Immune cell-to-cell communication studied through label-free microscopy combined with machine learning

2019 Fiscal Year Final Research Report

• Project/Area Number: 18K14695
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 44010:Cell biology-related
• Research Institution: Osaka University
• Principal Investigator: Pavillon Nicolas 大阪大学, 免疫学フロンティア研究センター, 特任助教(常勤) (80644525)
• Project Period (FY): 2018-04-01 – 2020-03-31
• Keywords: 免疫反応 / 細胞コミュニケーション / 無標識顕微鏡法 / 生細胞イメージング / 食細胞 / リンパ球 / ラマん分光学 / 定量位相

Outline of Final Research Achievements

We employ our previously developed label-free non-invasive imaging system based on the simultaneous acquisition of quantitative phase images and Raman spectroscopy, which are indicative of morphology and molecular content, respectively. We couple these measurements on live cells with machine learning to derive statistical models that describe immune activation.
We validate our approach on primary cells derived from murine samples, which are much more heterogenous than cell lines, and achieve over 90% accuracy. We also show that we can distinguish different cellular sub-types that are difficult to identify with standard methods.
We also investigate the activation and differentiation of T cell lymphocytes, and show that we can detect T cell activation with over 95% accuracy, and also distinguish different types of activation, namely TCR stimulation and chemical bypass.

Free Research Field

生細胞イメージング

Academic Significance and Societal Importance of the Research Achievements

細胞を分析する標準的な方法は高感度を果たすが、信号の検出のために標識の必要がある。標識等は細胞反応を変わり、生細胞の場合は表面受容体しか測定できない。細胞内分子を検知できるために固定がある。
無標識顕微鏡は生細胞の計測ができ、直接に細胞内分子を測定できる。我々はこの方法と機械学習で細胞の繊細な変化を検出できた。免疫細胞の色々な種類、例えば食細胞やT細胞のリンパ球に、反応を検出できることを証明した。