Impact of Peptide Structure on Colonic Stability and Tissue Permeability.

Most marketed peptide drugs are administered parenterally due to their inherent gastrointestinal (GI) instability and poor permeability across the GI epithelium. Several molecular design techniques, such as cyclisation and D-amino acid (D-AA) substitution, have been proposed to improve oral peptide drug bioavailability. However, very few of these techniques have been translated to the clinic. In addition, little is known about how synthetic peptide design may improve stability and permeability in the colon, a key site for the treatment of inflammatory bowel disease and colorectal cancer. In this study, we investigated the impact of various cyclisation modifications and D-AA substitutions on the enzymatic stability and colonic tissue permeability of native oxytocin and 11 oxytocin-based peptides. Results showed that the disulfide bond cyclisation present in native oxytocin provided an improved stability in a human colon model compared to a linear oxytocin derivative. Chloroacetyl cyclisation increased native oxytocin stability in the colonic model at 1.5 h by 30.0%, whereas thioether and N-terminal acetylated cyclisations offered no additional protection at 1.5 h. The site and number of D-AA substitutions were found to be critical for stability, with three D-AAs at Tyr, Ile and Leu, improving native oxytocin stability at 1.5 h in both linear and cyclic structures by 58.2% and 79.1%, respectively. Substitution of three D-AAs into native cyclic oxytocin significantly increased peptide permeability across rat colonic tissue; this may be because D-AA substitution favourably altered the peptide's secondary structure. This study is the first to show how the strategic design of peptide therapeutics could enable their delivery to the colon via the oral route.

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.

Reliably Filter Drug-Induced Liver Injury Literature With Natural Language Processing and Conformal Prediction.

Drug-induced liver injury describes the adverse effects of drugs that damage the liver. Life-threatening results were also reported in severe cases. Therefore, liver toxicity is an important assessment for new drug candidates. These reports are documented in research papers that contain preliminary in vitro and in vivo experiments. Conventionally, data extraction from publications relies on resource-demanding manual labeling, which restricts the efficiency of the information extraction. The development of natural language processing techniques enables the automatic processing of biomedical texts. Herein, based on around 28,000 papers (titles and abstracts) provided by the Critical Assessment of Massive Data Analysis challenge, this study benchmarked model performances on filtering liver-damage-related literature. Among five text embedding techniques, the model using term frequency-inverse document frequency (TF-IDF) and logistic regression outperformed others with an accuracy of 0.957 on the validation set. Furthermore, an ensemble model with similar overall performances was developed with a logistic regression model on the predicted probability given by separate models with different vectorization techniques. The ensemble model achieved a high accuracy of 0.954 and an F1 score of 0.955 in the hold-out validation data in the challenge. Moreover, important words in positive/negative predictions were identified via model interpretation. The prediction reliability was quantified with conformal prediction, which provides users with a control over the prediction uncertainty. Overall, the ensemble model and TF-IDF model reached satisfactory classification results, which can be used by researchers to rapidly filter literature that describes events related to liver injury induced by medications.

Machine learning to empower electrohydrodynamic processing.

Electrohydrodynamic (EHD) processes are promising healthcare fabrication technologies, as evidenced by the number of commercialised and food-and-drug administration (FDA)-approved products produced by these processes. Their ability to produce both rapidly and precisely nano-sized products provides them with a unique set of qualities that cannot be matched by other fabrication technologies. Consequently, this has stimulated the development of EHD processing to tackle other healthcare challenges. However, as with most technologies, time and resources will be needed to realise fully the potential EHD processes can offer. To address this bottleneck, researchers are adopting machine learning (ML), a subset of artificial intelligence, into their workflow. ML has already made ground-breaking advancements in the healthcare sector, and it is anticipated to do the same in the materials domain. Presently, the application of ML in fabrication technologies lags behind other sectors. To that end, this review showcases the progress made by ML for EHD workflows, demonstrating how the latter can benefit greatly from the former. In addition, we provide an introduction to the ML pipeline, to help encourage the use of ML for other EHD researchers. As discussed, the merger of ML with EHD has the potential to expedite novel discoveries and to automate the EHD workflow.