A miniaturized confocal microscope reveals subsurfaces in vivo

Confocal microscopy can create clear images at various depths inside a thick specimen. Nonethless, bulky bench-top devices prevent many in vivo applications, especially in live animals where the objects to view must be in a fixed position relative to the microscope. The device described here solves that problem.

Scientists would like to use in-vivo imaging to obtain morphological descriptions of tissues and track cell-cell interactions. Imaging in a whole, live animal allows visualization of these events in their natural environment without the use of tissue culture, specimen processing,or fixation, all of which carry a risk of producing artifacts. The same problems exist for making diagnoses in human tissue. In addition, tissue evaluation after organ puncture poses some potential drawbacks: the risk of bleeding or infection, time-consuming processing, and the fact that ex-vivo histology provides only a static snapshot of the tissue of interest instead of monitoring it continuously over time. Likewise, sampling errors in untargeted biopsies or punctures may cause false-negative or false-positive diagnoses. Still, microscopic examination remains the gold standard for a definitive diagnosis of many human diseases.

Given the potential benefits, many researchers want to miniaturize in vivo-imaging devices for research or clinical applications.¹ In many instances, confocal microscopy is the most promising approach. It enhances spatial resolution by combining focal-laser illumination with a pinhole technique to geometrically eliminate out-of-focus light. This permits the visualization of fine cellular details many micrometers below the tissue surface.² However, attempts to miniaturize confocal-scanning heads can incur significant compromises in resolution, sensitivity, or both, which limits the results of in vivo imaging.

Recently, Optiscan (Victoria, Austraia) developed a miniaturized, confocal-scanning head and integrated it into a flexible video colonoscope that allows microscopic imaging of colonic lesions in screening and surveillance colonoscopy in humans. This device also yields highly accurate in vivo predictions of histomorphological changes as compared to traditional histology.³

Now, further miniaturization has permitted the integration of this scanning head into a hand-held, rigid, mini-microscopy probe with an outer diameter of 7mm. The probe is attached to the laser source and detection unit by a flexible connector. The solid-state laser delivers an excitation wavelength of 488nm, with an infiltration depth of 0–250μm, and light emission is detected at 505–585nm. Two remote buttons control the actuation of the imaging-plane depth along the range of the z-axis. During imaging, laser-power output at the tissue surface can be adjusted from 0 to 1000μW to achieve appropriate tissue contrast. Serial images are collected at a scan rate of 0.8frames/s at 1024 × 1024 pixels or 1.6frames/s at 1024 × 512 pixels, approximating a 1000-fold magnification on a 19-inch (48.26cm) screen. While imaging, a review mode allows further magnification of details. Images are captured with the help of a foot pedal and are digitally stored as gray-scale images.

We evaluated this probe for morphological and functional imaging in healthy mice and in different rodent models of human diseases.⁴ Various dyes—fluorescein sodium, acriflavine hydrochloride, fluorescein isothiocyanate (FITC)-labelled Lycopersicon esculentum lectin, and FITC-labelled dextrans—were intravenously injected to serve as fluorescent contrast agents. Acriflavine hydrochloride was also applied topically. All of these dyes allowed easy visualization of tissue characteristics. Fluorescein sodium gives an excellent impression of the general tissue architecture, but not the nuclei. This is due to its pharmacological properties. Acriflavine hydrochloride stains nuclei and—to a lesser extent—cytoplasm. FITC-labelled dextrans of high molecular weight yields an angiographic view of the tissue on a microscopic basis, since in healthy tissue no relevant extravasation (cells moving from blood vessels to tissues) occurs. FITC-labelled L. esculentum lectin binds to glycoprotein moieties, and renders vessel walls complementary to the plasma labelling of dextrans. Taken together, these staining methods allow comprehensive visualization of connective tissue, cellular and subcellular details down to nuclear structures, vessel walls, and imaging of perfusion after plasma labelling.

In conclusion, this confocal mini-microscopy probe works easily. It establishes a stable interface between the confocal-imaging window at the tip of the probe and the tissue region of interest. Scanning along the range of the z-axis yields stacks of images of different planes, from the tissue surface to subsurface levels. This permits full-resolution scanning-confocal-fluorescent imaging, suitable for examining living animals in realtime without major readjustments during the procedure.