Lung cancer has a high mortality rate, but an early diagnosis can contribute to a favorable prognosis. A liquid biopsy that captures and detects tumor-related biomarkers in body fluids has great potential for early-stage diagnosis. Exosomes, nanosized extracellular vesicles found in blood, have been proposed as promising biomarkers for liquid biopsy. Here, we demonstrate an accurate diagnosis of early-stage lung cancer, using deep learning-based surface-enhanced Raman spectroscopy (SERS) of the exosomes. Our approach was to explore the features of cell exosomes through deep learning and figure out the similarity in human plasma exosomes, without learning insufficient human data. The deep learning model was trained with SERS signals of exosomes derived from normal and lung cancer cell lines and could classify them with an accuracy of 95%. In 43 patients, including stage I and II cancer patients, the deep learning model predicted that plasma exosomes of 90.7% patients had higher similarity to lung cancer cell exosomes than the average of the healthy controls. Such similarity was proportional to the progression of cancer. Notably, the model predicted lung cancer with an area under the curve (AUC) of 0.912 for the whole cohort and stage I patients with an AUC of 0.910. These results suggest the great potential of the combination of exosome analysis and deep learning as a method for early-stage liquid biopsy of lung cancer.

Summary Selective degeneration of midbrain dopaminergic (mDA) neurons is associated with Parkinson’s disease (PD), and thus an in-depth understanding of molecular pathways underlying mDA development will be crucial for optimal bioassays and cell replacement therapy for PD. In this study, we identified a novel Wnt1-Lmx1a autoregulatory loop during mDA differentiation of ES cells, and confirmed its in vivo presence during embryonic development. We found that the Wnt1-Lmx1a autoregulatory loop directly regulates Otx2 through the β-catenin complex and Nurr1 and Pitx3 through Lmx1a. We also found that Lmx1a and Lmx1b co-operatively regulate mDA differentiation with overlapping and cross-regulatory functions. Furthermore, co-activation of both Wnt1 and SHH pathways by exogenous expression of Lmx1a, Otx2 and FoxA2 synergistically enhanced the differentiation of ES cells to mDA neurons. Together with previous works, this study shows that two regulatory loops (Wnt1-Lmx1a and SHH-FoxA2) critically link extrinsic signals to cell-intrinsic factors and cooperatively regulate mDA neuron development.

Owing to the role of exosome as a cargo for intercellular communication, especially in cancer metastasis, the evidence has been consistently accumulated that exosomes can be used as a noninvasive indicator of cancer. Consequently, several studies applying exosome have been proposed for cancer diagnostic methods such as ELISA assay. However, it has been still challenging to get reliable results due to the requirement of a labeling process and high concentration of exosome. Here, we demonstrate a label-free and highly sensitive classification method of exosome by combining surface-enhanced Raman scattering (SERS) and statistical pattern analysis. Unlike the conventional method to read different peak positions and amplitudes of a spectrum, whole SERS spectra of exosomes were analyzed by principal component analysis (PCA). By employing this pattern analysis, lung cancer cell derived exosomes were clearly distinguished from normal cell derived exosomes by 95.3% sensitivity and 97.3% specificity. Moreover, by analyzing the PCA result, we could suggest that this difference was induced by 11 different points in SERS signals from lung cancer cell derived exosomes. This result paved the way for new real-time diagnosis and classification of lung cancer by using exosome as a cancer marker.