Automatic text classification of drug-induced liver injury using document-term matrix and XGBoost.

Introduction: Regulatory agencies generate a vast amount of textual data in the review process. For example, drug labeling serves as a valuable resource for regulatory agencies, such as U.S. Food and Drug Administration (FDA) and Europe Medical Agency (EMA), to communicate drug safety and effectiveness information to healthcare professionals and patients. Drug labeling also serves as a resource for pharmacovigilance and drug safety research. Automated text classification would significantly improve the analysis of drug labeling documents and conserve reviewer resources. Methods: We utilized artificial intelligence in this study to classify drug-induced liver injury (DILI)-related content from drug labeling documents based on FDA's DILIrank dataset. We employed text mining and XGBoost models and utilized the Preferred Terms of Medical queries for adverse event standards to simplify the elimination of common words and phrases while retaining medical standard terms for FDA and EMA drug label datasets. Then, we constructed a document term matrix using weights computed by Term Frequency-Inverse Document Frequency (TF-IDF) for each included word/term/token. Results: The automatic text classification model exhibited robust performance in predicting DILI, achieving cross-validation AUC scores exceeding 0.90 for both drug labels from FDA and EMA and literature abstracts from the Critical Assessment of Massive Data Analysis (CAMDA). Discussion: Moreover, the text mining and XGBoost functions demonstrated in this study can be applied to other text processing and classification tasks.

The Influence of Tissue Plasminogen Activator I/D Polymorphism on the tPA Response to Exercise.

The purpose was to determine if the Alu-insertion (I)/deletion (D) polymorphism of the tissue plasminogen activator (tPA) gene influences the tPA response to maximal exercise. Fifty male subjects (age = 23.6 ± 4.7 yrs) completed a maximal treadmill exercise test. Blood samples were drawn before and immediately after exercise for determination of plasma tPA antigen and activity. Isolated DNA was amplified via polymerase chain reaction, electrophoresed, and visually amplified to determine tPA genotype. Subjects were classified as possessing the D allele (D) (n = 28) or being homozygous for the I allele (I) (n = 22). Differences in tPA antigen and activity were assessed using a two-factor (genotype and time) repeated measures analysis of variance. There were significant main effects for time for tPA antigen and activity (p < 0.05), but no main effect for genotype. Furthermore, there was no genotype x time interaction due to a similar increase in tPA antigen in the D group (pre-exercise = 5.83 + 2.39 ng/ml, post-exercise = 21.88 + 7.38 ng/ml) and the I group (pre-exercise = 5.61 + 2.82 ng/ml, post-exercise = 19.05 + 7.67 ng/ml) and a similar increase in tPA activity in the D group (pre-exercise = 0.39 ± 0.19 IU/ml, post-exercise = 9.73 ± 4.22 IU/ml) and I group (pre-exercise = 0.45 ± 0.29 IU/ml, post-exercise = 9.76 ± 5.50 IU/ml). The I/D polymorphism of the tPA gene does not influence the tPA antigen nor tPA activity responses to maximal exercise in healthy, young, sedentary males.

Extracellular matrix fibronectin increases prostaglandin E2 receptor subtype EP4 in lung carcinoma cells through multiple signaling pathways: the role of AP-2.

We have previously demonstrated that fibronectin (Fn) stimulates the proliferation of non-small cell lung carcinoma (NSCLC) cell growth through the induction of cyclooxygenase-2 (COX-2) and prostaglandin E2 secretion. Here, we demonstrate that NSCLC cells express mRNA and protein for the prostaglandin E2 receptor EP4 and that Fn enhances its stimulatory effect by inducing the expression of EP4, but not of EP1, EP2, and EP3 receptor subtypes. The effect of Fn on EP4 was inhibited by an antibody against alpha5beta1 integrin and by inhibitors of phosphoinositide 3-kinase (wortmannin) and extracellular signal-regulated kinase (PD98095), but not by inhibitors of protein kinase C (calphostin C), of protein kinase A (H-89), or of mammalian target of rapamycin (rapamycin). A COX-2 small interfering RNA was also inhibitory. Fn significantly increased AP-2 binding activity in the promoter of the EP4 gene, and AP-2 antisense oligonucleotides blocked Fn-induced EP4 expression. Using full-length and mutated EP4 promoter constructs, we found that Fn stimulation of EP4 gene expression was inhibited when one AP-2 site (-1000 bp) was mutated. Fn induced nuclear AP-2alpha protein expression through multiple signaling pathways. Our results indicate that Fn-induced NSCLC cell proliferation is mediated through EP4. Furthermore, they show that Fn induces EP4 expression through the activation of alpha5beta1-dependent signals that include induction of extracellular signal-regulated kinase and phosphoinositide 3-kinase pathways as well as expression of COX-2. These events lead to activation of the transcription factor AP-2alpha, which interacts with specific regions in the EP4 gene promoter, leading to transcription of the EP4 gene.

Cyclooxygenase-2 gene transcription in a macrophage model of inflammation.

Infections involving LPS-bearing, Gram-negative bacteria can lead to acute inflammation and septic shock. Cyclooxygenase-2 (COX-2), the target of nonsteroidal anti-inflammatory drugs and selective COX-2 inhibitors, is importantly involved in these responses. We examined the dynamics of COX-2 gene expression in RAW264.7 murine macrophages treated with LPS as a model for COX-2 gene expression during inflammation. We established, using Northern blotting, nuclear run-on assays, and RT-PCR, that COX-2 transcriptional activation continues for at least 12 h after LPS treatment and involves at least three phases. Previous studies with murine macrophages identified an NF-kappaB site, a C/EBP site, and a cAMP response element-1 (CRE-1) as cis-acting elements in the COX-2 promoter. We identified three additional functional elements including a second CRE (CRE-2), an AP-1 site, and an E-box that overlaps CRE-1. The E-box mediates transcriptional repression whereas the other cis-elements are activating. Using electrophoretic mobility supershift and chromatin immunoprecipitation assays, we cataloged binding to each functional cis element and found them occupied to varying extents and by different transcription factors during the 12 h following LPS treatment. This suggests that the cis elements and their cognate transcription factors participate in a sequential, coordinated regulation of COX-2 gene expression during an inflammatory response. In support of this concept, we found, using inhibitors of Jun kinase and NF-kappaB p50 nuclear localization, that COX-2 gene transcription was completely dependent on phospho-c-Jun plus p50 at 6 h after LPS treatment but was only partially dependent on the combination of these factors at later treatment times.

Partnering between monomers of cyclooxygenase-2 homodimers.

Prostaglandin endoperoxide H synthases (PGHSs) 1 and 2 convert arachidonic acid to prostaglandin H2 in the committed step of prostanoid biosynthesis. These enzymes are pharmacological targets of nonsteroidal antiinflammatory drugs and cyclooxygenase (COX) 2 inhibitors. Although PGHSs function as homodimers and each monomer has its own COX and peroxidase active sites, the question of whether there is cross-talk between monomers has remained unresolved. Here we describe two heterodimers in which a native subunit of human PGHS-2 has been coupled to a subunit having a defect within the COX active site at some distance from the dimer interface. Native/G533A PGHS-2, a heterodimer with a COX-inactive subunit, had the same specific COX activity as the native homodimer. Native/R120Q PGHS-2, a heterodimer in which both subunits can oxygenate arachidonic acid but in which the R120Q subunit cannot bind the COX inhibitor flurbiprofen, was inhibited by flurbiprofen to about the same extent as native PGHS-2. These results imply that native PGHS-2 exhibits half-of-sites reactivity. Isothermal titration calorimetry established that only one monomer of the native PGHS-2 homodimer binds flurbiprofen tightly. In short, binding of ligand to the COX site of one monomer alters its companion monomer so that it is unable to bind substrate or inhibitor. We conclude that PGHS monomers comprising a dimer, although identical in the resting enzyme, differ from one another during catalysis. The nonfunctioning subunit may provide structural support enabling its partner monomer to catalyze the COX reaction. This subunit complementarity may prove to be characteristic of other dimeric enzymes having tightly associated monomers.

Activation of peroxisome proliferator-activated receptor beta/delta (PPARbeta/delta) increases the expression of prostaglandin E2 receptor subtype EP4. The roles of phosphatidylinositol 3-kinase and CCAAT/enhancer-binding protein beta.

The prostaglandin E2 receptor subtype EP4 has been implicated in the growth and progression of human non-small cell lung carcinoma (NSCLC). However, the factors that control its expression have not been entirely elucidated. Our studies show that NSCLC cells express peroxisome proliferator-activated receptor beta/delta (PPARbeta/delta) protein and that treatment with a selective PPARbeta/delta agonist (GW501516) increases EP4 mRNA and protein levels. GW501516 induced NSCLC cell proliferation, and this effect was prevented by PPARbeta/delta antisense or EP4 short interfering RNA (siRNA). GW501516 increased the phosphorylation of Akt and decreased PTEN expression. The selective inhibitor of phosphatidylinositol 3-kinase (PI3-K), wortmannin, and PPARbeta/delta antisense, abrogated the effect of GW501516 on EP4 expression, whereas that of the inhibitor of Erk did not. GW501516 also increased EP4 promoter activity through effects on the region between -1555 and -992 bp in the EP4 promoter, and mutation of the CCAAT/enhancer-binding protein (C/EBP) site in this region abrogated the effect of GW501516. GW501516 increased not only the binding activity of C/EBP to the NF-IL6 site in the EP4 promoter, which was prevented by the inhibitor of PI3-K, but also increased C/EBPbeta protein in a dose- and PPARbeta/delta-dependent manner. The effect of GW501516 on EP4 protein was eliminated in the presence of C/EBPbeta siRNA. Finally, we showed that pretreatment of NSCLC with GW501516 further increased NSCLC cell proliferation in response to exogenous dimethyl-prostaglandin E2 (PGE2) that was diminished in the presence of PPARbeta/delta antisense and EP4 siRNA. Taken together, these findings suggest that activation of PPARbeta/delta induces PGE2 receptor subtype EP4 expression through PI3-K signals and increases human lung carcinoma cell proliferation in response to PGE2. The increase in transcription of the EP4 gene by PPARbeta/delta agonist was associated with increased C/EBP binding activity in the NF-IL6 site of EP4 promoter region and C/EBPbeta protein expression that were mediated through both PI3-K/Akt and PPARbeta/delta signaling pathways.

HMG-CoA reductase inhibitors induce COX-2 gene expression in murine macrophages: role of MAPK cascades and promoter elements for CREB and C/EBPbeta.

Except functioning as lipid-lowering agents, HMG-CoA inhibitors, statins, are good tools to clarify the signaling role of small G proteins. In this study, we found in murine RAW264.7 macrophages, statins within 1-30 microM stimulated COX-2 gene transcription and PGE(2) formation, displaying potencies as lovastatin > fluvastatin > atorvastatin >> pravastatin. Transfection experiments with COX-2 promoter construct showed the necessity of C/EBPbeta and CRE promoter sites, but not NF-kappaB promoter site. Effects of statins on the activation of COX-2 promoter, induction of COX-2 protein, and PGE(2) production were all prevented by mevalonate and prenylated metabolites, FPP and GGPP. Consistent with the effect of statins, manumycin A, farnesyltransferase inhibitor, and geranylgeranyltransferase inhibitor increased PGE(2) production and COX-2 induction. Likewise, toxin B, an inhibitor of Rho family members, caused a prominent COX-2 induction. Results also indicated that tyrosine kinase, ERK, and p38 MAPK play essential roles in statin action. Taken together, these results not only demonstrate a unique action of statins in the upregulation of COX-2 expression in macrophages, but also suggest a negative role controlled by small G proteins in COX-2 gene regulation. Removal of this negative control by impairing G protein prenylation with statins leads to MAPKs activation and promotes COX-2 gene expression through the activation at CRE and C/EBPbeta sites.

The tomato homolog of CORONATINE-INSENSITIVE1 is required for the maternal control of seed maturation, jasmonate-signaled defense responses, and glandular trichome development.

Jasmonic acid (JA) is a fatty acid-derived signaling molecule that regulates a broad range of plant defense responses against herbivores and some microbial pathogens. Molecular genetic studies in Arabidopsis have established that JA also performs a critical role in anther and pollen development but is not essential for other developmental aspects of the plant's life cycle. Here, we describe the phenotypic and molecular characterization of a sterile mutant of tomato (jasmonic acid-insensitive1 [jai1]) that is defective in JA signaling. Although the mutant exhibited reduced pollen viability, sterility was caused by a defect in the maternal control of seed maturation, which was associated with the loss of accumulation of JA-regulated proteinase inhibitor proteins in reproductive tissues. jai1 plants exhibited several defense-related phenotypes, including the inability to express JA-responsive genes, severely compromised resistance to two-spotted spider mites, and abnormal development of glandular trichomes. We demonstrate that these defects are caused by the loss of function of the tomato homolog of CORONATINE-INSENSITIVE1 (COI1), an F-box protein that is required for JA-signaled processes in Arabidopsis. These findings indicate that the JA/COI1 signaling pathway regulates distinct developmental processes in different plants and suggest a role for JA in the promotion of glandular trichome-based defenses.

Bt: mode of action and use.

The insecticidal toxins from Bacillus thuringiensis (Bt) represent a class of biopesticides that are attractive alternatives to broad-spectrum "hard" chemistries. The U.S. Food Quality Protection Act and the European Economic Council directives aimed at reducing the use of carbamate and organophosphate insecticides were expected to increase the use of narrowly targeted, "soft" compounds like Bt. Here we summarize the unique mode of action of Bt, which contributes to pest selectivity. We also review the patterns of Bt use in general agriculture and in specific niche markets. Despite continued predictions of dramatic growth for biopesticides due to US Food Quality Protection Act-induced cancellations of older insecticides, Bt use has remained relatively constant, even in niche markets where Bt has traditionally been relatively high.