Mechanism-based organization of neural networks to emulate systems biology and pharmacology models.

Deep learning neural networks are often described as black boxes, as it is difficult to trace model outputs back to model inputs due to a lack of clarity over the internal mechanisms. This is even true for those neural networks designed to emulate mechanistic models, which simply learn a mapping between the inputs and outputs of mechanistic models, ignoring the underlying processes. Using a mechanistic model studying the pharmacological interaction between opioids and naloxone as a proof-of-concept example, we demonstrated that by reorganizing the neural networks' layers to mimic the structure of the mechanistic model, it is possible to achieve better training rates and prediction accuracy relative to the previously proposed black-box neural networks, while maintaining the interpretability of the mechanistic simulations. Our framework can be used to emulate mechanistic models in a large parameter space and offers an example on the utility of increasing the interpretability of deep learning networks.

Deep learning-enabled natural language processing to identify directional pharmacokinetic drug-drug interactions.

BACKGROUND: During drug development, it is essential to gather information about the change of clinical exposure of a drug (object) due to the pharmacokinetic (PK) drug-drug interactions (DDIs) with another drug (precipitant). While many natural language processing (NLP) methods for DDI have been published, most were designed to evaluate if (and what kind of) DDI relationships exist in the text, without identifying the direction of DDI (object vs. precipitant drug). Here we present a method for the automatic identification of the directionality of a PK DDI from literature or drug labels. METHODS: We reannotated the Text Analysis Conference (TAC) DDI track 2019 corpus for identifying the direction of a PK DDI and evaluated the performance of a fine-tuned BioBERT model on this task by following the training and validation steps prespecified by TAC. RESULTS: This initial attempt showed the model achieved an F-score of 0.82 in identifying sentences as containing PK DDI and an F-score of 0.97 in identifying object versus precipitant drugs in those sentences. DISCUSSION AND CONCLUSION: Despite a growing list of NLP methods for DDI extraction, most of them use a common set of corpora to perform general purpose tasks (e.g., classifying a sentence into one of several fixed DDI categories). There is a lack of coordination between the drug development and biomedical informatics method development community to develop corpora and methods to perform specific tasks (e.g., extract clinical exposure changes due to PK DDI). We hope that our effort can encourage such a coordination so that more "fit for purpose" NLP methods could be developed and used to facilitate the drug development process.

Calibration and Validation of a Mechanistic COVID-19 Model for Translational Quantitative Systems Pharmacology - A Proof-of-Concept Model Development for Remdesivir.

With the ongoing global pandemic of coronavirus disease 2019 (COVID-19), there is an urgent need to accelerate the traditional drug development process. Many studies identified potential COVID-19 therapies based on promising nonclinical data. However, the poor translatability from nonclinical to clinical settings has led to failures of many of these drug candidates in the clinical phase. In this study, we propose a mechanism-based, quantitative framework to translate nonclinical findings to clinical outcome. Adopting a modularized approach, this framework includes an in silico disease model for COVID-19 (virus infection and human immune responses) and a pharmacological component for COVID-19 therapies. The disease model was able to reproduce important longitudinal clinical data for patients with mild and severe COVID-19, including viral titer, key immunological cytokines, antibody responses, and time courses of lymphopenia. Using remdesivir as a proof-of-concept example of model development for the pharmacological component, we developed a pharmacological model that describes the conversion of intravenously administered remdesivir as a prodrug to its active metabolite nucleoside triphosphate through intracellular metabolism and connected it to the COVID-19 disease model. After being calibrated with the placebo arm data, our model was independently and quantitatively able to predict the primary endpoint (time to recovery) of the remdesivir clinical study, Adaptive Covid-19 Clinical Trial (ACTT). Our work demonstrates the possibility of quantitatively predicting clinical outcome based on nonclinical data and mechanistic understanding of the disease and provides a modularized framework to aid in candidate drug selection and clinical trial design for COVID-19 therapeutics.

Development of a Translational Model to Assess the Impact of Opioid Overdose and Naloxone Dosing on Respiratory Depression and Cardiac Arrest.

In response to a surge of deaths from synthetic opioid overdoses, there have been increased efforts to distribute naloxone products in community settings. Prior research has assessed the effectiveness of naloxone in the hospital setting; however, it is challenging to assess naloxone dosing regimens in the community/first-responder setting, including reversal of respiratory depression effects of fentanyl and its derivatives (fentanyls). Here, we describe the development and validation of a mechanistic model that combines opioid mu receptor binding kinetics, opioid agonist and antagonist pharmacokinetics, and human respiratory and circulatory physiology, to evaluate naloxone dosing to reverse respiratory depression. Validation supports our model, which can quantitatively predict displacement of opioids by naloxone from opioid mu receptors in vitro, hypoxia-induced cardiac arrest in vivo, and opioid-induced respiratory depression in humans from different fentanyls. After validation, overdose simulations were performed with fentanyl and carfentanil followed by administration of different intramuscular naloxone products. Carfentanil induced more cardiac arrest events and was more difficult to reverse than fentanyl. Opioid receptor binding data indicated that carfentanil has substantially slower dissociation kinetics from the opioid receptor compared with nine other fentanyls tested, which likely contributes to the difficulty in reversing carfentanil. Administration of the same dose of naloxone intramuscularly from two different naloxone products with different formulations resulted in differences in the number of virtual patients experiencing cardiac arrest. This work provides a robust framework to evaluate dosing regimens of opioid receptor antagonists to reverse opioid-induced respiratory depression, including those caused by newly emerging synthetic opioids.

Noise effect on the temporal patterns of neural synchrony.

Neural synchrony in the brain is often present in an intermittent fashion, i.e., there are intervals of synchronized activity interspersed with intervals of desynchronized activity. A series of experimental studies showed that this kind of temporal patterning of neural synchronization may be very specific and may be correlated with behaviour (even if the average synchrony strength is not changed). Prior studies showed that a network with many short desynchronized intervals may be functionally different from a network with few long desynchronized intervals as it may be more sensitive to synchronizing input signals. In this study, we investigated the effect of channel noise on the temporal patterns of neural synchronization. We employed a small network of conductance-based model neurons that were mutually connected via excitatory synapses. The resulting dynamics of the network was studied using the same time-series analysis methods as used in prior experimental and computational studies. While it is well known that synchrony strength generally degrades with noise, we found that noise also affects the temporal patterning of synchrony. Noise, at a sufficient intensity (yet too weak to substantially affect synchrony strength), promotes dynamics with predominantly short (although potentially very numerous) desynchronizations. Thus, channel noise may be one of the mechanisms contributing to the short desynchronization dynamics observed in multiple experimental studies.

Spike-Timing Dependent Plasticity Effect on the Temporal Patterning of Neural Synchronization.

Neural synchrony in the brain at rest is usually variable and intermittent, thus intervals of predominantly synchronized activity are interrupted by intervals of desynchronized activity. Prior studies suggested that this temporal structure of the weakly synchronous activity might be functionally significant: many short desynchronizations may be functionally different from few long desynchronizations even if the average synchrony level is the same. In this study, we used computational neuroscience methods to investigate the effects of spike-timing dependent plasticity (STDP) on the temporal patterns of synchronization in a simple model. We employed a small network of conductance-based model neurons that were connected via excitatory plastic synapses. The dynamics of this network was subjected to the time-series analysis methods used in prior experimental studies. We found that STDP could alter the synchronized dynamics in the network in several ways, depending on the time scale that plasticity acts on. However, in general, the action of STDP in the simple network considered here is to promote dynamics with short desynchronizations (i.e., dynamics reminiscent of that observed in experimental studies). Complex interplay of the cellular and synaptic dynamics may lead to the activity-dependent adjustment of synaptic strength in such a way as to facilitate experimentally observed short desynchronizations in the intermittently synchronized neural activity.