of other waves did not have significant changes besides T waves.

Another Xenopus was detected to have 3 minutes of arrhythmia when the seizures 340 happened. The ECG is shown in Fig. 6B . The arrhythmia happened right after the 341 occurrence of the first ictal event. Before and after this 3-minute interval of arrhythmia, 342 the R-R interval was around 2.8 seconds. When the arrythmia occurred, the heart beats 343 were in pairs as one fast heart rhythm and one slow rhythm. The R-R intervals were 1.6 344 seconds and 2.7 seconds, respectively. In contrast, the morphology of ECG did not have 345 noticeable changes for all subjects in the control group. EEG recordings on zebrafish were realized as well, but integrating them is still challenging.

Many of them used flexible electrodes to reduce the size and discomfort. Nevertheless, 360 most of electrophysiology recording systems were based on rodents or zebrafish. The 361 other animal models were lack of explorations.

During the first part of our work, we tried to place the ECG electrodes on different 363 locations to acquire the best signal. Initially, we considered the chest to be the best 364 location, similar to most other animals. However, we did not obtain ECG signals with 365 satisfying quality, probably because of their ventral fat. Surprisingly, the unique location 366 of Xenopus heart enabled the recording of good ECG signals dorsally. This finding 367 compelled us to combine the ECG and EEG recordings in one piece of sensor for 368 simultaneous recording. Another advantage that the Xenopus offered was their lack of 369 neck structure. This characteristic not only flattened their dorsum but also significantly 370 decreased the distance between the heart and the brain. The special body structure of 371 Xenopus enabled simultaneous ECG and EEG recordings by a set of electrodes 372 fabricated on a small piece of flexible substrate, which offered the opportunity to further 373 minimize the device size.

In our work, non-invasive electrophysiology was chosen. Due to the thin skin and 375 fat layer between the measuring spots and the signal sources, we were able to obtain 376 clear ECG and EEG signals. The drawback of the closed heart and brain was the 377 intermixing of ECG and EEG signals, but it was resolved by signal processing. In the 378 future, invasive electrophysiology with implanted electrodes may be an option as it may 379 provide signals with higher quality and stability. uncertainties for monitoring ECG alterations caused by drugs. Besides, higher heart rates 428 also led to fast drug absorption and dissipation in body. These concerns encouraged 429 researchers to find alternative models. Zebrafish was chosen by many researchers. 430 However, due to their small size, many challenges were still encountered. X. laevis had 431 a similar body size to rats, but their heart rate was much slower (11-52 in our monitoring), The filtered ECG. Now the PQRST complex are more recognizable. C. The typical ECG 620 of Xenopus, zebrafish and human recorded by the same system. 

Myocardial Iron Overload in an Experimental Model of Sudden Unexpected 477 Death in Epilepsy

Altered Cardiac Electrophysiology and 480 SUDEP in a Model of Dravet Syndrome

Performance analysis of Savitzky-Golay 482 smoothing filter using ECG signal

Zebrafish: an 485 emerging technology for in vivo pharmacological assessment to identify potential safety 486 liabilities in early drug discovery

Use of animals in experimental research: an ethical dilemma?

An amphibian with ambition: a new role for Xenopus in the 490 21st century

Epileptic seizure-induced 492 hypertension and its prevention by calcium channel blockers: a real-time study in 493 conscious telemetered rats

Effects of 496 pentylenetetrazol on GABA receptors expressed in oocytes of Xenopus laevis: extra-and 497 25 intracellular sites of action

Wearable multi-channel microelectrode 500 membranes for elucidating electrophysiological phenotypes of injured myocardium

Zebrafish as an animal model in epilepsy studies with multichannel EEG recordings

Electroencephalography in Freely Moving Mice

Comparison of methods for removing 509 electromagnetic noise from electromyographic signals

Animal models in translational medicine: 512 Validation and prediction

Burden of Arrhythmias in Epilepsy Patients: A Nationwide Inpatient Analysis of

Million Hospitalizations in the United States

Artifacts in EEG Using a Modified Independent Component Analysis Approach

Ictal tachycardia: The head-heart 520 26 connection

The right thalamus may 522 play an important role in anesthesia-awakening regulation in frogs

How can we identify ictal and interictal 524 abnormal activity?

Catecholamine-induced transient 526 myocardial dysfunction

In vivo imaging of 528 seizure activity in a novel developmental seizure model

Pentylenetetrazole-induced inhibition of recombinant gamma-aminobutyric acid type A 532 (GABA(A)) receptors: mechanism and site of action

Chemical toxins that cause seizures

Diagnostic Role of ECG Recording Simultaneously With EEG Testing

Cardiac Arrhythmia 539 Following an Epileptic Seizure

An EEG system to detect brain 541 signals from multiple adult zebrafish

ECG Diagnosis: Isolated Posterior Wall Myocardial Infarction

Critical review of current animal models of seizures and epilepsy used 545 in the discovery and development of new antiepileptic drugs

Zebrafish as tools for drug discovery

Stage-dependent cardiac regeneration in 550 <em>Xenopus</em> is regulated by thyroid hormone availability

Simultaneous Video-EEG-ECG

Monitoring to Identify Neurocardiac Dysfunction in Mouse Models of Epilepsy

Blood Pressure in Seizures and 556 Epilepsy

Cardiac effects of seizures

A Wireless 559 EEG Recording Method for Rat Use inside the Water Maze

Pentylenetetrazole-induced seizures in developing rats 562 prenatally exposed to valproic acid

The augmentor action of the sympathetic cardiac 564 nerves

Valproic Acid and Epilepsy: From Molecular Mechanisms 567 to Clinical Evidences

Hippocampal-dependent learning requires a functional circadian system

The early development and 573 physiology of &lt

Smoothing and differentiation of data by simplified least 576 squares procedures

Are electrocardiogram electrodes 578 acceptable for electroencephalogram bispectral index monitoring?

EEG in neurological conditions other than epilepsy: when does it 581 help, what does it add

EEG in the diagnosis, classification, and management of patients 584 with epilepsy

Removal of ECG interference 586 from the EEG recordings in small animals using independent component analysis

Electrophysiological assessment of plant status outside a Faraday 590 cage using supervised machine learning

Fragile X mental retardation protein knockdown in the developing 593

Xenopus tadpole optic tectum results in enhanced feedforward inhibition and behavioral 594 deficits

Potential Targets for Therapeutic Intervention in Epilepsy

Ictal and 598 interictal high frequency oscillations in patients with focal epilepsy

In vivo drug discovery in the zebrafish