Preclinical cardiac safety assessment of drugs

A novel deep sequential learning architecture for drug drug interaction prediction using DDINet

Drug drug Interactions (DDI) present considerable challenges in healthcare, often resulting in adverse effects or decreased therapeutic efficacy. This article proposes a novel deep sequential learning architecture called DDINet to predict and classify DDIs between pairs of drugs based on different mechanisms viz., Excretion, Absorption, Metabolism, and Excretion rate (higher serum level) etc. Chemical features such as Hall Smart, Amino Acid count and Carbon types are extracted from each drug (pairs) to apply as an input to the proposed model. Proposed DDINet incorporates attention mechanism and deep sequential learning architectures, such as Long Short-Term Memory and gated recurrent unit. It utilizes the Rcpi toolkit to extract biochemical features of drugs from their chemical composition in Simplified Molecular-Input Line-Entry System format. Experiments are conducted on publicly available DDI datasets from DrugBank and Kaggle. The model's efficacy in predicting and classifying DDIs is evaluated using various performance measures. The experimental results show that DDINet outperformed eight counterpart techniques achieving [Formula: see text] overall accuracy which is also statistically confirmed by Confidence Interval tests and paired t-tests. This architecture may act as an effective computational technique for drug drug interaction with respect to mechanism which may act as a complementary tool to reduce costly wet lab experiments for DDI prediction and classification.

Evaluating the proarrhythmic risk of delayed-action compounds in serum free cell culture conditions; serum-starvation accelerates/amplifies the effect of probucol on the KCNQ1 + KCNE1 channel

In vitro testing procedures for evaluating acute effects of compound on ion channels, utilizing heterologous expression systems (HES), are well-established, while slowly manifesting delayed effects remain challenging to detect. For this, immortalized HES are exposed to the compounds for a longer time, in general 24 h. As these cells proliferate every 12-20 h, we evaluated if the proliferation status, and by extension cell metabolism, influences the delayed compound response. The intervention of halting cell proliferation by excluding serum from the culturing medium was evaluated on CHO cells, stably expressing the KCNQ1 + KCNE1 channel complex that mediates the slow delayed rectifier potassium current (Iks). No abnormal changes in KCNQ1 + KCNE1 current were observed upon serum-starvation, except for a negative shift in the voltage dependence of channel activation (GV-curve) after 72 h. The delayed effect of probucol, a compound reported to interfere with Iks expression, was evaluated after 24 and 72 h of incubation. In serum-free conditions the inhibitory effect of probucol was increased fourfold after 24 h, compared to serum supplemented conditions. After 72 h, the current inhibition was similar between both culture conditions. Besides decreasing current expression, probucol shifted the GV-curve more positive combined with a shallower voltage response, changes that were more pronounced in serum-depleted conditions. The results indicated that serum-starvation had no substantial effect on the KCNQ1 + KCNE1 current in the tested CHO cells, but it amplified or accelerated the response to probucol, suggesting that halting cell proliferation is a method for enhancing the detection of delayed compound effects in HES.

Challenging the status quo: a framework for mechanistic and human-relevant cardiovascular safety screening

Traditional approaches to preclinical drug safety assessment have generally protected human patients from unintended adverse effects. However, these assessments typically occur too late to make changes in the formulation or in phase 1 and beyond, are highly dependent on animal studies and have the potential to lead to the termination of useful drugs due to liabilities in animals that are not applicable in patients. Collectively, these elements come at great detriment to both patients and the drug development sector. This phenomenon is particularly problematic in the area of cardiovascular safety assessment where preclinical attrition is high. We believe that a more efficient and translational approach can be defined. A multi-tiered assessment that leverages our understanding of human cardiovascular biology, applies human cell-based in vitro characterizations of cardiovascular responses to insult, and incorporates computational models of pharmacokinetic relationships would enable earlier and more translational identification of human-relevant liabilities. While this will take time to develop, the ultimate goal would be to implement such assays both in the lead selection phase as well as through regulatory phases.

Scientific and Regulatory Policy Committee Points to Consider: Integration of Clinical Pathology Data With Anatomic Pathology Data in Nonclinical Toxicology Studies

Integrating clinical pathology data with anatomic pathology data is a common practice when reporting findings in the context of nonclinical toxicity studies and aids in understanding and communicating the nonclinical safety profile of test articles in development. Appropriate pathology data integration requires knowledge of analyte and tissue biology, species differences, methods of specimen acquisition and analysis, study procedures, and an understanding of the potential causes and effects of a variety of pathophysiologic processes. Neglecting these factors can lead to inappropriate data integration or a missed opportunity to enhance understanding and communication of observed changes. In such cases, nonclinical safety information relevant to human safety risk assessment may be misrepresented or misunderstood. This "Points to Consider" manuscript presents general concepts regarding pathology data integration in nonclinical studies, considerations for avoiding potential oversights and errors in data integration, and focused discussion on topics relevant to data integration for several key organ systems, including liver, kidney, and cardiovascular systems.

Scientific and Regulatory Policy Committee Points to Consider: Integration of Clinical Pathology Data With Anatomic Pathology Data in Nonclinical Toxicology Studies

This article is temporarily under embargo.

Simultaneous triple-parametric optical mapping of transmembrane potential, intracellular calcium and NADH for cardiac physiology assessment

Investigation of the complex relationships and dependencies of multiple cellular processes that govern cardiac physiology and pathophysiology requires simultaneous dynamic assessment of multiple parameters. In this study, we introduce triple-parametric optical mapping to simultaneously image metabolism, electrical excitation, and calcium signaling from the same field of view and demonstrate its application in the field of drug testing and cardiovascular research. We applied this metabolism-excitation-contraction coupling (MECC) methodology to test the effects of blebbistatin, 4-aminopyridine and verapamil on cardiac physiology. While blebbistatin and 4-aminopyridine alter multiple aspects of cardiac function suggesting off-target effects, the effects of verapamil were on-target and it altered only one of ten tested parameters. Triple-parametric optical mapping was also applied during ischemia and reperfusion; and we identified that metabolic changes precede the effects of ischemia on cardiac electrophysiology.

In silico approaches in organ toxicity hazard assessment: Current status and future needs for predicting heart, kidney and lung toxicities

The kidneys, heart and lungs are vital organ systems evaluated as part of acute or chronic toxicity assessments. New methodologies are being developed to predict these adverse effects based on in vitro and in silico approaches. This paper reviews the current state of the art in predicting these organ toxicities. It outlines the biological basis, processes and endpoints for kidney toxicity, pulmonary toxicity, respiratory irritation and sensitization as well as functional and structural cardiac toxicities. The review also covers current experimental approaches, including off-target panels from secondary pharmacology batteries. Current in silico approaches for prediction of these effects and mechanisms are described as well as obstacles to the use of in silico methods. Ultimately, a commonly accepted protocol for performing such assessment would be a valuable resource to expand the use of such approaches across different regulatory and industrial applications. However, a number of factors impede their widespread deployment including a lack of a comprehensive mechanistic understanding, limited in vitro testing approaches and limited in vivo databases suitable for modeling, a limited understanding of how to incorporate absorption, distribution, metabolism, and excretion (ADME) considerations into the overall process, a lack of in silico models designed to predict a safe dose and an accepted framework for organizing the key characteristics of these organ toxicants.

High-Throughput Screening to Identify Chemical Cardiotoxic Potential

Cardiovascular (CV) disease is one of the most prevalent public health concerns, and mounting evidence supports the contribution of environmental chemicals to CV disease burden. In this study, we performed cardiotoxicity profiling for the Tox21 chemical library by focusing on high-throughput screening (HTS) assays whose targets are associated with adverse events related to CV failure modes. Our objective was to develop new hypotheses around environmental chemicals of potential interest for adverse CV outcomes using Tox21/ToxCast HTS data. Molecular and cellular events linked to six failure modes of CV toxicity were cross-referenced with 1399 Tox21/ToxCast assays to identify cardio-relevant bioactivity signatures. The resulting 40 targets, measured in 314 assays, were integrated via a ToxPi visualization tool and ranking system to prioritize 1138 chemicals based upon formal integration across multiple domains of information. Filtering was performed based on cytotoxicity and generalized cell stress endpoints to try and isolate chemicals with effects specific to CV biology, and bioactivity- and structure-based clustering identified subgroups of chemicals preferentially affecting targets such as ion channels and vascular tissue biology. Our approach identified drugs with known cardiotoxic effects, such as estrogenic modulators like clomiphene and raloxifene, anti-arrhythmic drugs like amiodarone and haloperidol, and antipsychotic drugs like chlorpromazine. Several classes of environmental chemicals such as organotins, bisphenol-like chemicals, pesticides, and quaternary ammonium compounds demonstrated strong bioactivity against CV targets; these were compared to existing data in the literature (e.g., from cardiomyocytes, animal data, or human epidemiological studies) and prioritized for further testing.

Drug-induced myocardial dysfunction - recommendations for assessment in clinical and pre-clinical studies

Introduction: Drug-induced myocardial dysfunction is an important safety concern during drug development. Oncology compounds can cause myocardial dysfunction, leading to decreased left ventricular ejection fraction and heart failure via several mechanisms. Cardiovascular imaging has a major role in the early detection and monitoring of cardiotoxicity. Echocardiography is the method of choice because of its widespread availability, low cost, and absence of radiation exposure. Cardiac magnetic resonance imaging can provide better reliability, reproducibility, and accuracy in the detection of drug-induced myocardial dysfunction. In addition, it enables assessment of myocardial edema, fibrosis, and necrosis. Cardiac serologic biomarkers such as troponins and B-type natriuretic peptides are used in combination with imaging during drug development. This article provides a general overview of each imaging modality and practical guidance for early detection and monitoring of cardiotoxicity.Areas covered: Cardiovascular imaging modalities and cardiac biomarkers for monitoring of cardiac function and early detection of drug-induced myocardial dysfunction in drug development.Expert opinion: Some new drugs especially in the oncology field, can cause myocardial dysfunction. Depending on the strength of pre-clinical or clinical data, CV imaging modalities and cardiac biomarkers play an important role in the early detection and mitigation plans for such drugs during their development.

Towards chamber specific heart-on-a-chip for drug testing applications

Modeling of human organs has long been a task for scientists in order to lower the costs of therapeutic development and understand the pathological onset of human disease. For decades, despite marked differences in genetics and etiology, animal models remained the norm for drug discovery and disease modeling. Innovative biofabrication techniques have facilitated the development of organ-on-a-chip technology that has great potential to complement conventional animal models. However, human organ as a whole, more specifically the human heart, is difficult to regenerate in vitro, in terms of its chamber specific orientation and its electrical functional complexity. Recent progress with the development of induced pluripotent stem cell differentiation protocols, made recapitulating the complexity of the human heart possible through the generation of cells representative of atrial & ventricular tissue, the sinoatrial node, atrioventricular node and Purkinje fibers. Current heart-on-a-chip approaches incorporate biological, electrical, mechanical, and topographical cues to facilitate tissue maturation, therefore improving the predictive power for the chamber-specific therapeutic effects targeting adult human. In this review, we will give a summary of current advances in heart-on-a-chip technology and provide a comprehensive outlook on the challenges involved in the development of human physiologically relevant heart-on-a-chip.

Drug-Drug Interaction Predicting by Neural Network Using Integrated Similarity

Drug-Drug Interaction (DDI) prediction is one of the most critical issues in drug development and health. Proposing appropriate computational methods for predicting unknown DDI with high precision is challenging. We proposed "NDD: Neural network-based method for drug-drug interaction prediction" for predicting unknown DDIs using various information about drugs. Multiple drug similarities based on drug substructure, target, side effect, off-label side effect, pathway, transporter, and indication data are calculated. At first, NDD uses a heuristic similarity selection process and then integrates the selected similarities with a nonlinear similarity fusion method to achieve high-level features. Afterward, it uses a neural network for interaction prediction. The similarity selection and similarity integration parts of NDD have been proposed in previous studies of other problems. Our novelty is to combine these parts with new neural network architecture and apply these approaches in the context of DDI prediction. We compared NDD with six machine learning classifiers and six state-of-the-art graph-based methods on three benchmark datasets. NDD achieved superior performance in cross-validation with AUPR ranging from 0.830 to 0.947, AUC from 0.954 to 0.994 and F-measure from 0.772 to 0.902. Moreover, cumulative evidence in case studies on numerous drug pairs, further confirm the ability of NDD to predict unknown DDIs. The evaluations corroborate that NDD is an efficient method for predicting unknown DDIs. The data and implementation of NDD are available at https://github.com/nrohani/NDD.

Furthering the state of knowledge on the electric properties of hemi-ellipsoidal single cells and cell patches on electrodes

The impedance of electrodes with adherent biological cells correlates with cell viability and proliferation. To model this correlation, we exploited the idea that the introduction of a highly conductive layer into the equatorial equipotential slice of a system with an oriented, freely suspended, single ellipsoidal cell may split the system into mirror-symmetrical halves without changing the field distribution. Each half possesses half of the system's impedance and contains a hemiellipsoidal cell attached to the conductive layer, which can be considered a bottom electrode. For a hemiellipsoidal adherent cell model (ACM) with standard electrical properties for the external and cellular media, the assumption of a bottom membrane and a subcellular cleft in the 100 nm range, as found in adherent cells, changed the potential distribution over a one-% range up to frequencies of 1 MHz. For simplicity, potential distributions for slices of spheroidal objects can be numerically calculated in 2D. The 2D distributions can be converted into three dimensions using simplified equations for the influential radii of spheroids. After the ACM approach was expanded to adherent cell patch models (APMs), the feasibility of our model modifications was tested using two criteria: the constancy of the equipotential plane touching the poles of ACMs or APMs and a comparison of the impedance, which could be numerically calculated from the overall current between the bottom electrode and a plane-parallel counter-electrode, with the impedance of the suspension obtained from Maxwell-Wagner's mixing equation applied to hemiellipsoidal cells.

Cardiotoxicity of ErbB2-targeted therapies and its impact on drug development, a spotlight on trastuzumab

INTRODUCTION

Trastuzumab, a therapeutic monoclonal antibody directed against ErbB2, is often noted as a successful example of targeted therapy. Trastuzumab improved outcomes for many patients with ErbB2-positive breast and gastric cancers, however, cardiac side effects [e.g., left ventricular dysfunction and congestive heart failure (CHF)] were reported in the early phase clinical studies. This finding, subsequently corroborated by multiple clinical studies, raised concerns that the observed cardiotoxicity induced by trastuzumab might adversely impact the clinical development of other therapeutics targeting ErbB family members. Areas covered: In this review we summarize both basic research and clinical findings regarding trastuzumab-induced cardiotoxicity and assess if there has been an impact of trastuzumab-induced cardiotoxicity on the development of other agents targeting ErbB family members. Expert opinion: There are a number of scientific gaps that are critically important to address for the continued success of HER2-targeted agents. These include: 1) elucidating the molecular mechanisms contributing to cardiotoxicity; 2) developing relevant preclinical testing systems for predicting cardiotoxicity; 3) developing clinical strategies to identify patients at risk of cardiotoxicity; and 4) enhancing management of clinical symptoms of cardiotoxicity.

Data-driven prediction of adverse drug reactions induced by drug-drug interactions

Background

The expanded use of multiple drugs has increased the occurrence of adverse drug reactions (ADRs) induced by drug-drug interactions (DDIs). However, such reactions are typically not observed in clinical drug-development studies because most of them focus on single-drug therapies. ADR reporting systems collect information on adverse health effects caused by both single drugs and DDIs. A major challenge is to unambiguously identify the effects caused by DDIs and to attribute them to specific drug interactions. A computational method that provides prospective predictions of potential DDI-induced ADRs will help to identify and mitigate these adverse health effects.

Method

We hypothesize that drug-protein interactions can be used as independent variables in predicting ADRs. We constructed drug pair-protein interaction profiles for ~800 drugs using drug-protein interaction information in the public domain. We then constructed statistical models to score drug pairs for their potential to induce ADRs based on drug pair-protein interaction profiles.

Results

We used extensive clinical database information to construct categorical prediction models for drug pairs that are likely to induce ADRs via synergistic DDIs and showed that model performance deteriorated only slightly, with a moderate amount of false positives and false negatives in the training samples, as evaluated by our cross-validation analysis. The cross validation calculations showed an average prediction accuracy of 89% across 1,096 ADR models that captured the deleterious effects of synergistic DDIs. Because the models rely on drug-protein interactions, we made predictions for pairwise combinations of 764 drugs that are currently on the market and for which drug-protein interaction information is available. These predictions are publicly accessible at http://avoid-db.bhsai.org. We used the predictive models to analyze broader aspects of DDI-induced ADRs, showing that ~10% of all combinations have the potential to induce ADRs via DDIs. This allowed us to identify potential DDI-induced ADRs not yet clinically reported. The ability of the models to quantify adverse effects between drug classes also suggests that we may be able to select drug combinations that minimize the risk of ADRs.

Conclusion

Almost all information on DDI-induced ADRs is generated after drug approval. This situation poses significant health risks for vulnerable patient populations with comorbidities. To help mitigate the risks, we developed a robust probabilistic approach to prospectively predict DDI-induced ADRs. Based on this approach, we developed prediction models for 1,096 ADRs and used them to predict the propensity of all pairwise combinations of nearly 800 drugs to be associated with these ADRs via DDIs. We made the predictions publicly available via internet access.

Electronic supplementary material

The online version of this article (doi:10.1186/s40360-017-0153-6) contains supplementary material, which is available to authorized users.

Cardiovascular Safety in Drug Development

As drug development becomes a long and demanding process, it might also become a barrier to medical progress. Drug safety concerns are responsible for many of the resources consumed in launching a new drug. Despite the money and time expended on it, a significant number of drugs are withdrawn years or decades after being in the market. Cardiovascular toxicity is one of the major reasons for those late withdrawals, meaning that many patients are exposed to unexpected serious cardiovascular risks. It seems that current methods to assess cardiovascular safety are imperfect, so new approaches to avoid the exposure to those undesirable effects are quite necessary. Endothelial dysfunction is the earliest detectable pathophysiological abnormality, which leads to the development of atherosclerosis, and it is also an independent predictor for major cardiovascular events. Endothelial toxicity might be the culprit of the cardiovascular adverse effects observed with a significant number of drugs. In this article, we suggest the regular inclusion of the best validated and less invasive endothelial function tests in the clinical phases of drug development in order to facilitate the development of drugs with safer cardiovascular profiles.

Age and gender dependent heart rate circadian model development and performance verification on the proarrhythmic drug case study

BACKGROUND

There are two main reasons for drug withdrawals at the various levels of the development path - hepatic and cardiac toxicity. The latter one is mainly connected with the proarrhythmic potency and according to the present practice is supposed to be recognized at the pre-clinical (in vitro and animal in vivo) or clinical level (human in vivo studies). There are, although, some limitations to all the above mentioned methods which have led to novel in vitro - in vivo extrapolation methods being introduced. With the use of in silico implemented mathematical and statistical modelling it is possible to translate the in vitro findings into the human in vivo situation at the population level. Human physiology is influenced by many parameters and one of them which needs to be properly accounted for is a heart rate which follows the circadian rhythm. We described such phenomenon statistically which enabled the improved assessment of the drug proarrhythmic potency.

METHODS

A publicly available data set describing the circadian changes of the heart rate of 18 healthy subjects, 5 males (average age 36, range 26-45) and 13 females (average age 34, range 20-50) was used for the heart rate model development. External validation was done with the use of a clinical research database containing heart rate measurements derived from 67 healthy subjects, 34 males and 33 females (average age 33, range 17-72). The developed heart rate model was then incorporated into the ToxComp platform to simulate the impact of circadian variation in the heart rate on QTc interval. The usability of the combined models was assessed with moxifloxacin (MOXI) as a model drug.

RESULTS

The developed heart rate model fitted well, both to the training data set (RMSE = 128 ms and MAPE = 12.3%) and the validation data set (RMSE = 165 ms and MAPE = 17.1%). Simulations performed at the population level proved that the combination of the IVIVE platform and the population variability description allows for the precise prediction of the circadian variation of drugs proarrhythmic effect.

CONCLUSIONS

It can be concluded that a flexible and practically useful model describing the heart rate circadian variation has been developed and its performance was verified.

A selective Akt inhibitor produces hypotension and bradycardia in conscious rats due to inhibition of the autonomic nervous system

Akt is a serine-threonine kinase that is amplified in a variety of human cancers, and as with other anticancer agents, some Akt inhibitors have produced functional cardiovascular effects such as marked hypotension that may limit their clinical benefit. Although identified in preclinical studies, the mechanism(s) responsible for these effects are often not fully characterized; potential targets include Akt signaling disruption in cardiac tissue, vascular smooth muscle, and/or autonomic system signaling. A selective Akt inhibitor was found to produce a rapid and marked hypotension and bradycardia in conscious rats. Isolated right atrial tissue and isolated thoracic aortic rings were used to examine direct effects of Akt inhibition on cardiac and vascular tissues, respectively. In addition, rats surgically prepared with telemetry units for monitoring blood pressure and heart rate were used to investigate potential effects on the autonomic nervous system (ANS). Whereas this Akt inhibitor did not produce any significant effect on atrial tissue, it did cause vasorelaxation of aortic rings. More significantly, in conscious rats, the Akt inhibitor inhibited the neural pressor response to the known nicotinic acetylcholine receptor (nAchR) agonist dimethylphenylpiperazinium (DMPP). In fact, the response observed was comparable to the response observed with the known ganglionic blocker hexamethonium. Thus, the hypotension and bradycardia produced by the Akt inhibitor is primarily due to blockade of nAchRs in autonomic ganglia. This finding highlights the importance of evaluating the ANS for cardiovascular effects associated with new chemical entities as well as suggesting a novel direct effect of an Akt inhibitor on nAchRs.

Echocardiographic speckle-tracking based strain imaging for rapid cardiovascular phenotyping in mice

RATIONALE

High-sensitivity in vivo phenotyping of cardiac function is essential for evaluating genes of interest and novel therapies in small animal models of cardiovascular disease. Transthoracic echocardiography is the principal method currently used for assessing cardiac structure and function; however, standard echocardiographic techniques are relatively insensitive to early or subtle changes in cardiac performance, particularly in mice.

OBJECTIVE

To develop and validate an echocardiographic strain imaging methodology for sensitive and rapid cardiac phenotyping in small animal models.

METHODS AND RESULTS

Herein, we describe a modified echocardiographic technique that uses speckle-tracking based strain analysis for the noninvasive evaluation of cardiac performance in adult mice. This method is found to be rapid, reproducible, and highly sensitive in assessing both regional and global left ventricular (LV) function. Compared with conventional echocardiographic measures of LV structure and function, peak longitudinal strain and strain rate were able to detect changes in adult mouse hearts at an earlier time point following myocardial infarction and predicted the later development of adverse LV remodeling. Moreover, speckle-tracking based strain analysis was able to clearly identify subtle improvement in LV function that occurred early in response to standard post-myocardial infarction cardiac therapy.

CONCLUSIONS

Our results highlight the utility of speckle-tracking based strain imaging for detecting discrete functional alterations in mouse models of cardiovascular disease in an efficient and comprehensive manner. Echocardiography speckle-tracking based strain analysis represents a method for relatively high-throughput and sensitive cardiac phenotyping, particularly in evaluating emerging cardiac agents and therapies in mice.

Successful drug development despite adverse preclinical findings part 2: examples

To illustrate the process of addressing adverse preclinical findings (APFs) as outlined in the first part of this review, a number of cases with unexpected APF in toxicity studies with drug candidates is discussed in this second part. The emphasis is on risk characterization, especially regarding the mode of action (MoA), and risk evaluation regarding relevance for man. While severe APFs such as retinal toxicity may turn out to be of little human relevance, minor findings particularly in early toxicity studies, such as vasculitis, may later pose a real problem. Rodents are imperfect models for endocrine APFs, non-rodents for human cardiac effects. Liver and kidney toxicities are frequent, but they can often be monitored in man and do not necessarily result in early termination of drug candidates. Novel findings such as the unusual lesions in the gastrointestinal tract and the bones presented in this review can be difficult to explain. It will be shown that well known issues such as phospholipidosis and carcinogenicity by agonists of peroxisome proliferator-activated receptors (PPAR) need to be evaluated on a case-by-case basis. The latter is of particular interest because the new PPAR α and dual α/γ agonists resulted in a change of the safety paradigm established with the older PPAR α agonists. General toxicologists and pathologists need some understanding of the principles of genotoxicity and reproductive toxicity testing. Both types of preclinical toxicities are major APF and clinical monitoring is difficult, generally leading to permanent use restrictions.