Sunkyu Kim

MOTIVATION: Biomedical text mining is becoming increasingly important as the number of biomedical documents rapidly grows. With the progress in natural language processing (NLP), extracting valuable information from biomedical literature has gained popularity among researchers, and deep learning has boosted the development of effective biomedical text mining models. However, directly applying the advancements in NLP to biomedical text mining often yields unsatisfactory results due to a word distribution shift from general domain corpora to biomedical corpora. In this article, we investigate how the recently introduced pre-trained language model BERT can be adapted for biomedical corpora. RESULTS: We introduce BioBERT (Bidirectional Encoder Representations from Transformers for Biomedical Text Mining), which is a domain-specific language representation model pre-trained on large-scale biomedical corpora. With almost the same architecture across tasks, BioBERT largely outperforms BERT and previous state-of-the-art models in a variety of biomedical text mining tasks when pre-trained on biomedical corpora. While BERT obtains performance comparable to that of previous state-of-the-art models, BioBERT significantly outperforms them on the following three representative biomedical text mining tasks: biomedical named entity recognition (0.62% F1 score improvement), biomedical relation extraction (2.80% F1 score improvement) and biomedical question answering (12.24% MRR improvement). Our analysis results show that pre-training BERT on biomedical corpora helps it to understand complex biomedical texts. AVAILABILITY AND IMPLEMENTATION: We make the pre-trained weights of BioBERT freely available at https://github.com/naver/biobert-pretrained, and the source code for fine-tuning BioBERT available at https://github.com/dmis-lab/biobert.

Unlike normal mammalian cells, which use oxygen to generate energy, cancer cells rely on glycolysis for energy and are therefore less dependent on oxygen. We previously observed that the c-Myc oncogenic transcription factor regulates lactate dehydrogenase A and induces lactate overproduction. We, therefore, sought to determine whether c-Myc controls other genes regulating glucose metabolism. In Rat1a fibroblasts and murine livers overexpressing c-Myc, the mRNA levels of the glucose transporter GLUT1, phosphoglucose isomerase, phosphofructokinase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, and enolase were elevated. c-Myc directly transactivates genes encoding GLUT1, phosphofructokinase, and enolase and increases glucose uptake in Rat1 fibroblasts. Nuclear run-on studies confirmed that the GLUT1 transcriptional rate is elevated by c-Myc. Our findings suggest that overexpression of the c-Myc oncoprotein deregulates glycolysis through the activation of several components of the glucose metabolic pathway.

UNLABELLED: Castration-resistant prostate cancer (CRPC) is the most aggressive, incurable form of prostate cancer. MDV3100 (enzalutamide), an antagonist of the androgen receptor (AR), was approved for clinical use in men with metastatic CRPC. Although this compound showed clinical efficacy, many initial responders later developed resistance. To uncover relevant resistant mechanisms, we developed a model of spontaneous resistance to MDV3100 in LNCaP prostate cancer cells. Detailed characterization revealed that emergence of an F876L mutation in AR correlated with blunted AR response to MDV3100 and sustained proliferation during treatment. Functional studies confirmed that AR(F876L) confers an antagonist-to-agonist switch that drives phenotypic resistance. Finally, treatment with distinct antiandrogens or cyclin-dependent kinase (CDK)4/6 inhibitors effectively antagonized AR(F876L) function. Together, these findings suggest that emergence of F876L may (i) serve as a novel biomarker for prediction of drug sensitivity, (ii) predict a "withdrawal" response to MDV3100, and (iii) be suitably targeted with other antiandrogens or CDK4/6 inhibitors. SIGNIFICANCE: We uncovered an F876L agonist-switch mutation in AR that confers genetic and phenotypic resistance to the antiandrogen drug MDV3100. On the basis of this fi nding, we propose new therapeutic strategies to treat patients with prostate cancer presenting with this AR mutation.