A team of researchers at the Georgia Institute of Technology and Emory University has developed a deep-ultraviolet (UV) microscopy method that can rapidly assess T cell viability, activation state, and subtype—all without the need for fluorescent labels or cell destruction. The work, published in BME Frontiers, offers a transformative approach for immune monitoring and cell-based therapy development.
T cells are central to the immune system, and their characterization is critical for understanding immune function, tracking disease progression, and optimizing adoptive T cell therapies such as CAR-T. However, current gold-standard methods, like flow cytometry, require fluorescent labeling, expensive equipment, and typically destroy the cells during measurement. This limits real-time monitoring and longitudinal studies of live cell cultures.
The new approach uses static deep-UV images captured at 255 nm, a wavelength strongly absorbed by nucleic acids, to generate high-contrast images of live T cells without any exogenous stains. By training a custom residual neural network on images from five human donors, the researchers achieved high accuracy in classifying T cells into three categories: activated, dead, and quiescent (naïve or contraction-phase). The model's predictions showed excellent agreement with flow cytometry, with an R² > 0.97 for both viability and activation percentage.
A more challenging task is subtyping CD4⁺ helper T cells from CD8⁺ cytotoxic T cells. Static morphological features alone proved insufficient. To overcome this, the team turned to dynamic deep-UV imaging, acquiring 500-frame time series at ~8 Hz. By analyzing pixel-wise temporal fluctuations in the frequency domain using phasor analysis and power-law fitting, they quantified intracellular activity. A second neural network, fed with four-channel inputs (UV absorption, phasor g, phasor s, and power-law slope), distinguished CD4⁺ from CD8⁺ T cells with ~90% accuracy.
Notably, CD4⁺T cells exhibited significantly higher intracellular dynamic activity than CD8⁺ cells, consistent with known metabolic differences—CD4⁺ cells rely more on glycolysis and oxidative phosphorylation and have more cytoplasmic mitochondria. Pseudocolorized images revealed that the activity difference is localized to the cytoplasm, not the nucleus, further supporting the link to metabolic machinery.
Deep-UV Microscopy's potential applications are vast, spanning immunology research, immune monitoring, and the development of emerging cell-based therapies. Its simplicity, speed, and high resolution make it an invaluable tool for optimizing adoptive T cell therapies, tracking disease progression, and enhancing our fundamental understanding of immune function.
