The light dose and wavelength of illumination proved to be essential parameters for maintaining cell viability in optical microscopy. Non-phototoxic light doses, i.e.
, those doses where plating efficiency decreased by less than 10% in comparison with non-irradiated controls, were up to 200 J/cm2
(increasing with excitation wavelength) for non-incubated cells, but decreased upon application of a fluorescent dye. Photosensitization and light-induced generation of cytotoxic reactive oxygen species, e.g., singlet oxygen or superoxide radicals, has been related to various intrinsic molecules, e.g., flavins [16
] or porphyrins [17
], as well as fluorescent dyes, e.g., organelle markers or photosensitizers used in photodynamic therapy (PDT) [19
]. This may explain the observed dependence of non-phototoxic light doses on excitation wavelength as well as on application of a certain fluorophore. The lowest range of non-phototoxic light doses of 0–0.25 J/cm2
was recently determined for the photosensitizer protoporphyrin IX (accumulation induced after incubation with 5-aminolevulinic acid (5-ALA), a precursor of porphyrin biosynthesis) [20
] which corresponded to only 0–2.5 s of solar irradiation. Reduction of phototoxicity by controlled light exposure in fluorescence microscopy has been well documented in the literature [21
]. Currently there is no experimental evidence that thermal damage might also affect cell viability at moderate powers of illumination used in the present study, since control measurements with an infrared camera (VarioScan, high resolution model 3021-ST; Jenoptik, Jena, Germany) did not show any temperature change above 0.2 K due to laser irradiation (633 nm; 200 J/cm2
, data not shown).
While cell viability was generally maintained in conventional wide field microscopy with highly sensitive camera systems, additional experiments proved that the limit for non-phototoxic light doses was easily attained in axially resolved microscopy with repetitive illumination of the same samples, e.g., in laser scanning microscopy or wide field microscopy with structured illumination. Therefore, the number of selected planes has either to be limited, or individual planes have to be illuminated without light exposure of the remaining parts of the sample. This has recently been achieved by Selective Plane Illumination Microscopy (SPIM) or Light Sheet Based Fluorescence Microscopy (LSFM) [7
Whether non-phototoxic light doses are the same upon cw and repetitive pulse excitation (e.g., in laser scanning microscopy) is still an open question. However, in all present experiments integral light dose (measured in J/cm2
) turned out to be the main parameter for colony formation, rates of photobleaching or morphological changes. Also Murray et al
] found equal amounts of photobleaching in a wide field and a laser scanning microscope, when the total dose of illumination was the same. Studies of further phenomena, e.g., cell damage due to non-linear interactions of very short, intensive light pulses in the picosecond to femtosecond range, or cell recovery upon prolonged light exposure, were not addressed by the present experiments.
Light doses upon TIR and epi-illumination were calculated as a product of intensity and exposure time. Due to comparably low absorption and scattering within a cell layer of about 10 μm thickness, light intensity upon epi-illumination remained almost unchanged all over the cells. Upon TIR illumination the intensity I0
of the electromagnetic field was enhanced by a transmission factor T≈3 on the glass surface, but decreased by about the same factor at an average distance Δ≈ 100 nm between the glass substrate and the cell membrane according to the relation I = T I0
, with d corresponding to the penetration depth of the evanescent wave which was again about 100 nm [14
]. Therefore, light intensity on the cell surface corresponded to the incident intensity, but decreased rapidly within the cell. In comparison with conventional epi-illumination, cells were expected to be less sensitive to TIR illumination, where only the plasma membrane and adjacent cellular sites are exposed to light. Experimentally, non-phototoxic light doses upon TIR and epi-illumination differed by a factor of about 3 for cells incubated with laurdan (s. above) as well as for untreated cells (preliminary result, not shown). Although this factor appears rather small, lower sensitivity towards TIR illumination is e.g., advantageous in variable-angle TIR microscopy, where cell-substrate topology can be calculated with nanometer precision from a certain number (typically 10–15) of TIR images [15
]. TIR microscopy is also a preferential method for single molecule detection in layers of about 100 nm thickness. In this case, however, the irradiance should be around 100 W/cm2
in order to absorb 10−8
and to excite each molecule several thousand times per second, such that about 100–1,000 fluorescence photons per second can be recorded. Cells incubated with very low concentrations of fluorophores may have similar sensitivity to light as non-incubated cells, i.e.
, the maximum non-phototoxic light dose is expected to be about 100–200 J/cm2
upon epi-illumination and 300–600 J/cm2
upon TIR illumination. Therefore, the duration of a single molecule experiment should not exceed a few seconds. This should be considered in super-resolution microscopy based on single molecule techniques, e.g., STORM, PALM or single molecule based energy transfer measurements.