Mid-infrared chemical imaging of intracellular tau fibrils using fluorescence-guided computational photothermal microscopy

Alzheimer's Disease (AD) is a significant global health concern, and tau protein aggregates are a key pathological feature. Characterizing the structure and morphology of these aggregates in their cellular context is crucial for understanding tauopathies and developing therapeutics. Existing techniques, such as X-ray crystallography, Cryo-EM, and Circular Dichroism (CD) spectroscopy, have limitations in analyzing intracellular tau aggregates due to sample preparation complexities, cost, and limited spatial resolution or sensitivity to β-sheet structures. Positron Emission Tomography (PET) and fluorescence imaging offer high-resolution intracellular imaging but lack the capability to quantify secondary protein structures. Vibrational spectroscopic imaging methods, including Raman and Infrared (IR)-based techniques, offer potential advantages but often suffer from low imaging speed, weak signals, photodamage, or limited spatial resolution. Mid-Infrared Photothermal (MIP) microscopy shows promise, but 3D chemical imaging of intracellular amyloid protein aggregates has not been previously reported. This study addresses these challenges by developing FBS-IDT, a cost-effective, high-resolution, and non-invasive method for 3D chemical imaging and spectroscopic analysis of intracellular tau fibrils.

The introduction thoroughly reviews existing methods for characterizing tau aggregates and other amyloid proteins. It discusses the limitations of structural biology tools (X-ray crystallography, Cryo-EM, NMR), spectroscopy methods (CD spectroscopy), and imaging techniques (PET, fluorescence imaging) in analyzing intracellular tau aggregates in their native environment. The review highlights the advantages and disadvantages of Raman and IR-based vibrational spectroscopic imaging, emphasizing the superior sensitivity and applicability of IR for detecting amyloid protein aggregates. The limitations of conventional IR spectroscopy and AFM-IR are discussed, leading to the introduction of MIP microscopy as a promising alternative. The review concludes by stating that while MIP microscopy has been used for 2D imaging, 3D chemical imaging of intracellular amyloid protein aggregates had not been reported prior to this work.

FBS-IDT combines 3D label-free chemical imaging with 2D single-photon fluorescence imaging. The chemical imaging modality utilizes a pump-probe MIP tomography imaging scheme. A pulsed, tunable mid-IR pump beam induces chemical-specific volumetric MIP effects, and a visible probe laser from a custom laser ring array images the induced 3D refractive index (RI) variations. The 2D fluorescence imaging modality acts as a guide star to differentiate amyloid protein aggregates from background signals. The 3D RI maps are reconstructed using the Intensity Diffraction Tomography (IDT) method, which relates the sample's properties to scattering information from 2D intensity images. 'Hot' and 'Cold' states are created by modulating the mid-IR laser, and the chemical-specific 3D RI variation map is obtained by subtraction. Repeating this process across the mid-IR fingerprint region generates 4D hyperspectral chemical images. Under fluorescence mode, a 488nm excitation laser illuminates the sample, providing a 2D guide-star image for site-specific mid-IR fingerprint spectra extraction from the 4D data. Peak deconvolution of amide I bands in these spectra allows protein secondary structure analysis, enabling 3D visualization of β-sheet structures. The FBS-IDT system uses a simple brightfield transmission microscope with added components (450nm laser ring array, tunable mid-IR QCL, 488nm excitation laser) for pump-probe detection. Time synchronization between the probe and mid-IR lasers is crucial for high SNR. The system achieves a high imaging speed of ~6Hz. Human epithelial cells (Tau RD P301S FRET Biosensor cells) were used, cultured with or without seeded tau fibril fractions. Samples were fixed, washed, and immersed in D2O PBS between CaF2 glasses for imaging.

The study successfully demonstrated label-free hyperspectral 3D chemical imaging of fixed human epithelial cells with and without seeded tau fibrils. Depth-resolved reconstructions showed clear visualization of protein and lipid distributions within the cells. The chemical-specific cellular features showed variations along different axis positions. Depth-resolved mid-IR fingerprint spectra were extracted from the tau fibrils, revealing chemical structure differences between tau fibrils and diffusive tau protein. 3D visualization of the β-sheet structures within the tau fibrils was achieved using the spectral ratio map method. A potential correlation between lipid accumulation and tau aggregate formation was observed by comparing cells with and without seeded tau fibrils. The results indicated that FBS-IDT is a powerful tool for characterizing the 3D chemical composition and secondary protein structure of intracellular amyloid protein aggregates in their native environment.

The findings demonstrate the successful application of FBS-IDT for 3D chemical imaging and spectroscopic analysis of intracellular tau fibrils, overcoming limitations of previous methods. The ability to visualize both lipids and proteins simultaneously provides insights into the potential correlation between lipid accumulation and tau aggregation. The depth-resolved mid-IR spectra and 3D visualization of β-sheet structures provide detailed structural information about the tau fibrils. The cost-effectiveness and high speed of FBS-IDT make it a practical tool for studying amyloid protein aggregation in cellular contexts. The results suggest that FBS-IDT could contribute to a better understanding of the mechanisms underlying tauopathies and facilitate the development of novel therapeutic strategies.

FBS-IDT offers a significant advancement in the study of intracellular amyloid protein aggregates. Its ability to provide high-resolution 3D chemical imaging and site-specific spectroscopic analysis, combined with cost-effectiveness, opens new avenues for investigating the complex interplay between amyloid proteins and other cellular components. Future research could explore the application of FBS-IDT to other amyloid proteins, different cell types, and in vivo studies. Further development could also include improvements in the speed and automation of the system.

The current study was performed on fixed cells, which might affect the structural integrity of the tau fibrils compared to live cells. The sample preparation method could be further optimized to improve the signal-to-noise ratio and reduce potential artifacts. The use of D2O PBS might not perfectly mimic the physiological environment of live cells. Future studies should investigate the use of FBS-IDT with live cells and more sophisticated sample handling techniques.