Although both confocal imaging and computer deconvolution can be effective in virtually eliminating out of focus light from thick objects, which is necessary for intravital imaging, these techniques do nothing to alleviate another major problem of intravital imaging, phototoxicity. When a fluorophore is excited, there is a probability that instead of decaying to a singlet state and emitting a fluorescence photon, intra-system crossing will occur to a triplet state. These relatively long lived states are very reactive and can damage living cells and bleach the fluorophore in both living and fixed cells. One of the most significant damage mechanisms is the generation of highly reactive singlet oxygens from triplet states. When a specimen is being observed in a fluorescence microscope, fluorophore is excited throughout the bulk of the sample, even though only one focal plane is being observed at any time. Most of the phototoxic load, therefore, comes from regions away from the thin focal plane being observed. This problem can be circumvented to some extent by soaking fixed specimens in antioxidants. However, this approach is not practical for living material and bleaching can reach very high rates and phototoxicity can become very extensive in living samples.

The technique of multi-photon excitation provides an elegant solution to the problem of unwanted out-of-focus excitation in which the excitation is confined only to the optical section being observed. The sample is illuminated with light of a wavelength which is approximately twice (or three times) the wavelength of the absorption peak of the fluorophore in use. In the case of fluorescein, which has an absorption peak at approximately 490 nm, 980 nm excitation could be used. Essentially no excitation of the fluorophore will occur at the 980 nm wavelength because it is so far removed from the peak excitation wavelength of the fluorophore. In addition, no bleaching will occur, nor will phototoxic products be generated in the bulk of the sample. A high peak power pulsed laser source is used in pulses that are shorter than a picosecond so that the peak mean power levels are moderate and do not damage the specimen. With this source, the photon density at the point of focus will be sufficiently high for significant numbers of two and three photon events to occur. In a laser scanning microscope, the point of focus will describe a raster point as it scans and only at this focal point is the photon density high enough during the brief pulse for two or more photons to be absorbed simultaneously by the fluorophore. The absorption of two photons of long wavelength is equivalent to the absorption of one photon of half the wavelength, resulting in fluorescence excitation. Thus, in a multiphoton fluorescence microscope, one has the ideal probe in which fluorescence excitation is confined to the focal point under observation. Out of focus interference is eliminated simply because it is never generated. Go to figure 4.

Furthermore, a single laser line of sufficiently long wavelength can be used to excite multiple fluorophores by either two or three photon events. For example, the red and green dyes would be excited by two photon absorption while the blue and UV dyes by three photon absorption. Spectral measurements have shown that the peak of excitation is either the same or blue shifted and broader from that predicted for single photon excitation. As a result, even if the laser is not tuned precisely to the peak absorption it is still possible to obtain sufficient excitation for imaging of different dyes simultaneously.

Therefore, multi-photon excitation fluorescence microscopy has important advantages over other imaging modes of microscopy, particularly for the study of live cells and/or for thick tissues:

a.) The lack of fluorophore excitation in regions away from the focal plane has the consequence that the bulk fluorophore bleaching and generation of toxic by-products is reduced to minimum levels during imaging.

b.) Multi-photon images are less prone to degradation by light scattering. This is because the longer wavelengths used for excitation suffer less scattering from microscopic refractive index differences within the sample, allowing much greater penetration of the tissues. In addition, as all the resolution is defined by the geometry of the excitation beam, the fluorescence emission is unaffected by light scattering. The insensitivity of multi-photon imaging to light scattering is particularly advantageous for the study of living specimens as live tissue is highly light scattering due to the presence of many refractive index interfaces.

d.) Due to the additive effects of a, b and c above, images can obtained from deeper within a sample using multi-photon imaging than with any other type of light microscopy. The limiting depth is uncertain but in the rat model optical sections have been obtained that are 300 um deep routinely.

e.) Because wavelengths of different colors of light originating from one point source focus at different focal planes, in standard multi-probe confocal microscopy registration of multiple probes along the Z axis is problematic. Because multi-photon microscopy excites at one discreet diffraction limited point and because emitted light is collected without being deconvolved to different focal planes, true registration in the Z axis is achieved.

An impressive demonstration of depth of imaging is shown in Figure 6 and its corresponding sequence 13 in which a 120 um thick z-series has been rendered into a 3D rotating projection which shows the interaction of tumor cells with a blood vessel in a live primary breast tumor. The blood space was labelled by injecting 2M dalton rhodamine-dextran into the blood space. The uniform signal to noise throughout this projection and the depth at which the optical planes begin and end indicate that usable images can be obtained to depths of more than 200 um in intact live tissue.

An additional advantage of multi-photon imaging is the ability to image extracellular matrix. It has been shown that fibronectin, elastin and collagen containing fibers that ordinarily fluoresce when excited with ultraviolet light will emit light as a result of multi-photon excitation. As shown in sequence 14, optical sectioning through a primary tumor allows visualization of both tumor cells and matrix fibers simultaneously. Matrix can also be imaged separately from GFP fluorescence with a narrow band pass filter (sequence 15). Matrix content, which can also be expressed as concentration since the pixel intensity was collected in a known volume, is correlated with metastatic potential. These results suggest that metastatic tumors either synthesize less or degrade more extracellular matrix.

Finally, advantage can be taken of the ability to image extracellular matrix and cells simultaneously in vivo to characterize interactions between cells and matrix within the primary tumor (sequences 14 and 16, below). By separating the cell and matrix channels as shown in sequence 16, points of contact between cells and matrix fibers can be visualized directly. The number of such contacts on a single cell and the number of cells making contact per imaging field can be counted and used as a measure of the amount of adhesion between cells and matrix in the primary tumor.

Figure 6: A multiphoton stereo-pair of a mammary tumor formed by the injection of GFP- labeled tumor cells. The blood vessel is visualized by the injection of Rhodamine-labeled dextran. The tumor is imaged to a depth of 120 um.