In this article, we introduce a novel neural architecture for the end-to-end Korean Part-of-Speech (POS) tagging problem. To address the problem, we extend the present recurrent neural network-based sequence-to-sequence models to deal with the key challenges in this task: rare word generation and POS tagging. To overcome these issues, Input-Feeding and Copying mechanism are adopted. Although our approach does not require any manual features or preprocessed pattern matching dictionaries, our best single model achieves an F-score of 97.08. This is competitive with the current state-of-the-art model (F-score 98.03), which requires extensive manual feature processing.

Cholangiocarcinoma has a poor prognosis and is refractory to conventional chemotherapy and radiation therapy. Improving survival of patients with advanced cholangiocarcinoma urgently requires the development of new effective targeted therapies in combination with chemotherapy. We previously developed a human monoclonal antibody (mAb) Ab417 that binds to both the human and mouse L1 cell adhesion molecule (L1CAM) with high affinities. In the present study, we observed that Ab417 exhibited tumor targeting ability in biodistribution studies and dose-dependent tumor growth inhibition in an intrahepatic cholangiocarcinoma (Choi-CK) xenograft mouse model. Regarding the mechanism of action, Ab417 was internalized into the tumor cells and thereby down-regulated membrane L1CAM, and inhibited tumor growth by reducing tumor cell proliferation in vivo. Gemcitabine inhibited the tumor growth in a dose-dependent manner in the Choi-CK xenograft model. However, cisplatin inhibited the tumor growth moderately and not in a dose-dependent way, suggesting that the tumors may have developed resistance to apoptosis induced by cisplatin. Combined treatment with Ab417 and gemcitabine or cisplatin exerted enhanced tumor growth inhibition compared to treatment with antibody or drug alone. The results suggest that Ab417 in combination with chemotherapy may have potential as a new therapeutic regimen for cholangiocarcinoma. Our study is the first to show an enhanced therapeutic effect of a therapeutic antibody targeting L1CAM in combination with chemotherapy in cholangiocarcinoma models.

Plasmid DNA-encoded antibodies, or DNA-based monoclonal antibodies (dMAbs), are delivered by intramuscular injection and in vivo electroporation (EP) and are effective in virus neutralization, although they have not been evaluated for tumor gene therapy. Here we investigated whether a dMAb was appropriate for tumor gene therapy. We constructed the expression plasmids coding for the heavy or light chain of a parental murine antibody of Herceptin with the antibody genes codon- and RNA-optimized and fused to the Kozak-IgE leader sequence in pVax1. Transfection of the plasmids into human muscle RD cells resulted in functional expression of the antibody, and this exhibited the same in vitro antiproliferative activity as Herceptin. A single intramuscular injection and in vivo EP of the plasmids (100 μg per head) resulted in high and sustained antibody expression in the sera of normal mice and in effective inhibition of tumor growth in nude mice bearing HER2-positive human breast carcinoma BT474 xenografts. The antitumor efficacy of the anti-HER2 dMAb was similar to that of four doses of intravenously injected 10 mg kg−1 Herceptin. The results demonstrate that the dMAb is effective in the treatment of HER2-positive breast cancer, suggesting that this dMAb may be applicable for tumor gene therapy.

For Korean language processing, morphological analysis is a critical component that requires extensive work. This morphological analysis can be conducted in an end‐to‐end manner without requiring a complicated feature design using a sequence‐to‐sequence model. However, the sequence‐to‐sequence model has a time complexity of O(n2) for an input length n when using the attention mechanism technique for high performance. In this study, we propose a linear‐time Korean morphological analysis model using a local monotonic attention mechanism relying on monotonic alignment, which is a characteristic of Korean morphological analysis. The proposed model indicates an extreme improvement in a single threaded environment and a high morphometric F1‐measure even for a hard attention model with the elimination of the attention mechanism formula.

Extrahepatic cholangiocarcinoma (ECC), a malignant tumor of biliary origin, has a poor prognosis with limited treatment options. The KRAS oncogene is the most commonly mutated gene in ECC and one of the factors that predicts a poor prognosis and low survival rate. L1 cell adhesion molecule (L1CAM) is expressed in ECC cells and acts as an independent poor prognostic factor in predicting patient survival. In this study we investigate the functional significance of L1CAM in ECC cells with activating KRAS mutation. We selected an ECC cell line, EGI-1, with activating KRAS mutation, and then confirmed its expression of L1CAM by RT-PCR, western blot analysis, and flow cytometry. The suppression of L1CAM expression (using a specific lentivirus-delivered shRNA) significantly decreased the migratory and invasive properties of EGI-1 cells, without altering their proliferation or survival. Analyses of signaling effectors in L1CAM-depleted and control EGI-1 cells indicated that L1CAM suppression decreased the levels of both phosphorylated MKK4 and total MKK4, together with c-Jun N-terminal kinase (JNK) phosphorylation. Further, exposure to a JNK inhibitor (SP600125) decreased migration and invasion of EGI-1 cells. These results suggest that L1CAM promotes cellular migration and invasion via the induction of MKK4 expression, leading to JNK activation. Our study is the first to demonstrate a functional role for L1CAM in ECC carrying the activating KRAS mutation. Given that KRAS is the most commonly mutated oncogene in ECC, L1CAM may serve as an attractive therapeutic target for ECC cells with activating KRAS mutation.