|Home | About | Journals | Submit | Contact Us | Français|
The successful implementation of photodynamic therapy (PDT)-based regimens depends on an improved understanding of the dosimetric and biological factors that govern therapeutic variability. Here, the kinetics of tumor destruction and regrowth are characterized by systematically varying benzoporphyrin derivative (BPD)-light combinations to achieve fixed PDT doses (M × J/cm2). Three endpoints were used to evaluate treatment response: 1.) Viability evaluated every 24 hours for 5 days post-PDT; 2.) Photobleaching assessed immediately post-PDT; and 3.) Caspase-3 activation determined 24-hours post-PDT. The specific BPD-light parameters used to construct a given PDT dose significantly impact not only acute cytotoxic efficacy, but also treatment durability. For each dose, PDT with 0.25 μM BPD produces the most significant and sustained reduction in normalized viability compared to 1 μM and 10 μM BPD. Percent photobleaching correlates with normalized viability for a range of PDT doses achieved within BPD concentrations. To produce a cytotoxic response with 10 μM BPD that is comparable to 0.25 μM and 1 μM BPD a reduction in irradiance from 150 mW/cm2 to 0.5 mW/cm2 is required. Activated caspase-3 does not correlate with normalized viability. The parameter-dependent durability of outcomes within fixed PDT doses provides opportunities for treatment customization and improved therapeutic planning.
Photodynamic therapy (PDT) is a light-mediated modality that involves activation of a photosensitizer (PS) to generate singlet oxygen and radical species that are toxic to cells. PDT is approved in the United States for ophthalmologic, dermatologic and oncologic applications (1–4) and is in clinical and pre-clinical evaluation for numerous cancers (3, 5–7) including intraperitoneal sarcomatoses and carcinomatoses (8–14). An array of parameters influence PDT outcomes such as PS concentration, fluence, irradiance, and the time course of cytotoxic response (15–17). These parameters provide important treatment-related challenges and opportunities that are also influenced by the therapeutic target site (3, 15–28). A disease and PS-specific characterization of outcomes that result from modulating treatment-related parameters within the framework of fixed PDT doses is necessary to provide an improved understanding of potential reasons for variability in therapeutic response.
Efforts by us (18, 19, 29–38) and others (6, 16, 23, 24, 39–46) have provided insights into better predicting PDT-related outcomes through improved dosimetry and modulation of treatment protocols to account for heterogeneities in PS levels and the biological characteristics of target tissue. Zhou et al. (19) have reported improved treatment consistency by adjusting fluences to compensate for variations in PS concentrations to achieve a fixed PDT dose (number of photons absorbed per unit tissue). This approach resulted in more consistent changes in tumor volume compared to the group that did not receive fixed PDT doses with a clinically-approved liposomal formulation of benzoporphyrin derivative monoacid ring-A (BPD) in an orthotopic model for prostate cancer (19). However, there is significant evidence that within a PDT dose, reciprocity only holds across limited PS concentrations and is highly dependent on treatment-related parameters and the characteristics of the disease site (19, 28, 47, 48). Patterson and colleagues have recently demonstrated (24) that singlet oxygen generation, photobleaching, and cell kill efficiency decrease significantly during PDT with high intracellular concentrations of BPD. The amount of oxygen consumed per photon absorbed decreased with increasing BPD concentration and singlet oxygen dose could only be correlated with photobleaching over a limited range of BPD concentrations under aerated conditions (24). These studies were conducted with tumor cells in suspension and evaluated acute PDT efficacy with the goal of improving dosimetry and treatment customization. A clinical study by Betz et al. (49) assessed the therapeutic effect of m-tetrahydroxyphenylchlorin (mTHPC)-based PDT in patients with basal-cell carcinoma and showed that low PS concentrations coupled with high fluences were effective in achieving high complete response rates with the added benefit of decreased general photosensitivity. The importance of considering the biological characteristics of the target disease is highlighted in a study published by Moesta et al. (28), which shows different rates of photofrin photobleaching and cytotoxic efficacy between two pancreatic cancer cell-lines at a given PDT dose. In addition, a lack of reciprocity across a broad range of PS-light combinations was observed (28).
An additional critical determinant of PDT efficacy is irradiance, which has been shown by us and others to significantly impact treatment outcomes (18, 21, 25, 33, 50). Gibson et al. were among the first to show increased tumor destruction with lower irradiances (51). Subsequent studies by Foster and colleagues have demonstrated that consideration of irradiance is important to improving PDT efficacy (21, 52). Using 3D spheroids of EMT6/Ro mouse mammary carcinoma cells, the authors showed that irradiances of 25 and 50 mW/cm2 provided better efficacy of Photofrin-based PDT compared to an irradiance of 200 mW/cm2 at a fluence of 60 J/cm2 (21). Similar results from our group and others have been reported in vivo with BPD-PDT (33) and aluminum pthalocyanine (AlPcS2)-based PDT (25). These findings collectively highlight the need for PS and target tissue-specific characterization of outcomes that result from modulating treatment-related parameters within the framework of fixed PDT doses.
The present study evaluates the effect of systematically modulating PS-light parameters in a three-dimensional (3D) model for micrometastatic ovarian cancer (OvCa) previously established in our group (53, 54). OvCa is the fifth most common cancer among women in the United States and causes more deaths than any other gynecologic malignancy (55–57). Despite significant advancements in surgical techniques and improvements in the medical management of advanced stage OvCa, survival rates remain dismal (55–57). There is a need to identify new mechanistically distinct therapeutic strategies that complement conventional treatments to improve outcomes from this lethal disease.
PDT is mechanistically-distinct from conventional therapies and is being evaluated for the treatment of an array of solid tumors including malignant intracranial cancers (58–63), head and neck cancers (64–66), recurrent prostate adenocarcinoma (67), bladder cancer (68), cholangiocarcinoma (69), pancreatic and bile duct cancers (70–73), as well as a variety of intraperitoneal carcinomatoses and sarcomatoses, including disseminated OvCa (8–14). Depending on the choice and concentration of PS, and other treatment-related parameters, a range of cell death pathways can be triggered by PDT including necrosis, apoptosis and autophagy (74–79). The mechanisms that trigger the various death pathways and molecular responses to PDT are complex and beyond the scope of this article. We refer the reader to a few, among many, elegant reviews and articles on this topic (3, 15, 17, 26, 74–77, 79–84). For BPD, the mitochondrion is a preferential, but not exclusive, site of sub-cellular localization (85–92). Light-mediated activation of BPD, can cause photodamage to the anti-apoptotic protein bcl-2 as well as the mitochondrion (20, 79, 86, 88, 93, 94). Depending on the PS concentration, target cells, PDT dose, and other biological and therapeutic variables, this photodamage triggers a rapid release of cytochrome c and the activation of a cascade of caspases (75, 86–88, 90–92). These caspases, such as the “effector” or “executioner” caspase 3, catalyze the hydrolytic reactions of apoptosis (75, 86–88, 90–92). This example of the potency and mechanistic breadth by which BPD-PDT confers cytotoxicity is illustrative of the factors that make PDT an important therapeutic tool in the development of rationally-designed treatment strategies for many diseases including metastatic OvCa. Leveraging these mechanistic assets, preclinical studies have shown that PDT reverses chemoresistance, synergistically enhances the efficacy of traditional chemotherapeutics and biologics, and reduces chemotherapy cycles (54, 95, 96). PDT disrupts and reduces the size of 3D ovarian micronodules, which addresses some of the key barriers to improved treatment efficacy in OvCa (53, 54). With these promising insights, it remains unclear which PDT parameters (e.g. PDT dose, PS-light combinations, and irradiances) will provide the most significant and sustained cytotoxic response in the treatment of multifocal tumors. An array of research tools and perspectives will be required to address the complexities of therapeutic variability.
Building on the elegant work of others to improve the predictability of PDT outcomes (6, 7, 15, 16, 24, 28, 75, 79, 97–102) as well as efforts by us (18, 19, 29, 30, 103), here the effects of modulating dose-related parameters on the extent and durability of PDT efficacy is assessed. A 3D model that restores important biological and architectural cues for adherent micrometastatic OvCa is used (Figure 1, left) (53, 54). A matrix of PS-light combinations that constitute three fixed PDT doses (Figure 1, middle) provides the framework to evaluate a few key determinants of cytotoxic response (Figure 1, right) among the many factors that warrant investigation. The effect of modulating PDT parameters is evaluated by administering BPD concentrations of 0.25 μM, 1 μM, or 10 μM (Figure 1, middle, blue) and appropriate fluences (0.125 J/cm2 – 40 J/cm2 at 150 mW/cm2) (Figure 1, middle, red) to achieve three fixed PDT doses (1.25, 5, or 10 μM × J/cm2) (Figure 1, middle, purple). Cytotoxic efficacy is evaluated every 24 hours for 5 days post-PDT using viability as a metric for therapeutic efficacy (Figure 1, right). Photobleaching is assessed immediately post-PDT as a surrogate measurement of the photodynamic events that produce concentration-dependent cytotoxic outcomes within fixed PDT doses. The impact of modulating irradiance on PDT efficacy is evaluated using viability 24 hours post-PDT as a cytotoxic endpoint. Additionally, activation of caspase-3 as a late stage apoptosis marker, is assessed 24 hours post-PDT.
Human ovarian carcinoma cells NIH:OVCAR5 (OVCAR5) were obtained from Fox Chase Cancer Center (Philadelphia, PA, USA), where they were characterized by microsatellite marker analysis. The cells were grown in RPMI 1640 (Roswell Park Memorial Institute) medium (Mediatech Inc., Herndon, Virginia, USA) and supplemented with 10% heat inactivated fetal calf serum (GIBCO Life Technologies, Grand Island, New York, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin. Growth factor reduced (GFR) Matrigel (Catalog number 354230, BD Biosciences, San Jose, California, USA) was used as a basement membrane in 3D cultures.
GFR Matrigel beds were prepared by pipetting 250 μL of GFR Matrigel solution at −4 °C on chilled 24-well plates. The plates were incubated for 30 minutes at 37 °C to allow for gelation. OVCAR5 cells were plated on the prepared beds on ice in black-walled 24-well plates at a density of 7500 cells in 1 mL of 2% GFR Matrigel medium. The 3D Ovarian cultures were then incubated at 37°C in an atmosphere of 5% CO2 for 10 days. The 3D culture was maintained with a 2% GFR-Matrigel RPMI 1640 solution and was changed every 3 to 4 days.
Ten days following plating, cultures were incubated with 0.25 μM, 1 μM and 10 μM BPD (QLT, Inc., Vancouver, British Columbia, Canada) in complete culture media for 90 min. The media were replaced with 2% GFR-Matrigel medium immediately prior to irradiation. Each well was irradiated with a 690-nm fiber-coupled diode laser (Model 7401; High Power Devices, Inc., North Brunswick, NJ, USA) using irradiances ranging from of 0.5–150 mW/cm2, which were measured using an Ophir Vega power meter with photodiode sensor (Ophir Optronics, Israel) prior to each experiment. Three-dimensional nodules were treated with total PDT doses ([PS] × fluence) of 1.25, 5, and 10 μM × J/cm2 at each BPD concentration (0.25 μM, 1 μM and 10 μM ). Irradiation times were calculated automatically by using custom-written IGOR scripts (Wavemetrics, Lake Oswego, OR, USA) and laser irradiations were controlled by a custom-built shutter system.
Photobleaching was evaluated in 10 day old 3D ovarian cancer nodules incubated with 0.25 μM, 1 μM, or 10 μM BPD for 90 minutes in complete growth media. Briefly, images of BPD fluorescence were acquired pre and immediately post-PDT using previously established settings (104). Percent photobleaching was estimated from the following formula as previously described (104):
OVCAR5 3D micronodules were sectioned using an HM 550 cryostat (Richard-Allen Scientific) as previously described (105). Samples were sectioned to a thickness of 35 μm and stained with DAPI (Sigma) to visualize cell nuclei. Images were acquired 24 hours after sectioning with the Olympus FV-1000 confocal microscope at an objective magnification of 40x (NA=0.75).
Cytotoxic response to PDT treatments on 3D cultures was assessed by an imaging-based methodology as previously described (106, 107). Briefly, cultures were stained, in situ, with calcein AM and ethidium bromide (to label live and dead cells respectively) prior to imaging using an Olympus FV-1000 confocal microscope. Images were batch-processed in MATLAB using a custom routine to segment the fluorescence channels and determine pixel intensities for the Calcein (live) and Ethidium bromide (dead) channels separately averaged within individual nodules or across the entire field of view. Subsequently, individual nodule viabilities and overall nodule viability for the viewing field was calculated from the mean intensities as reported by the normalized, and appropriately scaled ratio of fluorescent signals (calcein divided the sum of calcein plus ethidium bromide) as extensively described in previous publications (53, 54, 108–110).
Twenty-four hours following PDT treatment, 3D cultures in 24-well plate were carefully washed 3 times with PBS and then 1 ml of MatriSperse Cell Recovery Solution (BD Biosciences) was added into each well and incubated on ice for 30 min to dissolve the Matrigel bed and gently release the tumor acini. Detached tumor acini were collected and washed twice in 1 ml of fresh cold MatriSperse to remove the residue Matrigel. Pelleted acini were dissolved with protein extraction buffer provided in the ELISA kit for cleaved caspase-3 (Cell Signaling) by following the manufacturer’s suggested protocol. Ten μl of the soluble protein fraction were used for ELISA measurement of cleaved caspase-3, and the results were normalized to the total protein concentration of each sample, and plotted as fold-increase relative to no treatment controls. Values represent averages of duplicates for each group from 2 platings.
Comparisons of cytotoxic efficacy and BPD fluorescence were performed by one-way ANOVA or two-tailed t-test, as appropriate. Unlabeled comparisons indicate a two-tailed t-test. Linear regressions were evaluated by a least squares fit approach and correlation coefficients were determined using Pearson’s product-moment correlation method. Error bars indicate s.e.m. A P value of less than 0.05 was considered significant.
To assess the impact on cytotoxic response of modulating PS-light parameters within fixed PDT doses, OVCAR5 3D nodules were incubated with either 0.25 μM BPD (blue), 1 μM BPD (red), or 10 μM BPD (green) and treated with appropriate fluences to achieve three PDT doses: 1.25, 5, and 10 μM × J/cm2 (Figure 1). Relative to no treatment (black), the most significant and most sustained reduction in normalized viability was seen in nodules treated with 0.25 μM BPD-PDT compared to 1 μM and 10 μM BPD-PDT for all PDT doses (one-way ANOVA) (Figure 2A). The lowest viabilities over all PDT doses and all concentrations were observed 72 hours (0.09 ± 0.01) and 96 hours (0.10 ± 0.01) post-PDT in nodules treated with 0.25 μM BPD-PDT at a dose of 10 μm × J/cm2 (p<0.05; one-way ANOVA). These viabilities at 72 and 96 hours were not significantly different from each other. At a dose of 5 μm × J/cm2, viabilities of nodules treated with 0.25 μM BPD-PDT were significantly lower (0.16 ± 0.02 and 0.19 ± 0.02, respectively) than those treated with 1 μM BPD-PDT (0.30 ± 0.01 and 0.45 ± 0.04, respectively). At a PDT dose of 1.25 μm × J/cm2, normalized viabilities of nodules treated with 0.25 μM or 1 μM BPD-PDT were not significantly different for the first 72 hours post-PDT. Nodules treated with 10 μM BPD-PDT consistently showed the poorest response, with normalized viabilities that were significantly greater than 0.25 and 1 μM BPD-PDT for all doses across all time points (one-way ANOVA).
Figure 2B shows representative LIVE/DEAD images of OvCa 3D micronodules that were used to generate the values for normalized viability in Figure 2A. Images depict nodules evaluated 72 hours following PDT using different PS-light parameters to achieve a fixed dose of 10 μM × J/cm2. Relative to no treatment controls (1), increased killing is observed in nodules treated with 0.25 μM BPD-PDT (4) compared to either 1 μM BPD-PDT (3) or 10 μM BPD-PDT (2), as evidenced by increased ethidium bromide fluorescence (red) and decreased calcein fluorescence (green).
Photobleaching was measured in nodules treated using 0.25 μM (blue), 1 μM (red) or 10 μM (green) BPD at PDT doses of 1.25 μM × J/cm2 (circled), and 5 and 10 μM × J/cm2 (second and third points, respectively, on each correlation line within a concentration). Percent photobleaching (evaluated immediately post-PDT) correlates with normalized viabilities across PDT doses (assessed 24 hours post-PDT), within each BPD concentration (Figure 3). The goodness of fit values (r2) for linear regressions within each BPD concentration were 0.9962 for 0.25 μM BPD (blue), 0.9332 for 1 μM BPD (red), and 0.9982 for 10 μM BPD (green) (error bars indicate s.e.m. N≥4 wells for each data point). The p value for all correlations was less than 0.05. Representative unadjusted images of BPD fluorescence used to generate Figure 3 are shown in Supplemental Figure 1.
To evaluate concentration-dependent BPD distribution and uptake in 3D micronodules, 10 day old cultures were incubated with 0.25 μM, 1 μM and 10 μM BPD for 90 minutes, then cryosectioned, and imaged (Figure 4). (A) Representative fluorescence images of BPD (red) and DAPI (blue) along with corresponding BPD fluorescence intensity profiles (B) show that BPD is distributed throughout tumor micronodules at all concentrations: 0.25 μM (blue), 1 μM (red), and 10 μM (green). (C) BPD fluorescence intensity increased in a concentration-dependent manner. Mean fluorescence intensity was significantly lower at a concentration of 0.25 μM (2.07 × 105 ± 7.04 × 104)(blue), compared to 1 μM (1.23 × 106 ± 2.06 × 105)(red) and 10 μM BPD (1.46 × 107 ± 2.50 × 106) (green) (p<0.05, one-way ANOVA, n=5 for each concentration, error bars indicate s.e.m.). As shown in representative images in Supplemental Figure 2, BPD fluorescence was evenly distributed throughout the 3D nodules following PDT, irrespective of the PS concentration. Mean BPD fluorescence intensities varied significantly post-PDT (n=4 nodules for 0.25 μM BPD and n=6 for 10 μM BPD).
Informed by previously published findings (18, 21, 25, 33, 50), the impact of modulating irradiance on PDT efficacy was assessed in 3D OvCa nodules (Figure 5). A fixed PDT dose of 1.25 μM × J/cm2 was administered using 10 μM BPD (green) and irradiances ranging from 150 mW/cm2 to 0.5 mW/cm2. Normalized viability was significantly higher at an irradiance of 150 mW/cm2 (0.93 ± 0.01) than all lower irradiances (p<0.05, one-way ANOVA, N=6). An irradiance of 0.5 mW/cm2 was necessary to achieve a reduction in normalized viability with 10 μM BPD-PDT (0.69 ± 0.02) to levels that were not significantly different from 0.25 μM (0.63 ± 0.03)(blue) and 1 μM BPD-PDT (0.73 ± 0.03)(red). This reduction in normalized viability at the lower irradiance required a substantial increase in the irradiation time from approximately 1 second at 150 mW/cm2 to 250 seconds at 0.5 mW/cm2. An asymptotic exponential growth model was used to fit viability and irradiance for nodules treated with 10 μM BPD-PDT (r2=0.8411).
There was no relationship between activated caspase-3 levels and PDT dose for concentrations of 0.25 μM (blue) and 1 μM BPD (red) (Figure 6). A monotonic increase in caspase-3 activation was observed with increasing PDT dose using 10 μM BPD (green) (r2 = 0.995, P<0.05). Relative to no treatment controls, activated caspase-3 levels increased in all PDT treated groups. The most significant increase in activated caspase-3 levels was observed in the group treated with a PDT dose of 10 μM × J/cm2 BPD-PDT using 10 μM BPD (mean fold increase = 22.5 ± 2.0) (P<0.05, one-way ANOVA). Among the group treated with a PDT dose of 5 μM × J/cm2, the most significant increase in activated caspase-3 was seen in nodules treated with 1 μM BPD (mean fold increase = 20.7 ±1.3) (P<0.05, one-way ANOVA). For the groups treated with the lowest PDT dose, both 0.25 μM and 1 μM BPD induced the highest increases in activated caspase-3 (mean fold increase = 7.0 ±0.18 and 7.8 ±0.37, respectively) compared to 10 μM BPD (mean fold increase = 4.1±0.43).
Pre-clinical (54, 96, 111–118) and clinical (9–12, 119, 120) evidence suggests that PDT is a promising modality for the treatment of intraperitoneal malignancies, including disseminated OvCa. Due to the diverse biology of advanced stage disease and the narrow therapeutic index associated with treating complex sites such as the peritoneal cavity, PDT will most likely be part of a rationally-designed, multifaceted treatment plan (15, 81, 82, 118, 119). As with any treatment, the successful implementation of PDT-based regimens depends in part on improved dosimetric tools (6, 19, 24, 29, 39, 41, 97, 99, 112, 121, 122) along with a better understanding of the biological characteristics of the target tissue that govern optimal therapeutic response (28, 74, 75, 79, 101, 123, 124). Here, a systematic analysis of the tumoricidal effects that result from modulating PDT parameters is conducted to longitudinally characterize outcomes that may be relevant to developing comprehensive treatment strategies in the management of multifocal OvCa (53, 54, 125–129). The present study focuses on the following therapeutic and biological parameters among the many that warrant consideration: (i) the time-dependent evolution of cytotoxicity and regrowth resulting from systematically varying PS-light combinations to construct fixed PDT doses; (ii) the correlation between viability and photobleaching as a potentially relevant tool for implicit dosimetry; (iii) the effect of modulating irradiance on tumor viability; and (iv) the relationship between cytotoxicity and the activation of a marker for apoptotic cell death, caspase-3. Our findings demonstrate that the kinetics of tumor destruction and regrowth are significantly impacted by the parameters used to construct a given PDT dose. This study builds on efforts by us (18, 19, 29–38) and others (6, 16, 23, 24, 39–46) to improve the predictability of PDT outcomes through a better understanding of the biological and dosimetric factors that contribute to variability in treatment response.
PDT dosimetry is a complex and active area of study that can be approached from several perspectives, including implicit and explicit measurement of treatment-related variables, as described by Wilson and colleagues (24, 97, 122). Explicit dosimetry requires direct measurement of multiple independent photodynamic parameters that influence efficacy, such as local PS concentration, delivered light dose, and reactive oxygen species (24, 97). However, obtaining real-time measurements of each, or a combination of these parameters can be cumbersome, and reliable techniques and instrumentation are being investigated (40, 42, 130, 131). Implicit dosimetry may be a more feasible approach because it involves quantification of more readily accessible dosimetric parameters, such as PS photobleaching, which may be indicative of the photodynamic events that determine therapeutic efficacy (24, 97, 112, 121, 122). A study published by Ascencio et al. demonstrates that protoporphyrin IX photobleaching correlates strongly (r2=0.89) with necrotic score after intraperitoneal hexaminolevulinate (HAL)-based PDT of OvCa in rats (112). Dysart et al. have shown that photobleaching following mTHPC-based PDT can predict the viability of murine erythroleukemic cells in suspension (at PS concentrations below 2 μg/ml) (121). Weston et al. demonstrated that an implicit photobleaching-based metric correlates well with the viability of rat prostate adenocarcinoma cells in suspension after BPD-PDT (24). This correlation was true under oxygenated conditions for BPD concentrations of approximately 0.1 μM, 0.7 μM, and 3.5 μM (24). The study by Weston et al.(24) and the present study could only establish a correlation between cytotoxic efficacy and BPD photobleaching within limited ranges of BPD concentrations. Within the limited dose ranges evaluated in the present study (Figure 3), the rate of change in viability relative to percent photobleaching may be highest for 10 μM BPD-PDT compared to 0.25 and 1 μM BPD-PDT. The slopes for these correlations were not significantly different across BPD concentrations in this small, discrete dataset. It is worth noting that for 10 μM BPD-PDT the total change in viability and percent photobleaching occurred over a limited range, and additional studies are necessary to evaluate the scope of these relationships. These results are supported by the observation that photobleaching correlates with singlet oxygen production for limited ranges of BPD concentrations when the availability of local oxygen is not rate limiting (24).
Additional factors that may account for the concentration-dependent variability in therapeutic outcomes observed in the present study include PS aggregation, altered subcellular localization, differential photochemical mechanisms, microenvironmental influences, and self-shielding by the PS (20, 38, 44, 45, 100, 132, 133). Elegant studies by Aveline et al.(20, 38, 46, 134) have shown that the biological microenvironment can significantly impact PS photophysics and photochemistry. Local concentrations of PS and oxygen influence the photosensitizaion mechanisms of porphyrins in a cell. A reduction in the fluorescence and triplet state quantum yields is observed with increasing concentrations of many porphyrin PS, including BPD (38, 134). Foster and colleagues have demonstrated differential mechanisms of PS photobleaching that impact photodynamic dosimetry (44, 45, 135). Self-shielding is observed when sufficiently high PS concentrations prevent uniform light absorption (100). As shown in Supplemental Figure 2, BPD appears to be uniformly distributed across multiple cell layers following PDT treatment of 3D OvCa nodules at a dose of 5 μM × J/cm2 using either 0.25 μM (n=4) or 10 μM (n=6) BPD. These results from representative cryosectioned tumors per BPD concentration suggest that self-shielding at the size range evaluated here (~ 150 μm in diameter) may play only a small role in explaining the observed parameter-dependent discrepancies in cytotoxicity and photobleaching. Additional studies are necessary to characterize the potential role of concentration-dependent shielding that may be occurring at the subcellular level as well as in tumors that are outside the size range evaluated here.
In addition to these and other dosimetric approaches that may improve the predictability of treatment outcomes, an understanding of the molecular mechanisms that lead to cell death will be important considerations in PDT-related treatment planning. Caspase-3 is an executioner caspase that is activated by many death pathways and is thought to be a critical player in the events that lead to apoptosis (75, 101, 136). Oleinick and colleagues have shown differential kinetics and pathways for PDT-based destruction of breast cancer cells, dependent in part on the procaspase-3 status of the cells (101, 136). An increase in PDT-induced DNA fragmentation, among other hallmarks of apoptosis, were observed with increased expression of activated caspase-3 (136). Depending on the procaspase-3 status, there was also a differential sensitivity to PDT as evaluated by the WST-1 assay, which measures the ability of mitochondria to reduce a tetrazolium dye (136). However, the critical lethal events that compromise the ability of cells to divide and form colonies and ultimately lead to tumor destruction were independent of caspase-3 activation, both in monolayer and in vivo (75, 101, 136). These findings highlight the importance of understanding the metrics used to evaluate cytotoxic response within the context of the biological and microenviromental characteristics of the target tissue. The present study in a 3D OvCa model supports these findings, where no correlation was observed between caspase-3 activation and PDT dose (Figure 6) with 0.25 μM and 1 μM BPD-PDT. There was a correlation between PDT dose and caspase-3 activation with 10 μM BPD-PDT (Figure 6). Any potential relationship between caspase-3 activation and cytotoxicity seemed to be limited to low levels of cell killing with 10 μM BPD-PDT (Figure 2 and Figure 6). The nature and significance of this trend remains unclear and suggests a system-dependent activation of complex cell death pathways and survival processes that have been described previously (76, 90, 101, 136–138), and need to be explored further.
An understanding of the molecular characteristics of the target tissue may be important to determining the broader implications of the differential responses to PDT observed in a 3D tumor model using one tumor cell line. In contrast to the findings from the present study, Kessel and colleagues have shown that PDT with 10 μM BPD is significantly more effective than 1 μM BPD at reducing colony formation in a murine hepatoma cell line across a range of equivalent PDT doses (79). The authors suggest that in instances where autophagy serves as a protective mechanism against phototoxicity, the use of high concentrations of BPD could potentiate PDT efficacy (79). The basis for this idea is informed in part by a study from Donohue (139) et al. showing that BPD interferes with the sequestration and degradation of cytoplasmic material and inhibits the formation of autophagosomes in a concentration dependent manner. Osaki et al. have shown that BPD-PDT is differentially effective in four different types of rodent cell lines, due in part to variability in uptake and intracellular localization of the photosensitizer (140). A study by Moesta et al. (28) supports the importance of evaluating the effects of modulating PDT dose parameters in a manner that accounts for differences in the biological characteristics of the target tissue. Even within a given disease, variability between cell lines influences uptake and localization of the PS as well as the photophysical and photochemical events that determine overall response to treatment (28).
PDT has been shown to enhance the efficacy of traditional chemo- and biological-therapies, and should be included as part of rationally-designed combination treatment regimens (54, 95, 96, 141). Among the many factors that govern the successful development of PDT-based combinations, an understanding of the optimal dose and scheduling for the individual monotherapies based, in part, on the extent and durability of the cytotoxic response is required. A previous study from our laboratory has shown that PDT synergistically enhances carboplatin efficacy in a sequence-dependent manner (54). The synergistic regimen involved administration of carboplatin immediately after treatment with a PDT dose of 1.25 μM × J/cm2 delivered using 0.25 μM BPD and 5J/cm2 (54). The present study indicates that these same parameters confer the most significant reduction in the normalized viability, but the tumoricidal response continues to evolve for several days following treatment. It remains unclear whether the application of chemotherapy agents immediately after PDT or at the peak of PDT-induced cytotoxicity would maximize the cooperative interaction between the treatments. Additionally, weighing treatment-associated toxicity against increased efficacy in vivo will be an important consideration when determining the appropriate PDT dose-related parameters (142).
In the past few years, there has been an increasing focus towards the customization of targeted therapies informed by patient-specific molecular profiles of the target disease (143). Similarly, cytotoxic modalities such as PDT could benefit from therapeutic planning based on the identification of key molecular signatures that will guide the customization of treatment-related parameters. The present study demonstrates that using low concentrations of BPD and correspondingly high fluences to construct a PDT dose provides maximal tumoricidal efficacy and durability. It is important to note that while these results were assessed in a 3D OvCa model, other studies have identified differential PDT parameters that were optimal in other systems (22, 28, 79, 144). Percent photobleaching correlates with a reduction in normalized viability only within given BPD concentrations used to achieve a range of fixed PDT doses. These findings, supported by others (24, 97, 112, 121, 122), suggest that implicit dosimetry is a complex technique that must be informed in part by the specific parameters used to construct a given PDT dose. Building on the elegant work of others and efforts by us (18, 21, 25, 33, 50), the present study shows that low irradiance improves treatment efficacy for a PS-light combination that otherwise results in poor response. A correlation between PDT dose and activated caspase-3 was only observed with high BPD concentrations, where minimal killing was evident. Activated caspase-3 did not correlate with normalized viability across all BPD concentrations and time points evaluated, highlighting the complexity of death pathways that has been supported by other groups (76, 90, 101, 136–138). Collectively, these findings suggest that PDT delivery parameters could be customized based in part on a better understanding of the molecular signatures and biological characteristics of target sites.
At all PDT doses, 0.25 μM BPD photobleached significantly more than 1 and 10 μM BPD (p<0.05; one-way ANOVA). At a PDT dose of 1 μM × J/cm2, percent photobleaching of nodules treated with 0.25 μM BPD-PDT (13.9 ± 2.5%) was significantly higher than with nodules treated with 1 μM (0 ± 1.3%) and 10 μM BPD-PDT (−0.3 ± 1.8%) respectively (p<0.05; one-way ANOVA). This trend was also observed at a PDT dose of 5 μM × J/cm2 for 0.25 μM, 1.0 μM and 10 μM BPD (31.0 ± 4%, 10.7 ± 1.3% and 3.2 ± 1.8%, respectively) (p<0.05, one-way ANOVA) and 10 μM × J/cm2 (44 ± 5.8%, 23 ± 3.2% and 6 ± 1.3%, respectively) (p<0.05; one-way ANOVA) (N=8).
(a.) Representative fluorescence images (red, BPD; blue, DAPI) and (b.) BPD fluorescence intensity profiles of cryosections from 10 day-old 3D cultures treated show distribution of BPD throughout tumor micronodules incubated with 0.25 μM (blue), and 10 μM (green) BPD and treated with PDT at a dose of 5 μM × J/cm2 (n=4 nodules for 0.25 μM BPD and n=6 for 10 μM BPD).
Dr. Hasan and Dr. Celli wish to acknowledge support from the following grants from the National Cancer Institute at the National Institutes of Health: R01CA158415, R01CA160998 (TH) and K99CA155045 (JPC). Dr. Celli also acknowledges support from the Eleanor and Miles Shore Fellowship Program for Scholars in Medicine.