Fluorescent Protein Tomography

Fluorescent Proteins (FP) have become essential reporter molecules to elucidate the function of proteins within cells, the bio-distribution of immune and stem cells and for evaluation of drug candidates in vivo. Confocal and multi-photon microscopy have emerged as powerful methods for imaging superficial locations down to a few hundred microns and in limited field of views (typically 1-2 mm). Alternative technologies to date use macroscopic reflectance imaging utilizing sensitive color cameras that obtain "photographs" of fluorescence activity emanating up to a few millimeters under the skin surface of animals. Such images are often biologically informative but they are surface weighted and do not allow accurate quantitation of underlying fluorescence activity. Furthermore such images are often impaired by high skin auto-fluorescence.


Figure 1.
The experimental used for tomographic measurements in the visible setup showing the laser source, the light delivery optical fibers (thick red lines), the CCD detector, the optical scanning switch and the imaging plate. The two branches used for reflectance (r) and trans-illumination (t) are also shown. b) Reflectance image of a mouse with two subcutaneously grown GFP expressing tumors. c) Trans-illumination image corresponding to a superposition of the gray scale raw CCD images for all source positions. From images b) and c) it is clear that in trans-illumination the background signal is very low and limited to the noise of the CCD camera, whereas in reflectance the skin autofluorescence is limiting the sensitivity of the detection.

We have been developing technologies for quantitative tomography imaging for three dimensional reconstructions fluorescent protein activity in the visible light range. To allow accurate and robust detection of fluorescent proteins in whole animals it is important to develop methods that account for non-linear propagation effects of photons into tissues. Such technology has been recently reported for tomographic imaging in the far red and near-infrared 5, 6, but the translation of such methods to the visible light range have not been simple. This is because tissue in the visible offers significantly higher photon attenuation requiring different instrumentation than the one used in the near-infrared. Furthermore photon propagation characteristics change significantly between visible and near infrared due to the higher background absorption potentially in need of better theoretical forward models. We have developed an animal scanner shown in Figure 1 based on cryogenically cooled CCD camera technology and applied this technique to imaging the progression of lung cancers in vivo in a murine model, as well as imaging of gene delivery through a herpes virus vector. The developed technique should be widely useful for several in vivo imaging applications, drug delivery and developmental biology.

Fpt2 Fpt1

Figure 2.
Tomographic reconstruction of two subcutaneously grown GFP expressing tumors. Three coronal slices at three different depths are shown overlaid with the white light image. In the first slice corresponding approximately to the middle of the body of the mouse no tumors are reconstructed. Instead the two tumors are reconstructed only in the superficial slices corresponding to the depths under the skin down to 3mm.