Non-invasive NIR-IIa fluorescence imaging of blood vessels in the mouse brain. Image reproduced from Hong et al., Nature Publishing Group.

Typically, fluorescence imaging of the brain is performed in the visible or near-infrared regions from 400 to 900 nanometers (nm) and requires thinning or removal of the skull above the area of interest, which creates a cranial window. Nevertheless, the imaging depth is limited to around 1 mm because of light scattering. Hongjie Dai, Calvin Kuo and their colleagues at Stanford University show that they can improve penetration depth and resolution by imaging blood vessels using a narrow region in the near infrared that they call NIR-IIa and that ranges from 1,300 to 1,400 nm. This imaging technology does not require a cranial window. Dai says that the NIR-IIa wavelengths are the longest fluorescence wavelengths imaged so far.

Imaging in the NIR-IIa region has several benefits. Both light scattering and autofluorescence are reduced in this wavelength region. Therefore, higher penetration depths can be achieved compared to those of more traditional imaging with shorter wavelengths. In addition, light attenuation by water in the skin, skull and brain tissue is still low in this wavelength range, making it possible to image into the brain non-invasively through the intact skull.

To visualize blood vessels, Dai and his colleagues injected single-walled carbon nanotubes into the bloodstream of mice. “Carbon nanotubes can be tuned to different wavelengths by changing the nanotube diameter,” says Dai, which makes them ideal for a variety of imaging conditions. To assess the performance of their imaging regimen, Dai and his team used nanotubes that were conjugated to a dye that fluoresces in the near-infrared region at around 800 nm, called NIR-I. When these nanotubes were imaged through the skull in the NIR-I region, major blood vessels were blurry, and capillaries could not be discerned. In contrast, even small capillaries could be observed when imaged in the NIR-IIa region. In fact, the researchers routinely imaged deeper than 2 mm, with a resolution better than 10 μm. They think that the image quality is improved because around 50% fewer photons are scattered by the skull and scalp in the NIR-IIa region than in the NIR-I region.

With this imaging technique, it is also possible to image blood flow at high temporal resolution. From the resulting movies, Dai and his team extracted parameters that allowed them to classify the imaged vessels as either arteries or veins. They applied the technology to visualize impaired blood flow in mouse models for stroke and observed severely reduced perfusion in the regions affected by arterial occlusion.

Dai thinks that the imaging technology can be applied to a number of different areas such as tumor, skin, eye and brain imaging. He is excited about the possibility of applying this technology to functional brain imaging. “It would be necessary to develop a fluorescent agent in the NIR-II region that responds to brain activity ... like calcium sensor dyes,” says Dai. Because of the low autofluorescence in the NIR-IIa window, Dai envisions that imaging in this region could also be suitable for single-cell or even single-molecule imaging in vivo.