Galactodendritic Phthalocyanine Targets Carbohydrate-Binding Proteins Enhancing Photodynamic Therapy

Photosensitizers (PSs) are of crucial importance in the effectiveness of photodynamic therapy (PDT) for cancer. Due to their high reactive oxygen species production and strong absorption in the wavelength range between 650 and 850 nm, where tissue light penetration is rather high, phthalocyanines (Pcs) have been studied as PSs of excellence. In this work, we report the evaluation of a phthalocyanine surrounded by a carbohydrate shell of sixteen galactose units distributed in a dendritic manner (PcGal16) as a new and efficient third generation PSs for PDT against two bladder cancer cell lines, HT-1376 and UM-UC-3. Here, we define the role of galacto-dendritic units in promoting the uptake of a Pc through interaction with GLUT1 and galectin-1. The photoactivation of PcGal16 induces cell death by generating oxidative stress. Although PDT with PcGal16 induces an increase on the activity of antioxidant enzymes immediately after PDT, bladder cancer cells are unable to recover from the PDT-induced damage effects for at least 72 h after treatment. PcGal16 co-localization with galectin-1 and GLUT1 and/or generation of oxidative stress after PcGal16 photoactivation induces changes in the levels of these proteins. Knockdown of galectin-1 and GLUT1, via small interfering RNA (siRNA), in bladder cancer cells decreases intracellular uptake and phototoxicity of PcGal16. The results reported herein show PcGal16 as a promising therapeutic agent for the treatment of bladder cancer, which is the fifth most common type of cancer with the highest rate of recurrence of any cancer.


Introduction
Conventional photodynamic therapy (PDT) combines a nontoxic photosensitizer (PS), light irradiation at a specific wavelength and tissue molecular oxygen to produce cytotoxic reactive oxygen species (ROS) [1,2]. The molecular mechanisms underlying PDT are not clearly understood. However, it has been described that the generation of ROS will trigger signalling pathways that ultimately destroy the targeted tissue. Cell death in PDT may occur by apoptotic and by non-apoptotic mechanisms (e.g. necrosis), or even by a combination of the two mechanisms [2]. Additionally, studies suggest that cell death pathway induced after PDT depends on the PS and its intracellular localization, the PDT dose and the cell metabolic potential (e.g. its intrinsic antioxidant capacity) [2]. To enhance the specific deliver/target of PSs in cancer cells, third generation PSs have been synthesized, by conjugating them with biochemical motifs [3][4][5]. Among new third generation PSs, the advances in the past years concerning glycobiology have spurred the development of carbohydrate-based molecules for cancer treatment by PDT [3,4,[6][7][8][9][10][11][12][13][14].
Carbohydrates have a strong potential as PS-delivery systems, because they are biocompatible molecules with a rapid cellular uptake and specific recognition by lectin proteins, which play an important role in several biochemical signalling pathways implicated in cancer metastasis, cell growth and inflammation [15,16]. The exact interaction mechanism of PS-carbohydrate conjugates with cancer cells is still unknown. However, it is expected that the specific (non-covalent) binding of carbohydrates with lectins [16], promotes the accumulation of the glyco-conjugate inside cells by the endocytic pathway. In addition, the expression of certain carbohydrate-binding lectins (e.g. galectins) is higher in cancer cells than in non-tumoral cells [17].
Recently, the emerging role of dendrimers (with well-defined nano-scaled structures) in biological systems has highlighted their potential benefits for the preparation of new anticancer drugs [31][32][33]. Regarding dendritic units of specific carbohydrates, it is wellknown their multivalent interactions with lectins, promoting a synergistic increase in binding affinity [31]. The photodynamic efficiency of porphyrins conjugated with glycodendrimers has been reported in the literature [12,[34][35][36][37]. However, the in vitro PDT studies with the corresponding phthalocyanines (Pcs) are scarce.
Recently, we have reported the synthesis of a new Pc decorated with sixteen molecules of galactose (in a dendritic manner, PcGal 16 , Figure S1) [34]. PcGal 16 demonstrated strong absorbance in the red spectral region (600-800 nm), fluorescence emission bands at 734 and 805 nm, solubility in a phosphate buffered saline (PBS) solution and interaction with human serum albumin [34]. Additionally, PcGal 16 demonstrated photostability and ability to generate ROS after photoactivation. The present study was undertaken to validate the in vitro photodynamic efficacy of this PcGal 16 from the standpoint of its uptake by bladder cancer cells (HT-1376 and UM-UC-3, derived from transitional cell carcinoma) to interaction with carbohydrate-binding proteins; induction of phototoxicity, ROS production and activity of antioxidant enzymes after PDT. Our findings show that PcGal 16 has a strong photodynamic efficiency in an in vitro system of bladder cancer.

Cells culture and treatments
Human bladder cancer cell lines UM-UC-3 and HT-1376 derived from high-grade transitional cell carcinoma (from the American Type Culture Collection, ATCC, Manassas, VA, USA) were cultured in Eagle's Minimum Essential Medium (EMEM; ATCC) supplemented with 10% (v/v) of fetal bovine serum (Life Technologies, Carlsbad, CA, USA), 100 U/mL penicillin, 100 mg/mL streptomycin and 0.25 mg/mL amphotericin B (Sigma).
UM-UC-3 and HT-1376 cells were seeded at a density of 3610 4 and 4610 5 cells/well in 96-and 6-well culture plates (Orange Scientific, Braine-l'Alleud, Belgium), respectively. Twenty-four hours after plating, cells were incubated with the desired concentrations of PSs in the dark for the indicated period of time.
Photodynamic irradiation was carried out in fresh culture medium, devoid of PS, covering UM-UC-3 and HT-1376 cell monolayers and exposing them to red light (620-750 nm) delivered by an illumination system (LC-122 LumaCare, London). The light was delivered for 10 min or 40 min at a fluence rate of 2.5 mW/cm 2 or 10 mW/cm 2 , as measured with an energy meter (Coherent FieldMaxII-Top) combined with a Coherent Power-Sens PS19Q energy sensor [34]. Sham-irradiated cells, used as controls, consisted in cells kept in the dark for the same durations and under the same environmental conditions as the irradiated cells. In all treatments, triplicate wells were established under each experimental condition, and each experiment was repeated at least three times.

Cellular uptake of PcGal 16
After incubation with PcGal 16 in the dark, UM-UC-3 and HT-1376 cells were immediately washed with PBS buffer and lysed in 1% m/v sodium dodecyl sulfate (SDS; Sigma) in PBS buffer at pH 7.0. PcGal 16 intracellular concentration was determined by spectrofluorimetry using an IVIS Lumina XR equipment (Caliper Life Sciences, Hopkinton MA) with excitation and emission wavelengths set at 675 nm and Cy 5.5 (695-770 nm), respectively, and the results were normalized for protein concentration (determined by bicinchoninic acid reagent; Pierce, Rockford, IL, USA).
For microscopic evaluation, UM-UC-3 and HT-1376 bladder cancer cells were grown for 24 h on glass coverslips coated with poly-L-lysine (Sigma). The cells were incubated with 5 mM PcGal 16 for 2 h, at 37uC. After incubation, cells were fixed with 4% paraformaldehyde (PFA; Merck, Darmstadt, Germany) for 10 min at room temperature. The samples were then rinsed in PBS, and mounted in VectaSHIELD mounting medium containing 49,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, CA, Burlingame) for visualization under a confocal microscope (LSM 510, Carl Zeiss, Gottingen, Germany). For detection of PcGal 16 , the specimen was excited at 633 nm and its emitted light was collected between 653-750 nm. For DAPI detection, specimen was excited at 405 nm and its emitted light was collected between 430-500 nm.

Cell metabolic activity and membrane integrity
Trypan Blue dye exclusion. Cell membrane integrity after PcGal 16 incubation in the dark, irradiation, or both was determined by the trypan blue dye (Biowhittaker, Walkersville, MD, USA) exclusion test 24, 48 and 72 h after each treatment. Cells with intact membrane were counted on a Neubauer chamber after trypsinization and the cell viability of treated cells was normalized to that of the untreated cells.
MTT assay. Cell metabolic activity after PcGal 16 incubation in the dark, irradiation, or both was determined 24, 48 and 72 h after treatments by measuring the ability of bladder cancer cells to reduce 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide (MTT, Sigma), to a colored formazan using a microplate reader (Synergy HT, Biotek, Winooski, VT, USA). The data were expressed in percentage of control (i.e. optical density of formazan from cells not exposed to PcGal 16 ).
IC 50 values (i.e. concentration of PcGal 16 required to reduce cell viability by 50% as compared to the control cells) were calculated using non-linear regression analysis to fit dose-response curves in GraphPad Prism 5.0 software (La Jolla, CA, USA).

Detection of intracellular Reactive Oxygen Species (ROS) generation
Immediately after irradiation or sham-irradiation, cancer cells were washed twice with PBS and incubated with either 2 or 5 mM 29,79-dichlorodihydrofluorescein diacetate (H 2 DCFDA; Invitrogen Life Technologies, Carlsbad, CA, USA) for an additional 1 h period, at 37uC, protected from light. After incubation, cells were washed with PBS and lysed in 1% (m/v) SDS solution in PBS (pH 7.0). DCF fluorescence was determined using a microtiter plate reader (Synergy HT) with the excitation and emission filters set at 485/20 nm and 528/20 nm, respectively. Protein concentration was determined using the Pierce BCA Protein Assay Kit.
The ROS levels were also qualitatively evaluated by fluorescence microscopy. After PDT treatments, UM-UC-3 and HT-1376 human bladder cancer cells grown on coverslips were incubated with 5 mM of H 2 DCFDA in PBS buffer (in dark conditions). After washing steps and fixation in 4% (m/v) PFA, coverslips were mounted using VectaSHIELD mounting medium and the slides were visualized under a confocal microscope (LSM 710, Carl Zeiss).

Redox quenching studies
Immediately after PcGal 16 uptake, photodynamic treatment was performed with cell monolayers covered with culture medium containing 50 nM of redox quenchers sodium azide, L-histidine and L-cysteine obtained from Sigma. The effect of quenchers on cell viability was evaluated 24 h after PDT by the MTT viability assay.

TUNEL assay
Cell death was detected by terminal deoxynucleotidyltransfer-asedUTP nick end-labeling (TUNEL) assay, using the DeadEnd Fluorometric TUNEL System (Promega, Madison, WI, USA), according to the manufacturer's instructions. Briefly, 24 and 72 h after PDT treatment, bladder cancer cells were fixed in 4% (m/v) PFA and permeabilized with 0.2% v/v Triton X-100 in PBS solution. Cells were stained with TdT reaction cocktail for 60 min at 37uC. The nuclei were stained with DAPI and the cells were analyzed under a fluorescence microscope (Leica DFC350 FX, Leica Microsystems, Bannockburn, IL, USA). Tunel-positive DAPI-stained cells were counted in 10 randomly selected fields from three independent experiments. Percentage of dead cells was expressed as ratio of TUNEL-positive cell numbers to DAPIstained cell numbers.

Antioxidant enzyme activities
Cell homogenates were obtained immediately after PDT and centrifuged at 10,000 g for 10 min at 4uC. The supernatants were used for measurements of glutathione peroxidase (GPox), glutathione reductase (GR), glutathione S-transferase (GST), superoxide dismutase (SOD) and catalase (CAT) activities in 96-well plates using a Biotek Synergy HT spectrophotometer (Biotek). The activity was expressed as nmol of substrate oxidized per minute per mg of protein (mU/mg).
GPox activity was determined at 30uC, measuring the NADPH (Merck) oxidation at 340 nm. Supernatants were mixed with 1 mM of glutathione-reduced form (GSH; Sigma), 0.5 U/mL GR (Sigma), 0.18 mM NaDPH, 1 mM EDTA (Sigma) and 0.7 mM tert-butyl hydroperoxide (t-BOOH; Sigma) in 50 mM imidazole (Sigma) at pH 7.4. The activity was calculated using the NADPH extinction coefficient of 0.62 m 2 /mmoL. GR activity in cell supernatants was determined at 30uC by measuring the rate of NADPH oxidation at 340 nm in the presence of 3 mM glutathione-oxidised form (Sigma), 0.12 mM NADPH, and 2.5 mM EDTA, in 50 mM Hepes (pH 7.4). The activity was calculated using the NADPH extinction coefficient of 0.62 m 2 /mmoL.
GST activity was determined at 30uC by monitoring the formation of GSH conjugate with 1-chloro-2,4-dinitrobenzene (CDNB; Sigma) at 340 nm in the presence of 1 mM GSH and 1 mM CDNB in 50 mM Hepes (pH 7.4). The activity was calculated using the conjugate extinction coefficient of 0.96 m 2 / mmoL. SOD activity was determined at 25uC measuring the cytochrome c (Merck) reduction at 550 nm. The supernatants were mixed with 40 mM cytochrome c solution (0.05 M potassium phosphate, 0.5 mM EDTA, pH 7.8) containing 80 mM xanthine (Merck). To initiate the reaction, 2 U/mL xanthine oxidase (Merck) was added. The increase in cytochrome c absorbance at 550 nm was recorded. SOD activity was calculated considering that one unit of SOD activity represents the inhibition of 50% in the rate of increase in absorbance at 550 nm when compared with control (sample without SOD under the conditions of the assay).
CAT activity was determined at 25uC by monitoring the rate of hydrogen peroxide (0.04% w/w) decomposition in 0.05 M potassium phosphate, pH 7.0. One unit of catalase activity was defined by the enzyme quantity that produced an absorbance reduction of 0.43 per minute at 240 nm in this system.

Transfection assays
Galectin-1 or GLUT1 was depleted in human bladder cancer cells using a pool of three target-specific 20-25 nt siRNA (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). UM-UC-3 and HT-1376 bladder cancer cells were transfected in 6-or 96-well culture plates, at 60-80% confluence, with galectin-1 and GLUT1, respectively. Cells were also transfected with a scrambled siRNA in parallel as controls.
For each transfection, cells were treated for 5 h with 2.4 mM of siRNA in transfection medium (Santa Cruz) containing 0.5 mL/ cm 2 of transfection reagent (Santa Cruz). After incubation, complete media was added and the cells were incubated for 24 or 48 h. Galectin-1 or GLUT1 downregulation was evaluated 24 h or 48 h post-transfection by Western blotting. The uptake and PDT experiments were performed 24 h or 48 h posttransfection with GLUT1 hsiRNA or galectin-1 hsiRNA, respectively.
Immunofluorescence UM-UC-3 and HT-1376 human bladder cancer cells were grown on coverslips as previously described [20,21]. After treatment, cells were washed with PBS and fixed in 4% PFA. Cells were then permeabilized with 1% Triton X-100 in PBS (pH 7.4) and blocked with 5% bovine serum albumin in PBS buffer, before incubation with primary antibodies rabbit antigalectin-1 1:100 (Abcam) and rabbit anti-GLUT1 1:250 (Chemicon). The cells were then rinsed with PBS buffer and incubated with DAPI and secondary fluorescent antibodies. After washing, samples were imaged using a confocal microscope (LSM 710, Carl Zeiss).

Statistical analysis
The results are presented as mean 6 standard deviation (S.D.) with n indicating the number of experiments. Statistical significance among two conditions was assessed using the nonparametric Mann-Whitney test. Statistical significance among three conditions was assessed by the nonparametric Kruskal-Wallis test. Statistical significance among several conditions was assessed with the Friedman test. P-value was considered at the 5% level of significance to deduce inference of the significance of the data. All graphs and statistics were prepared using the GraphPad Prism 5.0 software.

PcGal 16 accumulates in cancer cells and is non-toxic in darkness
To study the cellular uptake of PcGal 16 , HT-1376 and UM-UC-3 bladder cancer cells have been incubated with increasing concentrations (0.5, 2.5, 5 and 9 mM) of PcGal 16 in PBS for up to 4 h. PcGal 16 intracellular accumulation was determined by quantitative spectrofluorimetry and fluorescence microscopy. As shown in Figure 1A, the uptake of PcGal 16 was both concentration-and time-dependent, reaching a plateau in less than 2 h. Addition of 5 mM PcGal 16 to HT-1376 and UM-UC-3 cells resulted in an intracellular concentration of 35316125.9 and 29736119.1 nmol PcGal 16 per mg of protein, respectively, after 2 h of incubation ( Figure 1A). This spectrofluorimetric data was confirmed by confocal microscopy showing that cells treated with PcGal 16 exhibit strong fluorescence, with occasional bright spots in the perinuclear region ( Figure 1B). PcF 16 , the non-conjugated Pc ( Figure S1), was used as control. No significant intracellular accumulation was observed when the cells were incubated with 0.5-9 mM PcF 16 (data not shown), showing that the uptake of the PcGal 16 by cancer cells is enhanced relatively to unconjugated PcF 16 . After confirmation of PcGal 16 uptake by bladder cancer cells, its cytotoxic effect in darkness was assessed by the MTT colorimetric assay ( Figure S2). No dark toxicity was observed in untreated cells (up to 4 h) in the presence of 0.45% or less DMSO in the incubation medium. Moreover, PcGal 16 showed no significant cytotoxicity at concentrations up to 9 mM up to 72 h after treatment ( Figure S2).

PcGal 16 induces cytotoxicity after photodynamic activation
To test the effect of light irradiation (red light at 620-750 nm delivered at 2.5 mW/cm 2 for 40 min, i.e. 6 J/cm 2 ) after PcGal 16 uptake on cell viability, MTT was performed 24 h after treatment ( Figure 2). No cytotoxicity was observed in the untreated shamirradiated cells (Figure 2A) or untreated irradiated cells in the presence of 0.45% (v/v) or less DMSO in PBS (data not shown). However, when HT-1376 and UM-UC-3 cells were incubated with PcGal 16 and then irradiated, there was an increased phototoxicity in a concentration-and uptake time-dependent manner ( Figure 2A). Data showed that PcGal 16 exerted a higher phototoxicity on UM-UC-3 cells compared to HT-1376 cells ( Figure 2A). Moreover, the percentage of cell death in treated cells compared to untreated cells was significantly influenced by the dose of light ( Figure 2B). The phototoxicity was higher in cells irradiated at 6 J/cm 2 than in cells irradiated at 1.5 J/cm 2 (cells irradiated with light at 2.5 mW/cm 2 for 40 min or 10 min, respectively). On the other hand, irradiation of cells with light at 10 mW/cm 2 for 10 min (i.e. 6 J/cm 2 ) resulted in induction of cell death in untreated control cells. In subsequent experiments, we then performed cells irradiation with light at 2.5 mW/cm 2 for 40 min. Based on the uptake results ( Figure 1A) and MTT data before ( Figure S2) and after PcGal 16 photoactivation (Figures 2A  and 2B), we estimate the lowest concentration of PcGal 16 and the lowest dose of light necessary to achieve high phototoxicity for both bladder cancer cell lines. When cells were incubated with 5 mM PcGal 16 for 2 h and then irradiated with light at 6 J/cm 2 (cells irradiated for 40 min with light at 2.5 mW/cm 2 ), we observed a significant increase in phototoxicity of HT-1376 and UM-UC-3 cells. The cells were also incubated with 5 mM of PcF 16 during 2 h and then irradiated. As shown in Figures S2 and 2, the phototoxicity was higher for PcGal 16 than for non-conjugated PcF 16 . Based on the critical role of ROS in causing cell death after PDT and considering the different PDT-induced phototoxicity observed in UM-UC-3 and HT-1376 cells, the intracellular production of ROS was evaluated immediately after PDT in the cells previously incubated with 5 mM PcGal 16 for 2 h. The application of PcGal 16 in combination with PDT led to a high significant augmentation of ROS in both bladder cancer cell lines compared with the control (Figures 2C and 2D). The ROS levels (DCF fluorescence fold increase per mg of protein) in HT-1376 and UM-UC-3 cells were 50.52612.77 and 74.88611.49, respectively, when 5 mM H 2 DCFDA was used for ROS detection ( Figure 2D).
To assess the contribution of ROS in PcGal 16 -mediated cell death, quenchers of ROS (histidine, sodium azide [38] and cysteine [39]) were added at non-toxic concentrations to the incubation medium when the cells were irradiated. Cell viability evaluated 24 h after treatment was dependent on the used scavenger and cell type ( Figure 2E). For the cell line UM-UC-3, all quenchers at the employed concentration partially decrease the PcGal 16 -PDT-induced phototoxicity. For the cell line HT-1376, none of the quenchers used in these experiments were able to reduce the phototoxicity induced by photoactivated PcGal 16 .
To assess whether PDT has a long-term phototoxic effect, we evaluated cell viability for up to 72 h after PDT treatment. In both cell lines, the results obtained with the MTT colorimetric assay (cell metabolic activity) were correlated with the loss of cell membrane integrity (trypan blue staining) (Figures 3A and 3B). Overall, UM-UC-3 and HT-1376 bladder cancer cells were unable to recover from the PDT-induced damage effects 48 or 72 h after treatment, for PcGal 16 concentrations above 5 mM. TUNEL data revealed that there is an induction of cell death in a time-dependent manner in the cells irradiated after incubation with PcGal 16 ( Figure 3C). Twenty-four hours after PDT with PcGal 16 , the percentage of TUNEL positive cells in UM-UC-3 cell line was 1.8 higher than that of the HT-1376 cells, but after 72 h there was almost the same percentage of TUNEL-positive cells in both cell lines. The concentrations of PcGal 16 necessary to inhibit the metabolic activity of UM-UC-3 and HT-1376 bladder cancer cells in 50% can be estimated from Figure 3A. These values, named as ''photocytotoxic concentrations'' (IC 50 ) are reported in Table 1. Data show that 24 h after PDT, IC 50 value is lower for UM-UC-3 when compared with HT-1376 cells and similar for these cell lines 72 h after PDT.

PcGal 16 induces antioxidant enzyme response after photodynamic therapy
Considering the different levels of ROS produced in the two bladder cancer cell lines after PDT with PcGal 16 , we investigated (immediately after PDT) the involvement of specific antioxidant enzymes [40] in the detoxification of ROS and/or resulting toxic products. For that, the activities of the three major antioxidant enzymes, SOD, CAT, and GPox were determined by spectroscopy [41]. SOD catalyses the dismutation of superoxide radical anions into hydrogen peroxide and molecular oxygen. Hydrogen peroxide is then removed by CAT when it is present at high concentrations or by GPox when present at low concentrations. Knowing about the indirect antioxidant function [40] of GR in the replenishment of gluthathione levels in reduced form (GSH) and of GST in the elimination of reactive compounds through their conjugation with GSH, their activities were also determined.
In UM-UC-3 control cells, the activities of GR, SOD and CAT were 1.5-fold, 1.9-fold and 1.5-fold higher, respectively, than in HT-1376 control cells (Table 2). There was no significant difference in the activities of GST and GPox between the control cells of the two cell lines. After PDT with PcGal 16 , there was a 1. Knockdown of galectin-1 and GLUT1 decreases the uptake and phototoxicity of PcGal 16 We investigated whether the presence of the dendritic galactose units around the core of Pc molecule could facilitate the   interaction of this PS with specific domains in the plasma membrane of cancer cells. We hypothesized that domains enriched in carbohydrate-binding proteins [42] could facilitate the interaction with PcGal 16 , enhancing somehow its cellular uptake, and therefore its photodynamic potential.
Galectin [18] and glucose [19] proteins are expressed in high levels in cancer cells and both have affinity for galactose molecules. Therefore, we have evaluated the protein levels of galectin-1 and GLUT 1 in UM-UC-3 and HT-1376 cells, by Western Blotting and immunofluorescence (Figures 4 and 5).
The galectin-1 protein levels were higher in UM-UC-3 than in HT-1376 control cells ( Figure 4A). To determine whether galectin-1 plays a role in the uptake of PcGal 16 by cancer cells, siRNA was used to knockdown galectin-1 within UM-UC-3 bladder cancer cells. The treatment of UM-UC-3 cells with a pool of three target-specific siRNAs maximally suppressed galectin-1 by <50% at 24 h and 48 h post-transfection ( Figure 4B), without affecting the expression of the housekeeping protein b-actin. The transfected cells were then treated with PcGal 16 48 h posttransfection. As shown in Figures 4C and 4D, transfected cells displayed a markedly decreased uptake and phototoxicity of PcGal 16 . The GLUT1 protein levels were higher in HT-1376 than in UM-UC-3 control cells ( Figure 5A). Therefore, HT-1376 bladder cancer cells were also treated with a pool of three targetspecific GLUT1 siRNAs. Application of GLUT1 siRNA suppressed GLUT1 by <50% and <90% at 24 h and 48 h posttransfection, respectively ( Figure 5B). Treatment of HT-1376 cells with PcGal 16 twenty-four hours post-transfection, resulted in a substantial decrease in the uptake and phototoxicity ( Figures 5C  and D).

PcGal 16 decreases the galectin-1 and GLUT1 protein levels
To further explore the role of galectin-1 and GLUT1 in the photodynamic effect induced by PcGal 16 , we determined the levels of these proteins before and after PDT. Both incubation of cancer cells with PcGal 16 (i.e. incubation of cancer cells with PcGal 16 in darkness) and PDT with PcGal 16 induced a decrease in galectin-1 as observed by Western Blotting and immunofluorescence ( Figures 4E, 4F and 4G). The decrease observed in galectin-1 was higher in UM-UC-3 cells as compared to HT-1376 cells and it was more evident after PDT. Using confocal fluorescence microscopy, we observed co-localization of PcGal 16 with galectin-1 inside bladder cancer cells ( Figure 4G).
Similar to what was observed for galectin-1, there was also a decrease in GLUT1 ( Figures 5E, 5F and 5G) both after PcGal 16 uptake and after PDT treatment in HT-1376 cancer cells. Furthermore, in these cancer cells it was higher after PDT than after PcGal 16 uptake in darkness. In UM-UC-3 cells, PcGal 16 was not able to reduce GLUT1 protein levels ( Figure 5F). In both bladder cancer cell lines there was co-localization of PcGal 16 with GLUT1 ( Figure 5G). Overall, these findings clearly indicate show the critical involvement of the carbohydrate-binding proteins in the potential of PcGal 16 as a therapeutic agent.

Discussion
Third-generation PSs such as Pc coupled to carbohydrates are interesting for PDT, because they can be recognized by glycoprotein-based membrane proteins that are overexpressed in tumors [6]. Besides the enhancement of cellular recognition, the presence of dendritic galactose molecules improves Pc solubility and biocompatibility [34]. We have recently reported the synthesis of a new Pc with dendrimers of galactose sugar (PcGal 16 ) that has valuable spectroscopic and photochemical properties [34]. In this  Table 2. Values of activity (mU/mg of protein) of antioxidant enzymes superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GPox), glutathione reductase (GR) and glutathione S-transferase (GST) determined after PDT. study, we showed that PcGal 16 is a nontoxic compound per se, and has high photocytotoxic efficiency in two bladder cancer cell lines, which is paralleled with its high ability to produce ROS and to induce oxidative stress ( Figure 6). The high intracellular uptake of the glycoconjugated PS, PcGal 16 , can be explained by the presence of carbohydrate cellular transporters or receptors present at the cell surface. Although the PcGal 16 uptake was quite similar in the two bladder cancer cell lines, the expression of carbohydrate-binding proteins GLUT1 and galectin-1 is different amongst them. Besides its role in the import and export of glucose [19], the isoform of glucose transporter GLUT1 also transports D-galactose [19] having lower affinity for it than for D-glucose. Studies have been suggested that the hydroxyl groups in C1, C3 and C4 positions of D-galactose are hydrogen bond acceptors for GLUT1 sugar uptake site [43]. Like other galectins, galectin-1 has a carbohydrate-recognition domain (CRD) able to recognize and bind b-galactose [17]. Our assays demonstrated that galectin-1 and GLUT1 are both expressed by UM-UC-3 and HT-1376 cells. However, HT-1376 cells present higher GLUT1 levels compared with UM-UC-3 cells, and the contrary was observed for galectin-1. Although a similar PcGal 16 uptake was observed in the two bladder cancer cell lines, both GLUT1 and galectin-1 may contribute for its specificity modulating the intracellular uptake. Knockdown of galectin-1 and GLUT1 in UM-UC-3 and HT-1376 cells, respectively, was associated with a marked decrease of PcGal 16 uptake and phototoxicity. Together, these data demonstrated that galectin-1 and GLUT1 contribute for the efficacy of PDT mediated by PcGal 16 .
Interestingly, although the similar uptake of PcGal 16 by UM-UC-3 and HT-1376 cells, the phototoxicity induced 24 h after PDT was higher in UM-UC-3 cells than in HT-1376 cells. Such lack of association between uptake and phototoxicity has been described [44,45]. We investigated whether the higher phototoxicity observed in UM-UC-3 cells was due to higher production of ROS and/or higher oxidative damage compared with that in HT-1376 cells. As expected, the ability of PcGal 16 to produce ROS was higher in UM-UC-3 than in HT-1376 cells.
In PDT, it has been described that ROS can be generated by two photochemical reactions [46,47]. In type-II photochemical reactions, the excited PS in its triplet state can transfer its energy to molecular oxygen leading to the formation of singlet oxygen. Type-I photochemical reactions happen when an excited PS reacts with a biological substrate forming radicals and radical ions. Treatment with ROS quenchers demonstrated that in UM-UC-3 cells, singlet oxygen should have a high effect since cell death was highly reduced with quenchers of singlet oxygen (sodium azide and histidine). Further studies are needed to gain insight into the contribution of specific ROS in PcGal 16 -mediated cell death after PDT.
Interestingly, we observed that PDT with PcGal 16 has a longterm phototoxic effect in both cancer cell lines. Cytotoxicity assays (MTT, trypan blue and TUNEL assays) performed 72 h after PDT demonstrated that UM-UC-3 cells were not able to recover. Moreover, in HT-1376 cells there was a marked induction of cell death occurring from 24 to 72 h after PDT with PcGal 16 . The three distinct cytotoxic methods used in the present work are widely applied in the study of cell death: MTT (indicator of metabolic activity), trypan blue staining (indicator of membrane integrity loss occurring in necrosis or in late stages of apoptosis) and TUNEL assay (indicator of DNA fragmentation, a key factor of apoptosis). Cell death in PDT may occur by apoptosis or necrosis, or even by a combination of the two mechanisms [2]. A more specific and comprehensive study is needed to understand the specific cell death pathways induced after PDT with PcGal 16 in the bladder cancer cells used in this study. The different cell death obtained 24 h after PDT in UM-UC-3 and HT-1376 cells can be partially explained by the different amount of ROS present in both cells lines after irradiation. In addition, the resistance exhibited by HT-1376 cells could be due to the presence of efficient protective mechanisms, at least in the first stages after photodynamic treatment. Cytoprotective mechanisms initiated by cancer cells after PDT are well-known [47]. The increase of antioxidant molecules (e.g. gluthathione, vitamin C and vitamin E) [48] and the induction of genes encoding proteins involved in apoptosis or in the repair of lesions [49] are two of the well-known cytoprotective mechanisms induced after PDT. Another one is based on the equilibrium between photo-oxidative impairment of cells by ROS versus elimination of ROS by the activity of cellular antioxidant enzymes. Recent studies have shown that PDT caused increasedantioxidant enzymes activity and expression [50]. Thus, PDT efficacy can be influenced by the antioxidant response of the enzymes SOD, the GSH system and CAT. Our data demonstrated that after PDT with PcGal 16 there was an increase in the activity of SOD, CAT and GR antioxidant enzymes in both cell lines, being higher in HT-1376 than in UM-UC-3 cells. This higher antioxidant defense of HT-1376 cells against ROS can explain the results obtained 24 h after treatment. However, it is hypothesized that this was not maintained for 72 h after PDT since for this time point there was a massive cell death. This not only suggests that in this cell line there is a temporal relationship between ROS levels and cell death, but shows that antioxidant enzymes activity is of greater importance in protecting HT-1376 cells for at least 24 h after PDT with PcGal 16 . Regarding the activity of antioxidant enzymes, in HT-1376 cells it was also observed a decrease in the activity of GST, which is an enzyme implicated in cells defense against oxidation products. This enzyme has been described as protecting cells from DNA desintegration and drug toxicity [51]. GST isoforms are overexpressed in multidrug resistant tumors having an important role in tumors drug resistance by direct detoxification or inhibition of the MAP kinase pathway [51]. Thus, the higher cell death observed in HT-1376 cells 72 h after treatment can be also related with the activity of GST. A decrease in the activity of GST can be associated with DNA fragmentation and cell death 72 h after treatment.
Understanding the role of galactose moieties in the recognition of the PS by cancer cells may allow the investigation and development of more focused therapeutic strategies. Thus, we investigated whether PcGal 16 could be directly recognized by specific carbohydrate-binding proteins present at the plasma membrane. Consistently, the photoactivated PcGal 16 was shown to co-localize and reduce the levels of the plasma membrane proteins galectin-1 and GLUT1. Moreover, the immunofluorescence and Western Blotting studies demonstrated that, although its non-dark toxicity, PcGal 16 decreases the levels of galectin-1 and GLUT1 proteins. A plausible explanation for the decreased levels of the galactose binding proteins, galectin-1 and GLUT1, after incubation with PcGal 16 can be the masking of the epitope, which can block antibody-epitope binding due to changes in protein conformation or, eventually, endocytosis of these proteins and subsequent degradation. Thus, the changes observed in the levels of galectin-1 and GLUT1 could be induced directly by the binding of PcGal 16 to the carbohydrate-binding proteins and/or indirectly by the generation of ROS after PDT with PcGal 16 .
Although significant progress has been made in research related with the role of galectins in cancer, the information underlying the molecular mechanisms that control the expression of these proteins in tumour cells is scarce. The interaction of PSs with galectins (namely galectin-1 and galectin-3) has been studied by spectroscopic studies [52] and molecular modeling analysis [6,27]; however, they have not been validated by in vitro studies. As far as we know, there are no in vitro reports indicating whether PSs can modulate the expression of carbohydrate-binding proteins such as galectin-1 and GLUT1. Knowing that galectin-1 expression is correlated with cell metastatic potential [18,53] and contributes to tumor progression and resistance after conventional cancer therapeutic modalities [18], the ability of PcGal 16 to reduce the levels of galectin-1 after its uptake and/or photoactivation prompted us to envisage PcGal 16 as a potential candidate for cancer treatment.
Knowing that the overexpression of GLUTs is involved in tumor glycolysis -one of the biochemical ''hallmarks'' of cancerthe efficiency of PcGal 16 as an efficient anti-cancer PS is also evidenced by its ability to reduce GLUT1. GLUT1 is an attractive target to consider in the development of new PSs because it is lower expressed in normal-epithelial tissues or benign epithelial cell tumors when compared with human cancer cells [54]. The function of GLUT1 in the tumorogenesis process has been demonstrated by in vitro and in vivo studies, where the overexpression of GLUT1 antisense resulted in the inhibition of HL60 leukaemia cells proliferation and MKN-45 derived xenografs, respectively [55,56]. Based on the results of the current study, we envisage PcGal 16 as a promising therapeutic agent for the treatment of bladder cancer. Further studies are warranted to investigate the selectivity and photototoxicity of this PS in an in vivo model of bladder cancer, to contribute to a possible impact on clinical practice.