Machine Learning Predicts Drug Metabolism and Bioaccumulation by Intestinal Microbiota

A study of machine learning to predict NRDS severity based on lung ultrasound score and clinical indicators

Objective

To develop predictive models for neonatal respiratory distress syndrome (NRDS) using machine learning algorithms to improve the accuracy of severity predictions.

Methods

This double-blind cohort study included 230 neonates admitted to the neonatal intensive care unit (NICU) of Yantaishan Hospital between December 2020 and June 2023. Of these, 119 neonates were diagnosed with NRDS and placed in the NRDS group, while 111 neonates with other conditions formed the non-NRDS (N-NRDS) group. All neonates underwent lung ultrasound and various clinical assessments, with data collected on the oxygenation index (OI), sequential organ failure assessment (SOFA), respiratory index (RI), and lung ultrasound score (LUS). An independent sample test was used to compare the groups' LUS, OI, RI, SOFA scores, and clinical data. Use Least Absolute Shrinkage and Selection Operator (LASSO) regression to identify predictor variables, and construct a model for predicting NRDS severity using logistic regression (LR), random forest (RF), artificial neural network (NN), and support vector machine (SVM) algorithms. The importance of predictive variables and performance metrics was evaluated for each model.

Results

The NRDS group showed significantly higher LUS, SOFA, and RI scores and lower OI values than the N-NRDS group (p < 0.01). LUS, SOFA, and RI scores were significantly higher in the severe NRDS group compared to the mild and moderate groups, while OI was markedly lower (p < 0.01). LUS, OI, RI, and SOFA scores were the most impactful variables for the predictive efficacy of the models. The RF model performed best of the four models, with an AUC of 0.894, accuracy of 0.808, and sensitivity of 0.706. In contrast, the LR, NN, and SVM models have lower AUC values than the RF model with 0.841, 0.828, and 0.726, respectively.

Conclusion

Four predictive models based on machine learning can accurately assess the severity of NRDS. Among them, the RF model exhibits the best predictive performance, offering more effective support for the treatment and care of neonates.

Exploring the interactions of JAK inhibitor and S1P receptor modulator drugs with the human gut microbiome: Implications for colonic drug delivery and inflammatory bowel disease

The gut microbiota is a complex ecosystem, home to hundreds of bacterial species and a vast repository of enzymes capable of metabolising a wide range of pharmaceuticals. Several drugs have been shown to affect negatively the composition and function of the gut microbial ecosystem. Janus Kinase (JAK) inhibitors and Sphingosine-1-phosphate (S1P) receptor modulators are drugs recently approved for inflammatory bowel disease through an immediate release formulation and would potentially benefit from colonic targeted delivery to enhance the local drug concentration at the diseased site. However, their impact on the human gut microbiota and susceptibility to bacterial metabolism remain unexplored. With the use of calorimetric, optical density measurements, and metagenomics next-generation sequencing, we show that JAK inhibitors (tofacitinib citrate, baricitinib, filgotinib) have a minor impact on the composition of the human gut microbiota, while ozanimod exerts a significant antimicrobial effect, leading to a prevalence of the Enterococcus genus and a markedly different metabolic landscape when compared to the untreated microbiota. Moreover, ozanimod, unlike the JAK inhibitors, is the only drug subject to enzymatic degradation by the human gut microbiota sourced from six healthy donors. Overall, given the crucial role of the gut microbiome in health, screening assays to investigate the interaction of drugs with the microbiota should be encouraged for the pharmaceutical industry as a standard in the drug discovery and development process.

Gut Microbiome Integration in Drug Discovery and Development of Small Molecules

Human microbiomes, particularly in the gut, could have a major impact on the efficacy and toxicity of drugs. However, gut microbial metabolism is often neglected in the drug discovery and development process. Medicen, a Paris-based human health innovation cluster, has gathered more than 30 international leading experts from pharma, academia, biotech, clinical research organizations, and regulatory science to develop proposals to facilitate the integration of microbiome science into drug discovery and development. Seven subteams were formed to cover the complementary expertise areas of 1) pharma experience and case studies, 2) in silico microbiome-drug interaction, 3) in vitro microbial stability screening, 4) gut fermentation models, 5) animal models, 6) microbiome integration in clinical and regulatory aspects, and 7) microbiome ecosystems and models. Each expert team produced a state-of-the-art report of their respective field highlighting existing microbiome-related tools at every stage of drug discovery and development. The most critical limitations are the growing, but still limited, drug-microbiome interaction data to produce predictive models and the lack of agreed-upon standards despite recent progress. In this paper we will report on and share proposals covering 1) how microbiome tools can support moving a compound from drug discovery to clinical proof-of-concept studies and alert early on potential undesired properties stemming from microbiome-induced drug metabolism and 2) how microbiome data can be generated and integrated in pharmacokinetic models that are predictive of the human situation. Examples of drugs metabolized by the microbiome will be discussed in detail to support recommendations from the working group. SIGNIFICANCE STATEMENT: Gut microbial metabolism is often neglected in the drug discovery and development process despite growing evidence of drugs' efficacy and safety impacted by their interaction with the microbiome. This paper will detail existing microbiome-related tools covering every stage of drug discovery and development, current progress, and limitations, as well as recommendations to integrate them into the drug discovery and development process.

Integrating the gut microbiome and pharmacology

The gut microbiome harbors trillions of organisms that contribute to human health and disease. These bacteria can also affect the properties of medical drugs used to treat these diseases, and drugs, in turn, can reshape the microbiome. Research addressing interdependent microbiome-host-drug interactions thus has broad impact. In this Review, we discuss these interactions from the perspective of drug bioavailability, absorption, metabolism, excretion, toxicity, and drug-mediated microbiome modulation. We survey approaches that aim to uncover the mechanisms underlying these effects and opportunities to translate this knowledge into new strategies to improve the development, administration, and monitoring of medical drugs.

Microbial metabolism marvels: a comprehensive review of microbial drug transformation capabilities

This comprehensive review elucidates the pivotal role of microbes in drug metabolism, synthesizing insights from an exhaustive analysis of over two hundred papers. Employing a structural classification system grounded in drug atom involvement, the review categorizes the microbiome-mediated drug-metabolizing capabilities of over 80 drugs. Additionally, it compiles pharmacodynamic and enzymatic details related to these reactions, striving to include information on encoding genes and specific involved microorganisms. Bridging biochemistry, pharmacology, genetics, and microbiology, this review not only serves to consolidate diverse research fields but also highlights the potential impact of microbial drug metabolism on future drug design and in silico studies. With a visionary outlook, it also lays the groundwork for personalized medicine interventions, emphasizing the importance of interdisciplinary collaboration for advancing drug development and enhancing therapeutic strategies.

Application of artificial intelligence approaches to predict the metabolism of xenobiotic molecules by human gut microbiome

A highly complex, diverse, and dense community of more than 1,000 different gut bacterial species constitutes the human gut microbiome that harbours vast metabolic capabilities encoded by more than 300,000 bacterial enzymes to metabolise complex polysaccharides, orally administered drugs/xenobiotics, nutraceuticals, or prebiotics. One of the implications of gut microbiome mediated biotransformation is the metabolism of xenobiotics such as medicinal drugs, which lead to alteration in their pharmacological properties, loss of drug efficacy, bioavailability, may generate toxic byproducts and sometimes also help in conversion of a prodrug into its active metabolite. Given the diversity of gut microbiome and the complex interplay of the metabolic enzymes and their diverse substrates, the traditional experimental methods have limited ability to identify the gut bacterial species involved in such biotransformation, and to study the bacterial species-metabolite interactions in gut. In this scenario, computational approaches such as machine learning-based tools presents unprecedented opportunities and ability to predict the gut bacteria and enzymes that can potentially metabolise a candidate drug. Here, we have reviewed the need to identify the gut microbiome-based metabolism of xenobiotics and have provided comprehensive information on the available methods, tools, and databases to address it along with their scope and limitations.

Machine learning model of imipenem-resistant Klebsiella pneumoniae based on MALDI-TOF-MS platform: An observational study

Background and Aim

Machine learning is an important branch and supporting technology of artificial intelligence, we established four machine learning model for the drug sensitivity of Klebsiella pneumoniae to imipenem based on matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) and compared their diagnostic effect.

Methods

The data of MALDI-TOF-MS and imipenem sensitivity of 174 cases of K. pneumoniae isolated from clinical specimens in the laboratory of microbiology department of Tianjin Haihe Hospital from 2019 January to 2020 December were collected. The mass spectrometry and imipenem sensitivity of 70 cases of imipenem-sensitive and 70 resistant cases were randomly selected to establish the training set model, 17 cases of sensitive and 17 cases of resistant cases were randomly selected to establish the test set model. Mass spectral peak data were subjected to orthogonal partial least squares discriminant analysis (OPLS-DA), the training set data model was established by machine learning least absolute shrinkage and selection operator (LASSO) algorithm, logistic regression (LR) algorithm, support vector machines (SVM) algorithm, neural network (NN) algorithm, the area under the curve (AUC) and confusion matrix of training set and test set model were calculated and selected by Grid search and 3-fold Cross-validation respectively, the accuracy of the prediction model was verified by test set confusion matrix.

Results

The R²Y and Q² of OPLS-DA were 0.546 and 0.0178. The AUC of the best training set and test set models were 0.9726 and 0.9100, 1.0000 and 0.8581, 0.8462 and 0.6263, 1.0000 and 0.7180 evaluated by LASSO, LR, SVM and NN model respectively. The accuracy of the LASSO, LR, SVM and NN model were 87%, 79%, 62%, and 68% in test set, respectively.

Conclusion

The LASSO prediction model of K. pneumoniae sensitivity to imipenem established in this study has a high accuracy rate and has potential clinical decision support ability.

The Interplay between Gut Microbiota and Oral Medications and Its Impact on Advancing Precision Medicine

Trillions of diverse microbes reside in the gut and are deeply interwoven with the human physiological process, from food digestion, immune system maturation, and fighting invading pathogens, to drug metabolism. Microbial drug metabolism has a profound impact on drug absorption, bioavailability, stability, efficacy, and toxicity. However, our knowledge of specific gut microbial strains, and their genes that encode enzymes involved in the metabolism, is limited. The microbiome encodes over 3 million unique genes contributing to a huge enzymatic capacity, vastly expanding the traditional drug metabolic reactions that occur in the liver, manipulating their pharmacological effect, and, ultimately, leading to variation in drug response. For example, the microbial deactivation of anticancer drugs such as gemcitabine can lead to resistance to chemotherapeutics or the crucial role of microbes in modulating the efficacy of the anticancer drug, cyclophosphamide. On the other hand, recent findings show that many drugs can shape the composition, function, and gene expression of the gut microbial community, making it harder to predict the outcome of drug-microbiota interactions. In this review, we discuss the recent understanding of the multidirectional interaction between the host, oral medications, and gut microbiota, using traditional and machine-learning approaches. We analyze gaps, challenges, and future promises of personalized medicine that consider gut microbes as a crucial player in drug metabolism. This consideration will enable the development of personalized therapeutic regimes with an improved outcome, ultimately leading to precision medicine.

Advancing oral delivery of biologics: Machine learning predicts peptide stability in the gastrointestinal tract

The oral delivery of peptide therapeutics could facilitate precision treatment of numerous gastrointestinal (GI) and systemic diseases with simple administration for patients. However, the vast majority of licensed peptide drugs are currently administered parenterally due to prohibitive peptide instability in the GI tract. As such, the development of GI-stable peptides is receiving considerable investment. This study provides researchers with the first tool to predict the GI stability of peptide therapeutics based solely on the amino acid sequence. Both unsupervised and supervised machine learning techniques were trained on literature-extracted data describing peptide stability in simulated gastric and small intestinal fluid (SGF and SIF). Based on 109 peptide incubations, classification models for SGF and SIF were developed. The best models utilized k-Nearest Neighbor (for SGF) and XGBoost (for SIF) algorithms, with accuracies of 75.1% (SGF) and 69.3% (SIF), and f1 scores of 84.5% (SGF) and 73.4% (SIF) under 5-fold cross-validation. Feature importance analysis demonstrated that peptides' lipophilicity, rigidity, and size were key determinants of stability. These models are now available to those working on the development of oral peptide therapeutics.

Machine learning using multi-modal data predicts the production of selective laser sintered 3D printed drug products

Three-dimensional (3D) printing is drastically redefining medicine production, offering digital precision and personalized design opportunities. One emerging 3D printing technology is selective laser sintering (SLS), which is garnering attention for its high precision, and compatibility with a wide range of pharmaceutical materials, including low-solubility compounds. However, the full potential of SLS for medicines is yet to be realized, requiring expertise and considerable time-consuming and resource-intensive trial-and-error research. Machine learning (ML), a subset of artificial intelligence, is an in silico tool that is accomplishing remarkable breakthroughs in several sectors for its ability to make highly accurate predictions. Therefore, the present study harnessed ML to predict the printability of SLS formulations. Using a dataset of 170 formulations from 78 materials, ML models were developed from inputs that included the formulation composition and characterization data retrieved from Fourier-transformed infrared spectroscopy (FT-IR), X-ray powder diffraction (XRPD) and differential scanning calorimetry (DSC). Multiple ML models were explored, including supervised and unsupervised approaches. The results revealed that ML can achieve high accuracies, by using the formulation composition leading to a maximum F1 score of 81.9%. Using the FT-IR, XRPD and DSC data as inputs resulted in an F1 score of 84.2%, 81.3%, and 80.1%, respectively. A subsequent ML pipeline was built to combine the predictions from FT-IR, XRPD and DSC into one consensus model, where the F1 score was found to further increase to 88.9%. Therefore, it was determined for the first time that ML predictions of 3D printability benefit from multi-modal data, combining numeric, spectral, thermogram and diffraction data. The study lays the groundwork for leveraging existing characterization data for developing high-performing computational models to accelerate formulation development.

Colonic drug delivery: Formulating the next generation of colon-targeted therapeutics

Colonic drug delivery can facilitate access to unique therapeutic targets and has the potential to enhance drug bioavailability whilst reducing off-target effects. Delivering drugs to the colon requires considered formulation development, as both oral and rectal dosage forms can encounter challenges if the colon's distinct physiological environment is not appreciated. As the therapeutic opportunities surrounding colonic drug delivery multiply, the success of novel pharmaceuticals lies in their design. This review provides a modern insight into the key parameters determining the effective design and development of colon-targeted medicines. Influential physiological features governing the release, dissolution, stability, and absorption of drugs in the colon are first discussed, followed by an overview of the most reliable colon-targeted formulation strategies. Finally, the most appropriate in vitro, in vivo, and in silico preclinical investigations are presented, with the goal of inspiring strategic development of new colon-targeted therapeutics.

Ileocolonic-Targeted JAK Inhibitor: A Safer and More Effective Treatment for Inflammatory Bowel Disease

Janus kinase (JAK) inhibitors, such as tofacitinib (Xeljanz) and filgotinib (Jyseleca), have been approved for treatment of ulcerative colitis with several other JAK inhibitors in late-stage clinical trials for inflammatory bowel disease (IBD). Despite their impressive efficacy, the risk of adverse effects accompanying the use of JAK inhibitors has brought the entire class under scrutiny, leading to them receiving an FDA black box warning. In this study we investigated whether ileocolonic-targeted delivery of a pan-JAK inhibitor, tofacitinib, can lead to increased tissue exposure and reduced systemic exposure compared to untargeted formulations. The stability of tofacitinib in the presence of rat colonic microbiota was first confirmed. Next, in vivo computed tomography imaging was performed in rats to determine the transit time and disintegration site of ileocolonic-targeted capsules compared to gastric release capsules. Pharmacokinetic studies demonstrated that systemic drug exposure was significantly decreased, and colonic tissue exposure increased at 10 mg/kg tofacitinib dosed in ileocolonic-targeted capsules compared to gastric release capsules and an oral solution. Finally, in a rat model of LPS-induced colonic inflammation, targeted tofacitinib capsules significantly reduced concentrations of proinflammatory interleukin 6 in colonic tissue compared to a vehicle-treated control (p = 0.0408), unlike gastric release tofacitinib capsules and orally administered dexamethasone. Overall, these results support further development of ileocolonic-targeted tofacitinib, and potentially other specific JAK inhibitors in pre-clinical and clinical development, for the treatment of IBD.

Gut Microbiome: The Interplay of an "Invisible Organ" with Herbal Medicine and Its Derived Compounds in Chronic Metabolic Disorders

Resembling a concealed "organ" in a holobiont, trillions of gut microbes play complex roles in the maintenance of homeostasis, including participating in drug metabolism. The conventional opinion is that most of any drug is metabolized by the host and that individual differences are principally due to host genetic factors. However, current evidence indicates that only about 60% of the individual differences in drug metabolism are attributable to host genetics. Although most common chemical drugs regulate the gut microbiota, the gut microbiota is also known to be involved in drug metabolism, like the host. Interestingly, many traditional herbal medicines and derived compounds are biotransformed by gut microbiota, manipulating the compounds' effects. Accordingly, the gut microbiota and its specified metabolic pathways can be deemed a promising target for promoting drug efficacy and safety. However, the evidence regarding causality and the corresponding mechanisms concerning gut microbiota and drug metabolism remains insufficient, especially regarding drugs used to treat metabolic disorders. Therefore, the present review aims to comprehensively summarize the bidirectional roles of gut microbiota in the effects of herbal medicine in metabolic diseases to provide vital clues for guiding the clinical application of precision medicine and personalized drug development.

Microbiome as an immune regulator in health, disease, and therapeutics

New discoveries in drugs and drug delivery systems are focused on identifying and delivering a pharmacologically effective agent, potentially targeting a specific molecular component. However, current drug discovery and therapeutic delivery approaches do not necessarily exploit the complex regulatory network of an indispensable microbiota that has been engineered through evolutionary processes in humans or has been altered by environmental exposure or diseases. The human microbiome, in all its complexity, plays an integral role in the maintenance of host functions such as metabolism and immunity. However, dysregulation in this intricate ecosystem has been linked with a variety of diseases, ranging from inflammatory bowel disease to cancer. Therapeutics and bacteria have an undeniable effect on each other and understanding the interplay between microbes and drugs could lead to new therapies, or to changes in how existing drugs are delivered. In addition, targeting the human microbiome using engineered therapeutics has the potential to address global health challenges. Here, we present the challenges and cutting-edge developments in microbiome-immune cell interactions and outline novel targeting strategies to advance drug discovery and therapeutics, which are defining a new era of personalized and precision medicine.

Design and manufacture of a lyophilised faecal microbiota capsule formulation to GMP standards

Faecal microbiota transplant (FMT) is an established and effective treatment for recurrent Clostridioides difficile infection (CDI) and has many other potential clinical applications. However, preparation and quality of FMT is poorly standardised and clinical studies are hampered by a lack of well-defined FMT formulations that meet regulatory standards for medicines. As an alternative to FMT suspensions for administration by nasojejunal tube or colonoscopy, which is invasive and disliked by many patients, this study aimed to develop a well-controlled, standardised method for manufacture of lyophilised FMT capsules and to provide stability data allowing storage for extended time periods. Faecal donations were collected from healthy, pre-screened individuals, homogenised, filtered and centrifuged to remove dietary matter. The suspension was centrifuged to pellet bacteria, which were resuspended with trehalose and lyophilised to produce a powder which was filled into 5 enteric-coated capsules (size 0). Live-dead bacterial cell quantitative PCR assay showed <10 fold viable bacterial load reduction through the manufacturing process. No significant loss of viable bacterial load was observed after storage at -80 °C for 36 weeks (p = 0.24, n = 5). Initial clinical experience demonstrated that the capsules produced clinical cure in patients with CDI with no adverse events reported (n = 7). We provide the first report of a detailed manufacturing protocol and specification for an encapsulated lyophilised formulation of FMT. As clinical trials into intestinal microbiota interventions proceed, it is important to use a well-controlled investigational medicinal product in the studies so that any beneficial results can be replicated in clinical practice.

Usefulness of Optimized Human Fecal Material in Simulating the Bacterial Degradation of Sulindac and Sulfinpyrazone in the Lower Intestine

The first aim of this study was to evaluate the usefulness of optimized human fecal material in simulating sulforeductase activity in the lower intestine by assessing bacterial degradation of sulindac and sulfinpyrazone, two sulforeductase substrates. The second aim was to evaluate the usefulness of drug degradation half-life generated in simulated colonic bacteria (SCoB) in informing PBPK models. Degradation experiments of sulfinpyrazone and of sulindac in SCoB were performed under anaerobic conditions using recently described methods. For sulfinpyrazone, the abundance of clinical data allowed for construction of a physiologically based pharmacokinetic (PBPK) model and evaluation of luminal degradation clearance determined from SCoB data. For sulindac, the availability of sulindac sulfide and sulindac sulfone standards allowed for evaluating the formation of the main metabolite, sulindac sulfide, during the experiments in SCoB. Both model compounds degraded substantially in SCoB. The PBPK model was able to adequately capture exposure of sulfinpyrazone and its sulfide metabolite in healthy subjects, in ileostomy and/or colectomy subjects, and in healthy subjects pretreated with metoclopramide by implementing degradation half-lives in SCoB to calculate intrinsic colon clearance. Degradation rates of sulindac and formation rates of sulindac sulfide in SCoB were almost identical, in line with in vivo data suggesting the sulindac sulfide is the primary metabolite in the lower intestine. Experiments in SCoB were useful in simulating sulforeductase related bacterial degradation activity in the lower intestine. Degradation half-life calculated from experiments in SCoB is proven useful for informing a predictive PBPK model for sulfinpyrazone.

Machine Learning in Drug Metabolism Study

Metabolic reactions in the body transform the administered drug into metabolites. These metabolites exhibit diverse biological activities. Drug metabolism is the major underlying cause of drug overdose-related toxicity, adversative drug effects and the drug's reduced efficacy. Though metabolic reactions deactivate a drug, drug metabolites are often considered pivotal agents for off-target effects or toxicity. On the other side, in combination drug therapy, one drug may influence another drug's metabolism and clearance and is thus considered one of the primary causes of drug-drug interactions. Today with the advancement of machine learning, the metabolic fate of a drug candidate can be comprehensively studied throughout the drug development procedure. Naïve Bayes, Logistic Regression, k-Nearest Neighbours, Decision Trees, different Boosting and Ensemble methods, Support Vector Machines and Artificial Neural Network boosted Deep Learning are some machine learning algorithms which are being extensively used in such studies. Such tools are covering several attributes of drug metabolism, with an emphasis on the prediction of drug-drug interactions, drug-target-interactions, clinical drug responses, metabolite predictions, sites of metabolism, etc. These reports are crucial for evaluating metabolic stability and predicting prospective drug-drug interactions, and can help pharmaceutical companies accelerate the drug development process in a less resourcedemanding manner than what in vitro studies offer. It could also help medical practitioners to use combinatorial drug therapy in a more resourceful manner. Also, with the help of the enormous growth of deep learning, traditional fields of computational drug development like molecular interaction fields, molecular docking, quantitative structure-toactivity relationship (QSAR) studies and quantum mechanical simulations are producing results which were unimaginable couple of years back. This review provides a glimpse of a few contextually relevant machine learning algorithms and then focuses on their outcomes in different studies.