In Vitro Comparison of Hypericin and 5-Aminolevulinic Acid-Derived Protoporphyrin IX for Photodynamic Inactivation of Medulloblastoma Cells

Background Hypericin (HYP) is a naturally occurring photosensitizer. Cellular uptake and photodynamic inactivation after incubation with this photosensitizer have neither been examined in medulloblastoma cells in vitro, nor compared with 5-aminolevulinic acid-derived protoporphyrin IX (5-ALA-derived PpIX). Methods In 3 medulloblastoma cell lines (D283 Med, Daoy, and D341 Med) the time- and concentration-dependent intracellular accumulation of HYP and 5-ALA-derived PpIX was analyzed by fluorescence microscopy (FM) and FACS. Photocytotoxicity was measured after illumination at 595 nm (HYP) and 635 nm (5-ALA-derived PpIX) in D283 Med cells and compared to U373 MG glioma cells. Results All medulloblastoma cell lines exhibited concentration- and time-dependent uptake of HYP. Incubation with HYP up to 10 µM resulted in a rapid increase in fluorescence intensity, which peaked between 2 and 4 hours. 5-ALA-derived PpIX accumulation increased in D283 Med cells by 22% over baseline after 5-ALA incubation up to 1.2 mM. Photocytotoxicity of 5-ALA-derived PpIX was higher in D283 Med medulloblastoma compared to U373MG glioma. The [lethal dose (light dose that is required to reduce cell survival to 50% of control)] of 5-ALA-derived PpIX was 3.8 J/cm2 in D283 Med cells versus 5.7 J/cm2 in U373MG glioma cells. Photocytotoxicity of HYP in D283 Med cells was determined at 2.5 µM after an incubation time of 2 h and an illumination wavelength of 595 nm. The value was 0.47 J/cm2. Conclusion By its 5-fold increase in fluorescence over autofluorescence levels HYP has excellent properties for tumor visualization in medulloblastomas. The high photocytotoxicity of HYP, compared to 5-ALA-derived PpIX, is convincingly demonstrated by its 8- to 13-fold lower . Therefore HYP might be a promising molecule for intraoperative visualization and photodynamic treatment of medulloblastomas.


Introduction
The treatment of medulloblastomas is based on surgery and radiochemotherapy [1]. Medulloblastomas often infiltrate extremely vulnerable adjacent areas, such as the brainstem and fourth ventricle. Thus, the surgeon often faces a dilemma during resection. Knowing that the extent of resection determines the response to therapy and survival time, surgeons favor extensive or complete resection of the tumor [2]. Yet, severe and often persistent neurological deficits caused by injuries to adjacent brain structures, such as cranial nerve nuclei in the brain stem may encounter. A method of visualizing the tumor intraoperatively might improve surgical outcomes.
The introduction of fluorescence-guided resection of glioblastomas using 5-aminolevulinic acid-derived protoporphyrin IX (5-ALA-derived PpIX) has improved radicality of tumor resections and, consequently, patient outcomes [3][4][5]. Fluorescence of malignant tissue after administration of 5-ALA is caused by the accumulation of its metabolite PpIX, which is involved in the last step of heme biosynthesis -insertion of ferrous iron into PpIX by ferrochelatase forming heme. In malignant tissue, this step is impeded resulting in the accumulation of PpIX [6]. Based on results in human cell culture studies, in which glioblastoma cell lines showed significantly higher fluorescence intensity after incubation with 5-ALA as compared with neurons, we became interested in whether medulloblastoma cells also accumulates PpIX.
Our recent studies have focused on hypericin (HYP), a naturally occurring photosensitizer (PS) that is found in plants of the genus Hypericum, of which Hypericum perforatum (St. John's wort) is the most common. HYP is a lipophilic molecule that is incorporated into the phospholipid bilayer of cell membranes and has already in the dark versatile pharmacological activities. These include antiviral, anticancer and antiangiogenic properties [7][8][9][10][11]. Takahashi et al. could show an inhibitory effect on proteinkinase C, which is involved in cell proliferation [12,13]. Malignant gliomas have, compared to glial cells, a high proteinkinase C activity [14]. HYP has excellent properties as a PS [15][16][17][18][19]. Photodynamic therapy (PDT) is based on the generation of reactive oxygen species (ROS -either superoxide radicals, type I reaction, or singlet oxygen molecules ( 1 O 2 ), type II reaction) by a photosensitizer (PS) in the presence of oxygen and light. HYP has a high triplet quantum yield and a high efficiency in the formation of ROS [16,17,20]. The excessive production of ROS leads to oxidative stress to many biomolecules, e.g. proteins, causing cell death by induction of apoptosis, necrosis or autophagy associated cell death [21]. Semiquinone anion radicals are also important reactive intermediates in HYP mediated PDT. The site of photodamage by semiquinone anions is the quinone reducing center (Q i ) off complex III [22]. In vivo, PDT causes vascular damage and effects tumor destruction additionally [23]. Efficacy of PDT in vitro depends on the uptake of photosensitizer and fluence rate [24]. Stylli et al. suggested that the intracellular concentration of photosensitizers in tumor tissue is a prognostic factor of therapeutic outcomes [25]. Thus, this parameter must be examined in vitro to develop therapeutic strategies for optimal treatment conditions.
In malignant glioma, the most frequent malignant brain tumor in adults, the value of HYP in tumor visualization and photodynamic therapy in vitro has been demonstrated in detail [26]. By this HYP may also be of great interest in the therapy of medulloblastomas. Based on the experience in glioblastoma, we performed this study to: Examine the time-and concentration-dependent fluorescence of medulloblastomas after administration of 5-ALA and HYP. II. Compare fluorescence microscopy (FM) and FACS with regard to the efficacy of measuring uptake kinetics of 5-ALA and HYP in vitro. III. Investigate the cytotoxicity and the photocytotoxicity as a function of photosensitizer concentration of 5-ALA and HYP in dependence of the light dose.
The fluorescence microscope was equipped with a PLAPO60x oil immersion objective (N.A. 1.4) and the following filter sets were used: to measure 5-ALA-derived PpIX fluorescence: band pass filter 370-390 nm (excitation), dichroic mirror 440 nm, long pass filter 455 nm (emission) -AHF Analysentechnik AG, Tübingen, Germany; to measure HYP fluorescence (U-MWG2 filter -Olympus, Hamburg, Germany): band pass filter 510-550 nm (excitation), dichroic mirror 570 nm, long pass filter 570 nm (emission). Images were obtained using a charge-coupled device camera (F-View II, Olympus Soft Imaging Solutions GmbH, Münster, Germany). Fluorescence intensity was assessed in 20 individual cells by the pixel grey scale intensities of single cells with the software analySIS (Soft Imaging Systems, Münster, Germany). For flow cytometric measurements cells were cultivated in 24well plates (NUNC) and incubated with 5-ALA or HYP. After rinsing with PBS (500 ml) cells were detached from the plates with Accutase TM (PAA, Cölbe, Germany). Cells from 4 identically treated wells were pooled and filtered through a metal mesh (100-112-mm pore size; Haver & Boecker, Oelde, Germany). After addition of propidium iodide (Sigma-Aldrich, Steinheim, Germany) dissolved in PBS (2 mg/ml; 10 ml per 1 ml cell suspension) fluorescence of HYP and 5-ALA treated cells was measured by flow cytometry. Fluorescence was excited at 488 nm. Photosensitizer fluorescence was detected at an emission wavelength of 585+/221 nm. To determine cell viability an emission wavelength of .670 nm was used. The data were analyzed with CellQuestTM (version 4.0.2, Becton Dickinson, Erembodegem-Aalst, Belgium) on 200,000 cells.

Cytotoxicity
To measure cytotoxicity (toxicity in the dark) of 5-ALA/5-ALAderived PpIX and HYP, cells were grown in 24-well plates (NUNC). 3610 4 cells/well were seeded and incubated with different PS concentrations (5-ALA: 0-15 mM; HYP: 0-20 mM) at 37uC (5% CO 2 ) for 2 h. After being rinsed with PBS cells were incubated with PS-free culture medium for 48 h (37uC/5% CO 2 ). Thereafter, culture media were discarded and cells were rinsed twice with isotonic saline. Cell survival was assessed by neutral red (NR) assay. Cultures were thus incubated (37uC) with an aqueous NR solution (Biochrom, Berlin, Germany) diluted with culture medium 1:25 resulting in final NR concentration of 0.012% and adjusted to pH 7.4) for 2-3 h (by addition of 500 ml NR solution to each well). Supernatants were discarded and cells were rinsed twice with isotone saline. NR was extracted from the cells by addition of 150 ml of a mixture of water/ethanol/acetic acid (1/1/ 0.02). After gentle agitation for 10 min, a 100 mL aliquot of each sample was transferred into a microtiter plate (Greiner, Frick-enhausen, Germany). Absorbance was measured at 570 nm (with a reference wavelength of 690 nm) on a plate reader (Lucy 1, anthos Mikrosysteme, Krefeld, Germany) and analyzed with the anthos WinRead software (version 2.3) [26][27][28]. Cell survival was calculated as the percentage of living cells incubated with the PS versus living cells of non-incubated controls. For each cell line, at least 3 independent experiments were performed in quadruplicate, resulting in at least 12 values for each case.

Photocytotoxicity
To measure the photocytotoxicity of 5-ALA-derived PpIX and HYP, cells were cultured in 4-well plates and incubated with the PS at non-cytotoxic concentrations for 2 h (5-ALA: 0-6 mM; HYP: 2.5 mM). After being rinsed with PBS and incubated with culture medium (without PS), the cells were illuminated with a dye laser (model 375, Spectra Physics, Mountain View, USA) that was pumped by an argon ion laser (model 2030, Spectra Physics).
Cells were irradiated under sterile conditions with energy densities (light doses) of 0 to 10 J/cm 2 at 635 nm (5-ALA-derived PpIX) or 0 to 1.5 J/cm 2 at 595 nm (HYP) through the cover plate of the 4-well plates. Irradiation was performed with the bare fibre used for laser light transmission. Light was transmitted through an optical diaphragm in order to ensure irradiation with the central part of the laser beam. Exposure times were varied to obtain the various light doses at constant power density (fluence rate) of 20 mW/cm 2 (5-ALA-derived PpIX) and 10 mW/cm 2 (HYP). All illumination experiments were monitored with a power meter (model TPM-310, Gentec, Quebec, Canada).
Cell survival was assessed photometrically (NR assay as described above) after an additional growth period of 48 h (37uC, 5% CO 2 ). Survival rates were calculated as percentages of the living cells after incubation and treatment by PDT compared to the non-illuminated controls. To compare the response of the different cell lines towards PDT and to compare efficacy of both

Statistical Analysis
For mathematical description of the concentration-and timedependent HYP and 5-ALA-derived PpIX uptake special compartmental models as described in more general detail in Jacquez were used [29].
The parameters were fitted by the nonlinear least square method after logarithmic transformation of the mean observations.
The  (2) for c = 0. Fluorescence intensity is described by model (3a) with b~0 at constant temperature. For temperature change at time T * , the initial increase in fluorescence intensity at low temperatures is described by FI(t)~m 1 t [a.u.].
For times after the temperature switch at time T Ã , the formula becomes.
The initial slope for 37uC m 0 where D 0 [mM] is the minimal incubation concentration, a 1 (dimensionless) determines the slope at the point of inflection, a 2 is the asymptotic D 50 [mM] as c I reaches infinity, and a 3 [mM 2 ] determines the approach to the asymptotic D 50 . The 95% confidence interval was calculated, and the cell lines were compared by z-test.

Accumulation of 5-ALA-derived PpIX
Concentration-dependent 5-ALA-derived PpIX fluorescence. PpIX fluorescence was measured by fluorescence microscopy in D283 Med medulloblastoma cells incubated with 5-ALA at concentrations between 0 and 6 mM for 2 h. Autofluorescence of the cells was assessed using non-incubated controls. Intracellular PpIX fluorescence increased from 204 to 250 a.u., i.e. by about 22% compared with autofluorescence levels, after incubation with 1.2 mM 5-ALA. However, incubation with higher 5-ALA concentrations (up to 6 mM) resulted only in a further 2% increase in PpIX accumulation (see Figure 1).

Accumulation of HYP
Concentration-dependent uptake by FM. Concentrationdependent uptake of HYP was examined in D283 Med, Daoy, and D341 Med cells by fluorescence microscopy. Cells were incubated at HYP concentrations between 0 and 20 mM for 2 h. Figure 2 shows the intracellular accumulation of HYP fitted by Formula (1). For all cell lines, fluorescence intensity increased almost linearly up to 2.5 mM HYP. However, fluorescence increase was less pronounced at higher concentrations. D341 Med and Daoy cells exhibited nearly identical fluorescence levels, whereas that of D283 Med cells was significantly higher. Figure 3A shows the intracellular HYP accumulation in three independent experiments of D283 Med incubated with different HYP concentrations by FM. Addionally, concentration-dependent HYP uptake was measured also by FACS in D283 Med cells under identical conditions as used in the FM experiment (Fig. 3B). Results obtained with both methods were fitted by Formula (1).    HYP fluorescence increased almost linearly up to incubation concentrations of 2,5 mM, whereas fluorescence increase was less pronounced at the higher concentrations (up to 20 mM). As depicted in Figure 3C,  Time-dependent HYP uptake. Time-dependent fluorescence intensity measurements of HYP uptake by FACS and FM are shown together in Figure 4. D283 Med cells were incubated for up to 6 hours with HYP at 2.5, 5, and 10 mM. For an incubation concentration of 10 mM, HYP uptake peaked at 2 h and 4 h for FM and FACS, respectively. For lower HYP concentrations (2.5 and 5 mM) a steady increase of cellular fluorescence was observed.
The parameter a, which describes the uptake rate, was estimated for the FACS and FM analyses (Table 1). After Bonferroni adjustment of p-values, there was no significant difference between the two methods at concentrations of 5 and 10 mM (Table 1).

Temperature Dependent Cellular Uptake
In order to elucidate, whether HYP uptake is an active, energydependent process, D283 Med cells were incubated with HYP (2.5 mM) at two different temperatures of 4uC and 37uC. Cellular HYP accumulation was almost negligible within the incubation period at 4uC (m 1 = 1.1). After raising the temperature to 37uC, HYP uptake started to increase (m 2 = 9.2). However, intracellular HYP fluorescence did not reach the same uptake rate (m 0~1 3:9) as observed when cells were incubated at 37uC for the entire period ( Figure 5). Table 2 summarizes the initial slopes and their confidence intervals for the two conditions. There is a significant difference between the rate of uptake at constant temperature and at delayed temperature increase: The rate of uptake at 37uC for constant temperature is 834 [a.u./h] and 552 [a.u./h] for delayed temperature rise (p,0.0001).
Photocytotoxicity 5-ALA-derived PpIX. Photocytotoxicity of 5-ALA-derived PpIX was determined in D283 Med cells at up to 6 mM 5-ALA (2h incubation for each) and an illumination wavelength of 635 nm. The light doses in all experiments were 0 to 10 J/cm 2 . As shown in Figure 6A, cell inactivation after incubation with 5-ALA concentrations higher than 0.5 mM depends mainly on the light dose. PDT with a concentration of 0.25 mM 5-ALA or less was dependent on light dose and 5-ALA concentration.
Under the same conditions, PDT was performed in U373 MG glioblastoma cells, incubated with 0 to 1 mM 5-ALA ( Figure 6B). In comparison to D283 Med cells, cell inactivation after incubation with 5-ALA concentrations higher than 0.25 mM was mainly dependent on the light dose. To compare 5-ALAderived PpIX photocytotoxicity in the different cell lines, LD 50 values were calculated according to Formula (4).  (Table 3) after incubation with 0.5 and 1 mM 5-ALA, i.e. medulloblastoma cells were more sensitive to PDT.
Photocytotoxicity of HYP. The photocytotoxicity of HYP in D283 Med cells was determined at a noncytotoxic concentration (2.5 mM), an incubation time of 2 h, and an illumination wavelength of 595 nm; the light doses were 0 to 1 J/cm 2 . The LD 50 value of HYP was 0.47 J/cm 2 (95% CI 0.426-0.515). Cell survival was fitted by Formula (5) as depicted in Figure 7.

Discussion
To improve the prognosis of medulloblastoma patients, new therapeutic strategies are necessary. In treating medulloblastomas, response to adjuvant therapy is amended by increasing the extension of resection. Thus, intraoperative visualization might be a suitable option. 5-ALA-derived PpIX has been successfully applied for the intraoperative visualization of malignant brain tumors improving the prognosis of patients [4].
HYP, a naturally occurring photosensitizer, has many advantages compared to other photosensitizers, e.g. 5-ALA. Recent in vitro and in vivo studies have demonstrated high efficacy of HYP in PDT for various tumors [26,[30][31][32][33]. Additionally, HYP inhibits cell proliferation and signal transduction, also in the dark [34,35]. No data on the visualization and PDT of medulloblastomas using HYP exist, prompting us to investigate this and compare it with 5-ALA-derived PpIX.

Accumulation of 5-ALA-derived PpIX
By FM, D283 Med cells showed a concentration-dependent accumulation of PpIX. However, fluorescence increase was about 20% as compared to autofluorescence after an incubation period of 2 h with concentrations up to 1.2 mM 5-ALA. Other groups have reported disparate results with regard to PpIX fluorescence after incubation with 5-ALA in vitro. Stummer incubated C6 glioma cells in vitro with 1 mM 5-ALA, observing a linear rise in PpIX fluorescence from 5 to 85 min by fluorescence spectroscopy [36]. Carre et al. noted increasing PpIX synthesis in C6 glioma cells but reported high intercell variability by confocal laser scanning microspectrofluorometry, exciting cell samples at 488 nm [37].
Great differences were seen in ex vivo glioblastomas after in vitro incubation with 5-ALA. PpIX synthesis ranged from undetectable to high [38]. Moan [41]. According to Amo et al. PDT-induced cell death seems to occur predominantly via apoptosis through 5-ALA induced PpIX in mitochondria [42].
The small amount of PpIX accumulation in medulloblastoma cells may be explained by the large nucleus to cytoplasm ratio that may limit the primarily perinuclear production of PpIX [41,43]. Another explanation for the lack of increase in fluorescence intensity might be the generation of non-fluorescent aggregates. According to Schneckenburger et al., porphyrins tend to form aggregates with poor fluorescence in solution as well as within cells [44]. However, the presented data demonstrates that PpIX production after 5-ALA incubation is quite comparable in medulloblastoma and glioblastoma cells.

HYP Uptake in Medulloblastomas
All 3 medulloblastoma cell lines showed time-and concentration-dependent HYP accumulation, based on fluorescence intensity measurements by FM and FACS. High incubation concentration of HYP (10 mM) led to a rapid increase in fluorescence within 2 h. For longer incubation times up to 6 h, HYP fluorescence declined by about 15% at high HYP concentrations. Incubation with lower HYP concentrations (e.g. 2.5 mM) reached saturation not until 6 hours, which agrees with findings from our earlier studies in glioblastoma cell lines [26,45]. Similarly to the presented data, Ali et al. examined HYP uptake in nasopharyngeal carcinoma cells (CNE2 and TW0-1) incubating them with 1.25 mM (CNE2) and 2.5 mM HYP (TW0-1). HYP fluorescence rose during the first 2 to 3 h and decreased after 4 h [26,46]. This phenomenon might be due to the aggregation of PS molecules at higher concentrations, resulting in lower fluorescence [44]. Further on the accumulation of HYP in the tumor, compared to normal tissue, depends whether HYP is dissolved in a solution or in coarse aggregates. HYP soluted in a mixture of polyethylene glycol, DMSO and water showed a superior tumour to tissue relation compared to HYP suspended as coarse aggregates [47]. Not only the incubation concentration but also the uptake time has some influence to the HYP aggregation and by this to the phototoxicity of HYP [48]. This is not only due to aggregation processes but also to dynamics of the intracellular traffic [49].
So far, the mechanism(s) by which HYP is taken up is unknown. The presented data demonstrate a temperature-dependent uptake of HYP in a medulloblastoma cell line. Quite similar results have been obtained previously for glioblastoma cells [26]. With this experiment it is not possible to investigate the transport mechanism(s) of HYP in detail, i.e. to distinguish between micropinocytosis or receptor-mediated transport mechanisms. However, it is appropriate to distinguish whether HYP is predominantly taken up by an active, energy-dependent process or by passive diffusion (energy-independent). Analyzing fluorescence images, Uzdensky et al. found no difference in HYP staining patterns in WiDr cells (human colon carcinoma cell line) after a 1 h incubation at 4uC or room temperature, concluding that the uptake occurred by diffusion [50]. However, the procedure of the experiment has not been described in detail. Active uptake of HYP has also been observed in Caco-2 cell monolayers, a model that is used to study intestinal absorption. Transcellular transport of HYP at 37uC was significantly higher compared with that at 4uC [51].
Photodynamic Inactivation 5-ALA-derived PpIX showed photodynamic inactivation of U373 MG and D283 Med cells upon irradiation at 635 nm with light doses up to 10 J/cm 2 after 2 h incubation with 5-ALA at a non-cytotoxic concentration (0.5 to 6 mM). The light dose that was needed to inactivate 50% of cells, also named LD 50 value, was significantly lower in the medulloblastoma cell line as compared to the glioblastoma cell line applying incubation concentrations of 0.5 and 1 mM 5-ALA. However, photodynamic inactivation of D283 Med cells (LD 50 44.6 J/cm 2 ) was less efficient as compared to U373 MG cells (LD 50 12.6 J/cm 2 ) at a lower 5-ALA incubation concentration (0.1 and 0.25 mM). Riesenberg et al. also observed disparate responses of several human bladder carcinoma cell lines in vitro to PDT with 0.15 and 0.6 mM 5-ALA and light doses ranging from 15 to 100 J/cm 2 [52].
In D283 Med cells, HYP exhibited high photocytotoxicity. HYP-PDT of D283 Med cells (2 h/2.5 mM) irradiated at 595 nm with light doses of 1 J/cm 2 resulted in .90% cell death and an LD 50 value of 0.47 J/cm 2 was calculated. For glioblastoma cell lines (U373 MG, LN229, and T98G), LD 50 values varied between 0.15 and 0.22 J/cm 2 after 2h-incubation with 2.5 mM HYP as reported previously [26]. These results demonstrated HYP is much more efficient for photodynamic treatment of medulloblastomas and gliomas in vitro as compared to 5-ALA-derived PpIX (LD 50 of 6.5 J/cm 2 at 0.25 mM). The high photodynamic efficacy of HYP is most likely due to its high triplet quantum yield and efficient induction of ROS [16,17,53]. High efficacy of HYP-PDT may at least compensate the lower penetration depth of 595 nm light into tissue as compared to the most effective wavelength used in 5-ALA-PDT (635 nm). However, more clarity can be gained only by in vivo experiments. For this purpose, orthotopic in vivo models are indispensable. Recently reported in vivo models, for instance the topic application of HYP and 5-ALA in a skin tumor model [54], are not sufficient to predict the therapeutic effects in brain tumors, where the blood brain barrier has also to be considered.

Conclusions
Medulloblastoma cell lines exhibited significant time-and concentration-dependent HYP accumulation (5-fold over autofluorescence levels) and HYP is probably taken up by an energy dependent mechanism. The high photocytotoxicity of HYP as compared to 5-ALA-derived PpIX (8-13-fold lower LD 50 values) is notable. Therefore, HYP might be a promising molecule for intraoperative visualization and photodynamic treatment of medulloblastomas. In case that these promising in vitro data are confirmed by in vivo experiments, HYP has the potential to widen the armamentarium of medulloblastoma treatment options in primary treatment and especially as an alternative treatment option in cases of tumor relapse after extensive radiation and chemotherapy.