The heme precursor 5-aminolevulinic acid (5-ALA) has been used for local (e.g., urinary bladder, skin, esophagus and colon) and systemic (brain) applications in humans, resulting in an epithelium-selective accumulation of its photosensitizing metabolite protoporphyrin IX (PPIX).1, 2 Epithelial selectivity is an important pharmacological property since it targets cancer therapy to cells of interest, avoiding the direct damage of photodynamic therapy (PDT) in adjacent stromal cells,3, 4 as has been found with other photosensitizers.5, 6
In bladder tumors, 5-ALA has also proven useful for diagnostic purposes since a high sensitivity of tumor detection and flat tumor precursors were found on fluorescence endoscopy by applying a filtered lamp system and detecting red fluorescent lesions, caused by intracellular PPIX accumulation.7–9 These findings are supported by genetic analyses of the urothelial lesions detected by PPIX fluorescence since they showed tumor-like genetic changes even in lesions such as hyperplasias, normally diagnosed as benign.10, 11
The fluorescence intensity ratio between tumor and normal surrounding urothelium seen in photodynamic diagnosis with 5-ALA further encourages the idea of intravesical treatment of tumors since, besides stromal cells, we can assume that non-neoplastic urothelium is damaged less than tumor cells by this type of treatment. First, clinical trials using PDT with 5-ALA in patients with urothelial cancer have shown positive results;12, 13 however, 5-ALA dose and incubation times for these trials were empirical.
To define parameters for the clinical situation, we (i) used a normal transfected urothelial cell line to represent normal urothelium and a cell line of a papillary well-differentiated tumor to represent an early phase of urothelial tumor development and (ii) applied this reproducible in vitro system to testing of fluorescence intensity, localization of PPIX and phototoxicity in dependence of cell type, 5-ALA concentration and incubation time.
MATERIAL AND METHODS
Cell lines and culturing
Two human cell lines, of varying urothelial differentiation, were used for all experiments. RT4 cells, originally derived from a recurrent papillary G1 tumor,14 and UROtsa cells,15 derived from normal urothelium and immortalized by a temperature-sensitive SV40 large T-antigen gene construct (provided by Dr. J.R.W. Masters, University College, London, UK), were used as a model for a highly differentiated papillary tumor vs. normal urothelium. Both cell lines were maintained as monolayer cultures in RPMI-1640 (Biochrom, Berlin, Germany) without phenol red supplemented with 5% FCS (Sigma, Deisenhofen, Germany), 1% (v/v) L-glutamine and 1% (v/v) sodium pyruvate (both from GIBCO, Eggenstein, Germany) at 37°C in a humidified atmosphere containing 5% carbon dioxide. Cells were detached for subculturing and experiments using 0.1% trypsin/0.04% EDTA (GIBCO) in PBS (Biochrom) for 3 to 5 min. All experiments were performed using plateau-phase cells. To reach plateau phase, growth-defined cell densities were seeded into culture dishes and allowed to grow for the times indicated: RT4, 22,000 cells/cm2, 9 days; UROtsa, 5,600 cells/cm2, 10 days.
5-ALA incubation and illumination
Stock solutions of 5-ALA (Synopharm, Barsbüttel, Germany) were prepared in de-ionized water at a concentration of 10 μg/ml and stored at –20°C. For each experiment, stock solution was diluted in culture medium without FCS. Subsequent to 5-ALA incubation, culture medium was removed and the cell monolayer rinsed with PBS to remove remaining FCS. Cells were incubated with 5-ALA at concentrations of 100 and 200 μg/ml in culture medium without FCS for 3 and 1 hr at a total volume of 250 μl/cm2. In subsequent handling, care was taken to avoid exposure of cells to ambient light. After incubation, the 5-ALA solution was removed and fresh medium without FCS added. Cells incubated with the solution containing 100 μg/ml 5-ALA for 3 hr were illuminated with 1.5 J/cm2, and cells incubated with the solution containing 200 μg/ml 5-ALA for 1 hr were illuminated with 2.5 J/cm2. All illuminations were performed using incoherent white light (400–700 nm) of a high-pressure xenon arc lamp (Karl Storz, Tuttlingen, Germany) at a power density of 50 mW/cm2. After PDT, medium without FCS was removed and fresh culture medium with FCS added.
For flow-cytometric measurements of cellular PPIX content, cells were seeded into 6-well dishes as described above. After incubation for 1 or 3 hr, cells were detached, spun at 200 g for 5 min and resuspended in 1 ml PBS. Red fluorescence of PPIX was detected in the FL3 channel (650 nm longpass [lp]) of a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) after gating for forward and right-angle light scatter. Mean PPIX red fluorescence was corrected for autofluorescence by calculating the ratio of the mean intensity values of 5-ALA–incubated and sham-treated control cells.
For fluorescence microscopic determination of cellular PPIX localization, cells were seeded into petriPERM culture dishes (In Vitro Systems and Service, Osterode, Germany) at identical densities, as described above. Cells were incubated for 1 hr with 200 μg/ml 5-ALA or 3 hr with 100 μg/ml 5-ALA. For the last 30 min of incubation, cells were stained for mitochondria using MitoTracker green (Molecular Probes, Eugene, OR) at a concentration of 100 nM (dissolved in PBS). Cellular PPIX localization was visualized using an Axiovert S 100 microscope (Zeiss, Oberkochen, Germany) equipped with a mercury short-arc lamp (HBO50W, Osram, Munich, Germany), including specific filter sets for PPIX red fluorescence (excitation 360 ± 20 nm, emission 630 ± 30 nm) and mitochondrial green fluorescence (excitation 480 ± 15 nm, emission 535 ± 20 nm). Images were obtained with a Plan-Apochromat lens (63× 1.4) and recorded with a high-resolution (4,096 gray levels, 1,317 × 1,035 pixels, 6.8 × 6.8 μm pixel size) and Peltier element-cooled (–10°C) charge-coupled device camera (Princeton, Philadelphia, PA). Each color had to be recorded and digitally processed (filtering, contrast enhancement) separately using the Metamorph software package (Universal Imaging/Visitron Systems, Puchheim, Germany). Finally, corresponding images were superimposed.
Two of the main targets of PDT were analyzed, to characterize cellular damage: plasma membrane as the main location of sensitizer accumulation (see PPIX localization) and mitochondria as the main location of PPIX generation. In addition, general cellular morphology was examined by fluorescence and phase-contrast microscopy. Phototoxic effects were investigated 0 to 48 hr after PDT (1, 2, 4, 24 and 48 hr), with 0 hr serving as untreated control.
Cells, seeded into 6-well dishes as described above, were used for flow-cytometric analysis of membrane integrity with propidium iodide (PI) exclusion of vital cells. After 5-ALA incubation and white light illumination (see above), cells were detached, spun at 200 g and resuspended in 1 ml PBS. A stock solution of PI (1 mg/ml in PBS) was prepared, 3 μl of which were added to the cell suspension, leading to a final concentration of 3 μg/ml. Following a short incubation interval (3 to 5 min at room temperature), the cell suspension was measured on a FACSCalibur flow cytometer. After gating for forward vs. right-angle light scatter, the percentage of cells showing red PI fluorescence (measured in channel FL3, 650 nm lp) was determined.
Additionally, cell volume and membrane integrity of an aliquot of cells was analyzed using a cell counter and analyzer system (CASY1; Schärfe, Reutlingen, Germany). This system combines the principle of measuring changes in the defined resistance of a capillary during the passage of a cell (an intact cell acts as resistance) and analysis of the pulse area. Calculation of cell volume, diameter and, in consequence, membrane integrity is possible using this method.
For the determination of mitochondrial activity, cells were seeded into 6-well culture dishes as described above. A stock solution of 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1; Molecular Probes, Leiden, the Netherlands) of 500 μg/ml in 70% methanol was prepared. JC-1 is a J aggregate–forming dye that accumulates in mitochondria and builds multimeres, depending on the potential across the inner mitochondrial membrane, generally accepted as an indicator of mitochondrial activity. Fluorescence emission of these aggregates after 488 nm laser excitation is shifted to the red (emission maximum 590 nm, detected in channel FL2, 585/42 nm bandpass (bp)) and can easily be separated from monomer fluorescence (emission maximum 527 nm, detected in channel FL1, 530/30 nm bp) on a FACSCalibur flow cytometer using adequate instrument and correct compensation settings (FL1 monomer, 310 V; FL2 aggregate, 259 V; FL1-% FL2, 2, 0; FL2-% FL1, 20, 9). Detached cells were centrifuged at 200 g, the supernatant was removed, cells were resuspended in 1 ml PBS and 10 μl of stock solution were added to a final concentration of 5 μg/ml. After a 15 min incubation at room temperature, the cell suspension was measured without washing on a FACSCalibur flow cytometer.
Cells were seeded into petriPERM culture dishes and, after 5-ALA incubation and irradiation, stained with acridine orange (Sigma) at a final concentration of 13 μg/ml for 5 min to visualize cellular alterations using an inverted microscope (for details, see PPIX localization, above). Resulting images were processed and superimposed as described above.
For analysis of morphological alterations, cells were examined by phase-contrast microscopy. A defined number of cells was seeded into petriPERM culture dishes and, after 5-ALA incubation and irradiation, examined using the microscope described above (see PPIX localization) fitted with a 12 V/100 W tungsten halogen phase-contrast illumination device. Images were obtained with a phase-contrast lens (Plan-Neofluar 40x/0.75) and processed as described above (PPIX localization).
Each experiment was performed in triplicate at least. Data are presented as means ± SD. Differences were tested for statistical significance using the 2-sided t-test. p values <0.05 were considered statistically significant.
Flow-cytometric analysis of PPIX red fluorescence allows determination of relative sensitizer content of 5-ALA–incubated and sham-treated control cells and comparison of the 2 cell lines used. Both cell lines, RT4 tumor and UROtsa normal urothelial cells, showed significantly higher PPIX levels compared with sham-treated control cells after 1 and 3 hr incubations (p < 0.0008 in all cases). Cellular PPIX levels appeared to increase linearly over incubation time (0–3 hr). Relative PPIX content in RT4 cells was approximately 2- to 3-fold higher than in UROtsa cells for 1 and 3 hr incubations (p < 0.006 for both conditions). This confirms the increase in the amount of sensitizer up to 3 hr incubation. Compared to sham-treated control cells, RT4 tumor cells accumulated 18 and 68 and UROtsa 8 and 22 arbitrary units of PPIX after 1 and 3 hr incubations, respectively (Fig. 1).
Fluorescence microscopic images of RT4 tumor cells show very similar sensitizer localization after both 1 and 3 hr incubations. Most of the cellular PPIX was located in membranes, i.e., plasma membranes, and regions close to membranes. Some was also diffusely distributed in the cytoplasm of RT4 cells (Fig. 2, left column). In contrast, UROtsa cells showed different PPIX localization after 1 and 3 hr incubations. Short incubation times (1 hr) led to punctuate distribution of PPIX in the cytoplasm but no significant co-localization with mitochondria. Long incubation periods (3 hr) resulted in membranous and diffuse cytoplasmic localization of PPIX (Fig. 2, right column) as seen in RT4 cells.
Based on the above results, we investigated plasma membrane and mitochondria for phototoxicity, which are the main sites of sensitizer localization (plasma membrane) and PPIX formation in the heme biosynthetic pathway (mitochondria). Different phototoxic effects are to be expected for different PPIX localizations, i.e., for UROtsa cells after 1 and 3 hr incubations.
Membrane integrity was examined by flow-cytometric analysis with PI exclusion 0 to 48 hr after PDT. Both cell lines showed an increasing number of PI-positive cells within the experimental period. While the number of dead cells reached a plateau after about 4 hr in RT4 (3 hr incubation; Fig. 3, closed triangles), 1 hr incubation resulted in a continuously increasing number of dead cells up to 48 hr (Fig. 3, black circles). At the end of the experimental period, tumor cells showed nearly identical levels of PI-positive cells (70% to 80%) for both incubation conditions examined (differences not significant). The number of PI-positive UROtsa cells increased continuously using a 3 hr incubation time (Fig. 3, gray triangles), while 1 hr incubation resulted in only slightly higher numbers of dead cells during early periods after PDT compared to sham-treated control cells (Fig. 3, open circles). Compared with RT4, PDT of UROtsa resulted in significant differences in the percentage of dead cells at the end of our experiments (p < 0.02), indicating an improvement of differential phototoxicity on tumor and normal urothelial cells using shorter incubation times.
Figure 4 shows representative flow-cytometric measurements 0 to 48 hr after PDT, and our results of mitochondrial analysis are summarized in Figure 5. Most of the response to PDT appeared within the first hours (2 to 4 hr), and this effect was common for all cell lines and incubation and illumination conditions used. While the collapse of the mitochondrial transmembrane potential was complete in RT4 cells 2 to 4 hr following PDT using 3 hr incubation, it continued until the end of the experimental period using 1 hr incubation. Forty-eight hours after PDT, RT4 tumor cells showed a slightly different (not significant) but similar response (Figs. 4, 5), indicating a slower process of phototoxic action. In contrast, as observed in our analysis of PI exclusion, the differences in the response of normal UROtsa cells to PDT using the 2 incubation conditions were much more distinctive. The short incubation time of 1 hr leaves cells nearly unaffected, with only a small population demonstrating damaged mitochondria 2 hr after illumination. At the end of our observations, UROtsa cells exhibited no significant difference in mitochondrial function compared to sham-treated control cells (Figs. 4, 5) applying the short incubation time (1 hr). Using 3 hr incubation led to a markedly increased population of cells with mitochondria with collapsed transmembrane potential. This occurred as early as 1 hr after PDT and reached its maximum after 4 hr. However, even 48 hr after PDT, UROtsa cells showed a significant difference (p < 0.02) compared to cells treated with the short incubation (Figs. 4, 5).
To confirm our findings concerning differential phototoxicity of tumor and normal urothelial cells, we analyzed phase-contrast and fluorescence microscopic images 0 to 48 hr after PDT using the 2 different incubation conditions. On fluorescence microscopic images, intact cells were characterized by homogenous green fluorescence of nuclei and punctate red fluorescence of lysosomes due to accumulation of acridine orange. Necrotic cells lost both their green homogenous nuclear and red punctate lysosomal fluorescence. Cells that underwent apoptosis showed typical nuclear fragmentation (seen as bright green spots in the nuclei), mostly retaining their red lysosomal fluorescence.
For both cell lines examined, different pathways to cell death were observed following the 2 different treatment modalities. The short incubation interval induced morphological signs of apoptosis in RT4 tumor cells (Fig. 6, middle column of RT4) as well as in UROtsa cells, seen as nuclear fragmentation and/or cell detachment. These processes appeared to take place more slowly in tumor cells, which reach their maximum damage 24 to 48 hr after PDT (cell loss, fragmented nuclei and/or loss of red lysosomal fluorescence), whereas normal cells exhibited a population with nuclear fragmentation and subsequent cell detachment only 4 hr after PDT (data not shown). No cell damage was seen at later time points, indicating complete tumor cell kill with minimized damage to normal urothelial cells as a result of PDT using 1 hr incubation. In contrast, 3 hr incubation led to necrotic cell death (Fig. 6, right column of each cell line); and signs of necrotic cell death (loss of homogenous green and punctate red fluorescence, cell swelling, plasma membrane disruption and release of cellular contents into surrounding medium), which appeared 2 to 4 hr after illumination (data not shown), could be seen 48 hr after PDT in both cell lines (Fig. 6, right column). At the end of our experimental period, complete tumor cell kill could be observed accompanied by a marked fraction of lethally damaged UROtsa cells.
Forty-eight hours after PDT, tumor cells are completely killed under each treatment modality examined, but different responses were observed in UROtsa cells. While short incubation times did not show significant cell damage in monolayers, long incubation times resulted in a marked population of dead cells.
Our results were obtained with established cell lines grown as plateau cultures. Previous experiments have shown that important aspects of the in vivo growth situation are mimicked by these cell lines and can partially be explained by differences in PPIX metabolism between normal and tumor cells.16 Especially in plateau culture state, the SV-40–transfected normal urothelial cell line is able to enter quiescence and to up-regulate iron and ferrochelatase, in comparison to its exponential growth state.16 In contrast, animal models, e.g., rats with chemically induced bladder tumors, do not show the marked difference in PPIX fluorescence and differential toxicity17 observed in human tissue and are of only limited use for therapeutic strategies in humans.
Data obtained with the human cell lines indicate that, besides absolute PPIX amount, the intracellular sensitizer distribution plays a decisive role in cellular phototoxicity and photodynamic efficacy. 5-ALA–induced cellular PPIX mediates phototoxicity via the generation of reactive oxygen species, i.e., singlet oxygen.18, 19 As a consequence, a correlation between the amount of sensitizer and phototoxicity is to be expected. This was confirmed within one cell line as well as between different cell lines in earlier experiments (data not shown). Furthermore, cellular damage should be restricted to the location of PPIX due to the short lifetime and range of the PDT mediator singlet oxygen.20, 21
In addition to different cellular PPIX concentrations caused by differences in PPIX metabolism,16 several authors have described the crucial role of intracellular sensitizer localization in photodynamic effectiveness.22–24 The most discussed PDT-sensitive cell organelles using various photosensitizers, e.g., 5-ALA–induced PPIX, methylene blue derivative (MBD) and others, are mitochondria.25–28 Mitochondrial photodamage leads to apoptotic cell death via release of mitochondrial content, i.e., cytochrome c. Cytochrome c mediates apoptosis via activation of downstream caspases in the apoptotic cascade, i.e., caspase-3.26, 29 In our experiments, we clearly demonstrated breakdown of mitochondrial activity in a significant fraction of photodynamically treated cells for all treatment modalities used (except for normal urothelial cells after 1 hr 5-ALA incubation and PDT). This suggests induction of apoptosis in irradiated cells via the mechanism mentioned above. However, only cells incubated with 5-ALA for 1 hr showed the typical signs of apoptosis (cell detachment, nuclear fragmentation) following white light illumination, whereas 3 hr incubation led to necrosis, characterized by early cell membrane damage and loss of cellular integrity. In accordance with other studies, we can therefore hypothesize that rapid necrosis can inhibit or delay induction of apoptosis, resulting in rapid necrotic cell death without typical morphological signs of apoptosis.22 Thus, we conclude that sensitizer localization in plasma membranes and regions close to membranes after 3 hr incubation leads to early loss of membrane integrity and results in necrosis following PDT.
The photosensitizer PPIX was localized similarly following 1 and 3 hr incubations in RT4 tumor cells, i.e., in cell membranes and regions next to membranes. Thus, similar phototoxic processes should take place and result in similar photodamage 48 hr after illumination. However, completely different pathways to cell death occurred after 1 and 3 hr incubations in RT4 cells. While 3 hr incubation led to clear necrosis, typical signs of apoptosis were observed following PDT after 1 hr incubation. These data indicate that inhibition of apoptosis by rapid initiation of necrosis occurs only if a critical level of membrane photodamage is reached. Otherwise, apoptotic processes triggered by mitochondrial photodamage can take place. However, complete tumor cell kill was achieved with both incubation and illumination conditions.
Significant differences in cell viability after PDT were seen in UROtsa cells using the 2 treatment modalities. While PDT after 3 hr incubation resulted in a marked fraction of dead cells, 1 hr incubation left most cells undisturbed. Furthermore, modes of cell death differed between the 2 treatment modalities, as seen in RT4 tumor cells. Shorter incubation intervals (1 hr) led to different localization patterns in UROtsa cells with lower PPIX content in or near cell membranes compared to 3 hr incubation. Together with lower cellular sensitizer content, this results in completely different modes of cell death. PDT after 3 hr incubation led to necrosis, as indicated by early loss of membrane integrity. PDT after 1 hr incubation led to reduced membrane photodamage in UROtsa cells caused by a lower ratio of PPIX content between cell membrane and intracellular compartments. Therefore, cell membrane integrity is preserved and cells can undergo apoptosis through release of cytochrome c from mitochondria. In consequence, damage to normal urothelial cells could be minimized or reduced to apoptosis, which is more gentle to surrounding tissue by avoiding extensive inflammatory reactions.
In conclusion, complete tumor cell kill could be achieved for both 1 and 3 hr incubation times, but significant differences in the extent of photodamage were seen in normal urothelial cells. These data may be used to enhance differential photodynamic effectiveness between tumor and normal cells in clinical PDT, thus avoiding negative side effects.30, 31 Our data also provide the basis for planning combination PDT and chemotherapy, as in preliminary in vitro studies,32 since different mechanisms of toxicity can be combined in a defined manner to optimize the ratio between toxicity and side effects.