A novel anti-inflammatory hybrid 3-ibuprofenyl-copalic acid (3-IbuCA) was synthesized from 3-hydroxy-copalic acid isolated from Amazonian copaiba oil (Copaifera multijuga Hayne), and the anti-inflammatory ibuprofen. After full characterization, several assays to verify its anti-inflammatory effects were performed in vitro, in vivo and in silico (molecular docking). Induced fit docking was performed to observe the interactions with the enzymes cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2). In vitro tests of cytotoxicity and tumor necrosis factor (TNF)-α inhibition, and in vivo tests of pleurisy, protein expression and gastrocytotoxicity were performed. Molecular docking studies with COX-1 and 2 showed binding free energies (ΔG) of −2.2 and −7.8 kcal mol −1 , respectively, while for mofezolac and indomethacin, the binding free energies ΔG presented values of −8.5 and −10.1 kcal mol −1 , which makes 3-IbuCA selective for COX-2 inhibition. This hybrid showed no toxicity against human macrophage at concentrations up to 2 µM, and inhibited TNF-α production in lipopolysaccharide (LPS)-stimulated macrophages. In the pleurisy assays, 3-IbuCA reduced the total leukocytes and mononuclear cells, which was followed by reduction of p-IKBα (phosphorylated nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha) protein expression. Compared with ibuprofen alone, the hybrid caused less gastric damage. Thus, the docking, together with in vitro and in vivo studies suggest that this novel hybrid has potential as a new anti-inflammatory agent.

No abstract

A typical information extraction pipeline consists of tokenor span-level classification models coupled with a series of pre-and postprocessing scripts. In a production pipeline, requirements often change, with classes being added and removed, which leads to nontrivial modifications to the source code and the possible introduction of bugs. In this work, we evaluate sequence-to-sequence models as an alternative to token-level classification methods for information extraction of legal and registration documents. We finetune models that jointly extract the information and generate the output already in a structured format. Post-processing steps are learned during training, thus eliminating the need for rule-based methods and simplifying the pipeline. Furthermore, we propose a novel method to align the output with the input text, thus facilitating system inspection and auditing. Our experiments on four real-world datasets show that the proposed method is an alternative to classical pipelines.