Graphene Oxide and Graphene Quantum Dots as Delivery Systems of Cationic Porphyrins: Photo-Antiproliferative Activity Evaluation towards T24 Human Bladder Cancer Cells

The development of new photodynamic therapy (PDT) agents designed for bladder cancer (BC) treatments is of utmost importance to prevent its recurrence and progression towards more invasive forms. Here, three different porphyrinic photosensitizers (PS) (TMPyP, Zn-TMPyP, and P1-C5) were non-covalently loaded onto graphene oxide (GO) or graphene quantum dots (GQDs) in a one-step process. The cytotoxic effects of the free PS and of the corresponding hybrids were compared upon blue (BL) and red-light (RL) exposure on T24 human BC cells. In addition, intracellular reactive oxygen species (ROS) and singlet oxygen generation were measured. TMPyP and Zn-TMPyP showed higher efficiency under BL (IC50: 0.42 and 0.22 μm, respectively), while P1-C5 was more active under RL (IC50: 0.14 μm). In general, these PS could induce apoptotic cell death through lysosomes damage. The in vitro photosensitizing activity of the PS was not compromised after their immobilization onto graphene-based nanomaterials, with Zn-TMPyP@GQDs being the most promising hybrid system under RL (IC50: 0.37 μg/mL). Overall, our data confirm that GO and GQDs may represent valid platforms for PS delivery, without altering their performance for PDT on BC cells.


Introduction
Bladder cancer (BC) is the tenth most diagnosed type of cancer worldwide, with a higher occurrence in men [1]. It commonly begins as a non-muscle invasive form in 70% of cases, and it can be managed by complete transurethral resection, following repeated instillation of bacillus Calmette-Guerìn (BCG) [2]. However, cases of recurrence are not uncommon, and treatment options are limited. Moreover, recurrence of BC can evolve towards a more invasive form that can rapidly affect the muscular tissues (muscle-invasive BC, MIBC) and spread into metastasis [3]. Successful management of BC requires the combination of multiple therapeutic approaches, including chemotherapy and radiotherapy. These treatments often involve several serious side effects that can affect the patient's life quality [4].

Synthesis of Zn-TMPyP
Zn-TMPyP was synthesized directly from the free-base TMPyP. The complexation was accomplished through the reaction of TMPyP (50.0 mg, 4.21 × 10 −2 mmol) with an excess of 1.5 equiv of Zn(II) acetate dihydrate (13.9 mg, 6.32 × 10 −2 mmol) in a solution of CHCl 3 :MeOH (3:1) at 50 • C for 1.5 h. After the terminus of the reaction, the porphyrin was crystallized firstly in a mixture of propanol and diethyl ether, while a second crystallization step was carried out in a mixture of MeOH and hexane (90% yield).

Characterization of Porphyrin@GO and Porphyrin@GQDs Hybrids
The abovementioned porphyrins (TMPyP, Zn-TMPyP, and P1-C 5 ) were non-covalently functionalized with GO (GO powder, Graphenea, San Sebastian, Spain) or GQDs (1.0 mg mL −1 dispersion of cyan luminescent GQDs in H 2 O, Merck, Darmstadt, Germany). The average thickness of the individual GO flakes was found to be 1.5-1.9 nm (3 layers) by atomic force microscopy (AFM), while for GQDs, the manufacturer indicated a 1-2.0 nm thickness (3-4 layers). UV-VIS absorption spectra were recorded on a Jasco V-560 spectrophotometer and a Varian Cary 50 spectrophotometer, using quartz cells (1 cm × 1 cm). Fluorescence spectra were recorded on Jasco FP-8300 and Varian Eclipse spectrofluorimeters, with excitation and emission slits set at 10 and 5 nm, respectively. Raman studies were performed in a combined Raman-AFM confocal microscope WITec alpha300 RAS+, through the excitation with a 532 nm line of a Nd:YAG laser (2 s, 10 acquisitions, 1 mW or 3 mW laser power for porphyrin@GO or porphyrin@GQDs, respectively). A Hitachi HD-2700 STEM microscope operating at 200 kV was employed to collect the STEM images on carbon-coated Cu grids.
The interactions between porphyrins and GO were monitored by spectrophotometric and spectrofluorimetric titrations at room temperature. Briefly, aliquots of a GO aqueous suspension (1.0 mg mL −1 or 3.0 mg mL −1 ) were successively added to an aqueous solution of porphyrin (2.0 µm or 6.0 µm). GO was added until three consecutive and similar measurement values of absorption red-shift and fluorescence were obtained, meaning that optimal interactions were achieved. In total, around 50 µL of GO was added to 1.0 mL of porphyrin. The synthesis of the porphyrin@GQDs hybrids was monitored oppositely, using the porphyrin solutions as a titrating agent (rather than the carbon nanomaterial), considering that GQDs have a stronger luminescence emission compared to porphyrins fluorescence. Several volumes of a porphyrin solution (stock 42 µm) were added to a solution of GQDs (0.1 µg mL −1 ). The fluorescence spectra of the hybrids were recorded at excitation at 330 nm, with 10 min delay between each addition. These spectrofluorometric titrations were further extended to other concentrations of porphyrin and carbon nanomaterials to allow the determination of the ideal binding ratio, which was used to prepare the treatment solutions for cellular assays. Porphyrins and graphene derivatives were mixed, considering the previously determined binding ratio, and placed on an orbital shaker for 30 min to allow the two components to interact. The solution was then diluted to a final 1.0 mg mL −1 concentration using PBS supplemented with 2% antibiotic/antimycotic solution and used to prepare the treatment solutions.

Light Exposure
Two irradiation devices were used for the irradiation of cell cultures during the assays: (i) a UVH436F medical lamp (Waldmann, Villingen-Schwenningen, Germany) was used for BL irradiation, with an irradiance of 17 mW cm −2 and a centered emission peak at 417 nm; (ii) a custom-build Oslon+ LED array system (Osram, Munich, Germany) was used to perform RL irradiation at an irradiance of 12 mW cm −2 . The selected LED array emitted light with a narrow band, centered at 625 ± 20 nm. BL and RL irradiances were measured using a Waldmann Variocontrol radiometer, equipped with blue_v and PDT1200 sensors.

Cell Cultures and Cell Treatment
A T24 human bladder cancer cell line (ATCC/LGC Standards, Milan, Italy) was maintained in minimum essential media (MEM) culture medium (Life Technologies, Monza, Italy), supplemented with 10% FBS, 1% alanyl-glutamine, 1% non-essential amino acids, and 1% antibiotic/antimycotic solution (Sigma-Aldrich, Milan, Italy) at 37 • C, in a humidified atmosphere containing 5% CO 2 . Cells (5 × 10 4 ) were seeded in 96-well plates in proliferation medium. After 24 h, cells were treated with increasing concentrations of non-immobilized porphyrins, porphyrin@GO, and porphyrins@GQD hybrids, in phenol red-free MEM containing 3% fetal bovine serum (FBS). After 1 h (non-immobilized porphyrins) and 4 h (porphyrin hybrids) of incubation, respectively, the medium was replaced with PBS and cells were treated with 2.5 J cm −2 of BL and 25.0 J cm −2 of RL. After irradiation, PBS was replaced with complete proliferation media. After 24 h, cell viability and cell death mechanisms were assessed. Dark control experiments were conducted, incubating cell cultures for 72 h with non-immobilized porphyrins or their hybrids with GO or GQDs, in complete proliferation media.

ROS and Singlet Oxygen Generation
The ROS assay was performed using the DCFDA-Cellular Reactive Oxygen Species Detection Assay Kit (Abcam, Cambridge, UK) according to the manufacturer's instructions.
Briefly, the cells (2.5 × 10 4 cells/well) were seeded into each well of a 96-well clear bottom black plate (Corning, New York, NY, USA) in the proliferation medium. After 24 h, cultures were incubated with 20 µm 2 ,7 -dichlorofluorescein diacetate (DCFDA) for 45 min and then treated with various concentrations of porphyrins/porphyrin hybrids for 1 h. Afterwards, the plate was irradiated with 2.5 J cm −2 of BL or 25.0 J cm −2 of RL. Data were collected using a Victor 3 multimodal microplate reader (λ exc /λ em = 485/535). Cultures treated with 100 and 200 µm tert-butyl hydrogen peroxide (TBHP) were taken as positive controls.
Singlet Oxygen was determined using Singlet Oxygen Sensor Green, using a protocol developed by Gollmer et al. [34]. Briefly, cells (2.5 × 10 4 cells/well) were seeded into each well of a 96-well clear bottom black plate (Corning, New York, NY, USA) in the proliferation medium. After 24 h, cells were washed with SMM (Standard Maintenance Medium, an aqueous solution of 140 mM NaCl, 3.5 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 10 mM glucose, and 10 mM HEPES, pH 7.35) and then incubated with 1 µm Singlet Oxygen Sensor Green (SOSG) for 2 h before being treated with various concentrations of porphyrins/porphyrin hybrids for 1 h. Afterwards, the plate was irradiated with 2.5 J cm −2 of BL or 25.0 J cm −2 of RL. Data were collected using a Victor 3 multimodal microplate reader (λ exc /λ em = 485/535).

Cell Viability and Mechanisms of Cellular Death
Cell viability was assessed by resazurin assay. Briefly, cells were incubated with resazurin for 4 h, and the fluorescence of resulting resorufin produced by healthy cells was measured using a Victor 3 multimodal microplate reader (Perkin Elmer, Waltham, MA, USA). Data are expressed as mean ± the standard deviation (SD) of the mean. The difference between groups was evaluated using analysis of variance (ANOVA) with Graphpad Prism 9.0. Mechanisms of cell death were investigated by flow cytometry (BD FACSCanto II, BD, East Rutherford, NJ, USA) and the BD Pharmingen FITC Annexin V Apoptosis Detection Kit I, following the manufacturer's indication. Briefly, cells were detached after 24 h after treatment, washed, and double-stained with Annexin V-FITC (AnnV-FITC, apoptosis marker) and 7-aminoactinomicyn D (7-AAD, necrotic marker) in the supplied binding buffer. Unmarked samples and samples individually marked with AnnV-FITC and 7-AAD were used for setting analysis parameters.

Intracellular Localization
T24 cells were grown on imaging dishes and, after 24 h, incubated at 37 • C with a final concentration of 10.0 µm of the three porphyrins and respective hybrids for 1 h and 4 h, respectively. Lysosomes were labeled using Lysotracker Green DND-26 (Life Technologies, Monza, Italy) for 15 min. Samples were imaged using an SP5 confocal laser scanning microscope (Leica, Wetzlar, Germany).

Synthesis and Characterization of the Tetracationic Porphyrins
The synthetic routes used to obtain the tetracationic porphyrins (TMPyP, Zn-TMPyP, and P1-C 5 ) are summarized in Schemes 1 and 2. These three cationic porphyrins were selected because they are readily internalized by cells in comparison to the non-cationic counterparts, and their positively charged structure allows non-covalent interactions to both GO and GQDs at physiological pH.
Access to the cationic porphyrin TMPyP and to the Zn(II) complex (Zn-TMPyP) required the previous preparation of the neutral 5,10,15,20-tetra(4-pyridyl)porphyrin (TPyP), which was obtained in 11% yield via acid-catalyzed condensation of pyrrole with 4-pyridinecarboxaldehyde, according to well-established literature procedures (Scheme 1) [29,35]. The quaternization of TPyP pyridyl units with iodomethane then afforded the desired free-base TMPyP in 70% yield. The Zn-TMPyP was obtained in 90% yield after the core metalation of the free-base with Zn(II) acetate.   The synthesis of the tetracationic dibenzoporphyrin P1-C5 (Scheme 2) was based on the procedure described by Wang et al. for preparing this type of π-extended porphyrins [31,32], but using the 5,10,15,20-tetraphenylporphyrin (TPP) as a scaffold. The tetrabromination of this porphyrin at opposite β-pyrrolic positions and the core metalation with Ni(II) was followed by a four-fold Heck reaction between Ni-β-Br 4 TPP and 4-vinylpyridine. This synthetic step was carried out in a mixture of N,N-dimethylformamide (DMF) and toluene, using Pd(OAc) 2 as a catalyst, PPh 3 as a ligand, and K 2 CO 3 as a base to afford the key intermediate Ni-P1 in 48% yield. The demetalation of this porphyrin was carried out in acidic conditions (in 52% yield), and then the quaternization of the pyridyl units from P1 was achieved with iodopentane to afford the final cationic porphyrin P1-C 5 in 74% yield.
All porphyrin derivatives were characterized by nuclear magnetic resonance (NMR) spectroscopy, electrospray ionization mass spectrometry (ESI-MS), and UV-VIS spectroscopy (see Materials and Methods section and Figures S1-S9 in Supplementary Materials). In particular, the UV-VIS spectra of the porphyrins in phosphate buffer saline solution (PBS) show an intense Soret band (S 0 →S 2 transitions), which peaked at 423 nm and 438 nm for TMPyP and Zn-TMPyP, respectively ( Figure 1a); moreover, the Q bands (S 0 →S 1 transitions) lie between 550 and 650 nm. The Soret band of the P1-C 5 porphyrin is red-shifted (to 465 nm) in comparison to TMPyP and Zn-TMPyP due to extensive electronic delocalization of the aromatic structure, which lowers the required energy for electronic transitions from the fundamental state to the excited states of the porphyrin. The molar  (Table 1), as well as at the wavelength emission of the two lamps used for the visible light irradiation in the biological experiments (417 nm for BL and 625 nm for RL). Furthermore, the three porphyrins show fluorescence emission maxima between 600 and 750 nm (Figure 1b). Scheme 2. Synthesis of P1-C5.

Characterization of Neat Carbon Nanomaterials
The commercial GO and GQDs selected to prepare the new hybrids with the cationic porphyrins were characterized by different techniques (UV-VIS, fluorescence and Raman spectroscopies, and electron microscopy) in order to evaluate the influence of the structural composition and morphological features in the electronic and optical properties of the final hybrid nanomaterials.
Both carbon nanomaterials absorb strongly in the UV region ( Figure 2a). GO displays two main peaks around 230 nm (π-π* electronic transitions of the graphitic sp 2 domains), with a shoulder around 300 nm (n-π* transitions), while for GQDs such electronic transitions originate absorption bands peaked at 249 nm and had a shoulder at 287 nm. In addition, a typical broad absorption band with peaks around 370 and 390 nm can be detected in the absorption spectrum of GQDs, which is ascribed to the absorption of different surface functional groups [36][37][38].
The GQDs solutions exhibit a bright cyan-blue fluorescence under UV light irradiation (insets Figure 2b) at a concentration as low as 0.1 µg mL −1 , while GO aqueous suspensions do not show fluorescence in these conditions ( Figure S10). As highlighted in Figure 2b, the fluorescence intensity and maximum emission wavelength of GQDs at this concentration depend on the excitation wavelength (λ exc , 310-430 nm). It has been reported that excitationdependent emission fluorescence in GQDs may be ascribed to particle size effects and their surface properties. This is in line with changes in the location of the main emission peak, λ1, which shows a 15.5 nm shift, depending on the λ exc . The shifts observed for the other emission peaks (λ2~419 nm and λ3~398.5 nm) are negligible. The maximum intensity for λ1 was recorded at λ exc = 370 nm, while the minimum intensity occurred at λ exc = 330 nm.
Both carbon nanomaterials absorb strongly in the UV region ( Figure 2a). GO displays two main peaks around 230 nm (π-π* electronic transitions of the graphitic sp 2 domains), with a shoulder around 300 nm (n-π* transitions), while for GQDs such electronic transitions originate absorption bands peaked at 249 nm and had a shoulder at 287 nm. In addition, a typical broad absorption band with peaks around 370 and 390 nm can be detected in the absorption spectrum of GQDs, which is ascribed to the absorption of different surface functional groups [36][37][38]. Two strong Raman bands were observed for GO and GQDs ( Figure 2c): (i) the G band (ca. 1600 cm −1 ), characteristic of a carbon network with sp 2 hybridization, arising mainly from the in-plane C-C deformations; (ii) the disorder-induced D band (around 1350 cm −1 ), activated by symmetry breaking at defects and edges of the graphene lattice, namely oxygen groups on sp 3 -carbon moieties [39][40][41]. Moreover, both GO and GQDs show a broad feature with low intensity at about 2500-3100 cm −1 , corresponding to 2D and combination bands, which is more pronounced for the GQDs sample, indicating a multilayered and disordered arrangement of the carbon nanosheets [38,41,42]. The average thickness of the individual GO flakes was found to be 1.5-1.9 nm (3 layers) by atomic force microscopy (AFM, Figure S11), while for GQDs the manufacturer indicated a 1-2.0 nm thickness (3-4 layers) [42][43][44].
The morphology of the carbon nanomaterials was analyzed by scanning transmission electron microscopy (STEM, Figure S11). GO sheets are observed as thin folded sheets of low contrast, while GQDs appear as quasi-spherical nanoparticles with higher contrast. The average diameter of the GQDs (26.0 ± 3.8 nm), measured as the circles circumscribing the dots, was larger than expected (~10 nm). This might be due to the coalescence of the GQDs that had probably occurred to some extent during the STEM analysis.

Synthesis and Characterization of Non-Covalent Hybrids of Porphyrin@Carbon Nanomaterials
Herein, we report the multifunctional hybrids obtained from each tetracationic porphyrin and the nanomaterials GO and GQDs. The interactions between both components were monitored by spectrophotometric and spectrofluorometric titrations at room temperature, and the ensuing hybrids porphyrin@GO and porphyrin@GQDs (where porphyrin stands for TMPyP, Zn-TMPyP, or P1-C 5 ) were characterized by Raman spectroscopy and electron microscopy.

Porphyrin@GO Hybrids
The spectrophotometric and spectrofluorometric titrations of porphyrins with GO typically led to a red-shift of the porphyrin's absorbance and the quenching of its original fluorescence ( Figure 3).
After the non-covalent functionalization with porphyrins, the GO sheets became wrinkled and decorated with darker flakes, which presumably correspond to over stacked porphyrin layers, as illustrated in Figure 4 for the TMPyP@GO sample. The STEM images suggest that the whole surface of GO sheets is covered by such porphyrin layers although there is a tendency for the accumulation of porphyrins at the edges or wrinkles of GO sheets. This can be explained because the aromatic groups in the GO sheets induce a homogeneous distribution of the porphyrins over the GO surfaces through π-π stacking but still, there is a preferential distribution of porphyrin through electrostatic interactions between the deprotonated carboxyl groups at the edges of GO structure and the positively charged pyridyl groups of the porphyrin.  Upon addition of successive aliquots of GO to TMPyP, the original Soret band of TMPyP (λ max 423 nm) was red-shifted to 437 nm (∆λ 14 nm), and its absorption intensity decreased ( Figure 3(a1)). In addition to these changes, a decrease of the porphyrin emission band intensity at 653 nm was observed ( Figure 3(a2)). These observations support the formation of TMPyP@GO hybrids through non-covalent interactions caused by the molecular flattening of the four cationic methylpyridinium moieties of the porphyrin molecules onto the GO sheets [45,46].
Zn-TMPyP displayed a slightly lower red-shift of the Soret band (from 437 nm to 450 nm, ∆λ 13 nm) when compared to TMPyP, and the final absorbance of the hybrid material did not decrease (Figure 3(b1)). Among the porphyrins, Zn-TMPyP is more fluorescent than TMPyP, but the addition of small aliquots of GO suspension caused a pronounced quenching effect in both systems (Figure 3(b2)). The strong affinity of Zn-TMPyP towards GO may have the contribution of the coordination of Zn(II) to oxygen groups present at the GO surfaces [47].
In order to perform the experiments in similar absorbance ranges, the concentrations of P1-C 5 and GO were increased, respectively, to 6.0 µm and 3.0 mg mL −1 , thus maintaining the ratio of the respective amounts used in the previous studies. The addition of GO to P1-C 5 caused a small red-shift (from 465 to 468, ∆λ 3 nm) of the Soret band (Figure 3(c1)), but the fluorescence was also quenched (Figure 3(c2)). The major spectroscopic differences observed in the titration of P1-C 5 in relation to the other charged porphyrins investigated can be explained by its chemical structure, which contains two fused aromatic rings bearing the positively charged pyridyl units at opposite β-positions and phenyl rings at the mesopositions. One possible explanation for the negligible UV-VIS changes after GO addition is that P1-C 5 is more planar than TMPyP and its substituents are not rotated to lie flat in the GO as they do in TMPyP.
After the non-covalent functionalization with porphyrins, the GO sheets became wrinkled and decorated with darker flakes, which presumably correspond to over-stacked porphyrin layers, as illustrated in Figure 4 for the TMPyP@GO sample. The STEM images suggest that the whole surface of GO sheets is covered by such porphyrin layers, although there is a tendency for the accumulation of porphyrins at the edges or wrinkles of GO sheets. This can be explained because the aromatic groups in the GO sheets induce a homogeneous distribution of the porphyrins over the GO surfaces through π-π stacking, but still, there is a preferential distribution of porphyrin through electrostatic interactions between the deprotonated carboxyl groups at the edges of GO structure and the positively charged pyridyl groups of the porphyrin.  Several Raman studies were carried out in order to assess the above hypothesis on the type of chemical interactions underlying the formation of the hybrids ( Figure 5). We have previously demonstrated that the binding of TMPyP to the edges of GO causes an intensification of the porphyrin's Raman band at 978 cm −1 in the TMPyP@GO hybrids, associated with an in-plane bending of the positively-charged pyridyl groups ( Figure S12) [45]. Thus, TMPyP is attached to the edges of GO, mostly via electrostatic interactions involving the methylated pyridines and the deprotonated carboxyl groups of GO.
On the other hand, the Zn-TMPyP@GO hybrid ( Figure 5a) shows a vibrational mode around 1006 cm −1 , which corresponds to the vibration of the porphyrin's core. This suggests that the interaction between Zn-TMPyP and GO occurs mainly through the porphyrin core.
To the best of our knowledge, there are no Raman assignments of the vibrational modes for P1-C5 porphyrin. Still, in the P1-C5@GO hybrid (Figure 5b), both the D and G bands of GO are broadened and more well-resolved, containing additional peaks that might be ascribed to the vibrational modes of individual P1-C5, proving the successful interaction between the two components of the hybrid. Several Raman studies were carried out in order to assess the above hypothesis on the type of chemical interactions underlying the formation of the hybrids ( Figure 5). We have previously demonstrated that the binding of TMPyP to the edges of GO causes an intensification of the porphyrin's Raman band at 978 cm −1 in the TMPyP@GO hybrids, associated with an in-plane bending of the positively-charged pyridyl groups ( Figure S12) [45]. Thus, TMPyP is attached to the edges of GO, mostly via electrostatic interactions involving the methylated pyridines and the deprotonated carboxyl groups of GO.  Several Raman studies were carried out in order to assess the above hypothesis on the type of chemical interactions underlying the formation of the hybrids ( Figure 5). We have previously demonstrated that the binding of TMPyP to the edges of GO causes an intensification of the porphyrin's Raman band at 978 cm −1 in the TMPyP@GO hybrids associated with an in-plane bending of the positively-charged pyridyl groups (Figure S12 [45]. Thus, TMPyP is attached to the edges of GO, mostly via electrostatic interactions involving the methylated pyridines and the deprotonated carboxyl groups of GO. On the other hand, the Zn-TMPyP@GO hybrid ( Figure 5a) shows a vibrational mode around 1006 cm −1 , which corresponds to the vibration of the porphyrin's core. This sug gests that the interaction between Zn-TMPyP and GO occurs mainly through the porphy rin core.
To the best of our knowledge, there are no Raman assignments of the vibrationa modes for P1-C5 porphyrin. Still, in the P1-C5@GO hybrid (Figure 5b), both the D and G bands of GO are broadened and more well-resolved, containing additional peaks tha might be ascribed to the vibrational modes of individual P1-C5, proving the successfu interaction between the two components of the hybrid.

Porphyrin@GQDs Hybrids
The synthesis of porphyrin@GQDs hybrids was monitored by optical measurements as noted above, but unlike for the GO system, aliquots of the porphyrin solution were added to the GQDs solution (see Experimental section). This alteration was required be On the other hand, the Zn-TMPyP@GO hybrid ( Figure 5a) shows a vibrational mode around 1006 cm −1 , which corresponds to the vibration of the porphyrin's core. This suggests that the interaction between Zn-TMPyP and GO occurs mainly through the porphyrin core.
To the best of our knowledge, there are no Raman assignments of the vibrational modes for P1-C 5 porphyrin. Still, in the P1-C 5 @GO hybrid (Figure 5b), both the D and G bands of GO are broadened and more well-resolved, containing additional peaks that might be ascribed to the vibrational modes of individual P1-C 5 , proving the successful interaction between the two components of the hybrid.

Porphyrin@GQDs Hybrids
The synthesis of porphyrin@GQDs hybrids was monitored by optical measurements, as noted above, but unlike for the GO system, aliquots of the porphyrin solution were added to the GQDs solution (see Experimental section). This alteration was required because the GQDs show more intense luminescence as compared to the porphyrins. These studies were monitored by spectrofluorometric titrations, as shown in Figure 6; note that the GQDs (0.1 µg mL −1 ) suspensions show low absorbance, which precluded the monitoring by spectrophotometric titrations ( Figure S13). All the samples have been excited at 330 nm with 10 min breaks between each measure to promote the solutions' stability and decrease the effect of GQDs' self-quenching through time (26%, Figure S14). In all cases, the porphyrins outweighed the self-quenching of GQDs ( Figure 6). By increasing the total porphyrin concentration up to 4.0 µM, TMPyP causes a 73% quenching of the initial GQDs' fluorescence, followed by P1-C 5 (68%) and Zn-TMPyP (62%). This occurs because the absorption bands of the porphyrins (Figure 1) partially overlap with the emission bands of the GQDs at the selected excitation wavelength; therefore, the inner filter effect of porphyrins on GQDs results in substantial fluorescence quenching of the GQDs throughout the titrations, particularly for TMPyP.
ceutics 2021, 13, x 12 the GQDs (0.1 µ g mL −1 ) suspensions show low absorbance, which precluded the mon ing by spectrophotometric titrations ( Figure S13). All the samples have been excited a nm with 10 min breaks between each measure to promote the solutions' stability and crease the effect of GQDs' self-quenching through time (26%, Figure S14). In all cases porphyrins outweighed the self-quenching of GQDs ( Figure 6). By increasing the porphyrin concentration up to 4.0 μм, TMPyP causes a 73% quenching of the initial GQ fluorescence, followed by P1-C5 (68%) and Zn-TMPyP (62%). This occurs because th sorption bands of the porphyrins (Figure 1) partially overlap with the emission band the GQDs at the selected excitation wavelength; therefore, the inner filter effect of por rins on GQDs results in substantial fluorescence quenching of the GQDs throughou titrations, particularly for TMPyP. The conjugation of GQDs with porphyrins led to a slight increase of the particl erage size but still within a comparable size range, as exemplified in Figure 7a (28.4 nm) for the TMPyP@GQDs hybrids. The Raman spectrum of the TMPyP@GQDs hy shows features that are also observed in the corresponding spectra of the free sampl GQDs and TMPyP (Figure 7b). However, the G band peak assigned to the GQDs is sh to 1645 cm -1 in relation to the free sample (1600 cm −1 ), while the D band (1350 cm peaked at the same wavenumber as the original GQDs sample. Additionally, some Ra The conjugation of GQDs with porphyrins led to a slight increase of the particle average size but still within a comparable size range, as exemplified in Figure 7a (28.4 ± 4.4 nm) for the TMPyP@GQDs hybrids. The Raman spectrum of the TMPyP@GQDs hybrid shows features that are also observed in the corresponding spectra of the free samples of GQDs and TMPyP (Figure 7b). However, the G band peak assigned to the GQDs is shifted to 1645 cm −1 in relation to the free sample (1600 cm −1 ), while the D band (1350 cm −1 ) is peaked at the same wavenumber as the original GQDs sample. Additionally, some Raman bands ascribed to the porphyrin seem enhanced in the hybrid, as compared to the free sample, such as the bands at: 1006 cm −1 (stretching mode involving the α-and mesocarbons of the porphyrin core, ν C α -C m ); 1551 cm −1 (ν C β -C β ); and 1453 cm −1 (ν C α -C β ). These observations suggest that the TMPyP molecules are associated with the GQDs by non-covalent interactions. A similar interpretation can be extended to the Raman data obtained for the other two hybrids (Zn-TMPyP@GQDs and P1-C 5 @GQDs, Figure S15).

Photo-Antiproliferative Activity of non-Immobilized Porphyrins on T24 Human Bladder Cancer Cells
Before testing the photo-antiproliferative activity of the non-immobilized porphyrins, their photostability was investigated under increasing irradiations of both RL (0-35 J cm −2 ) and BL (0-20 J cm −2 ) lights. While TMPyP and Zn-TMPyP proved to be photostable under both light irradiations (data not shown), P1-C5 demonstrated a high instability under both light irradiations, as shown by the progressive decrease of its absorption spectrum during irradiation ( Figure S16). Still, none of the spectra evidenced an isosbestic point, which suggested the absence of stable products formed during the photodegradation process.
The light doses selected for irradiating the T24 cells were 2.5 J cm −2 for BL and 10 times higher for RL (25.0 J cm −2 ); indeed, the higher RL dose was necessary for the activation of the compounds, in line with their molar absorption coefficients at 417 nm and 625 nm (Table 1). Moreover, both light doses were chosen as non-toxic for the cells when applied alone (without any porphyrin or porphyrin@GO/GQDs, data not shown).
The obtained results under both irradiations when the T24 human BC cells were incubated for 1 h with the three cationic porphyrins are represented in Figure 8.

Photo-Antiproliferative Activity of non-Immobilized Porphyrins on T24 Human Bladder Cancer Cells
Before testing the photo-antiproliferative activity of the non-immobilized porphyrins, their photostability was investigated under increasing irradiations of both RL (0-35 J cm −2 ) and BL (0-20 J cm −2 ) lights. While TMPyP and Zn-TMPyP proved to be photostable under both light irradiations (data not shown), P1-C 5 demonstrated a high instability under both light irradiations, as shown by the progressive decrease of its absorption spectrum during irradiation ( Figure S16). Still, none of the spectra evidenced an isosbestic point, which suggested the absence of stable products formed during the photodegradation process.
The light doses selected for irradiating the T24 cells were 2.5 J cm −2 for BL and 10 times higher for RL (25.0 J cm −2 ); indeed, the higher RL dose was necessary for the activation of the compounds, in line with their molar absorption coefficients at 417 nm and 625 nm (Table 1). Moreover, both light doses were chosen as non-toxic for the cells when applied alone (without any porphyrin or porphyrin@GO/GQDs, data not shown).
The obtained results under both irradiations when the T24 human BC cells were incubated for 1 h with the three cationic porphyrins are represented in Figure 8.
TMPyP showed significant antiproliferative activity against T24 human BC cells (Figure 8 Finally, P1-C5 proved to be more active under red light (Figure 8, bottom): according to its absorption spectrum, this effect was higher under RL (IC50 of 0.14 ± 0.06 μm) at the concentration range of 0-0.5 μm, while the IC50 under BL was 2.20 ± 0.02 μm at the concentration range of 0-5.0 μm.
As noted previously, P1-C5 has been shown to be photo-unstable at increasing red and blue light doses ( Figure S16), although its photodegradation seems to proceed without the formation of stable photoproducts. Considering the doses of BL and RL used for the experiments on cells, its instability under 25.0 J cm −2 of RL could be noteworthy for Finally, P1-C 5 proved to be more active under red light (Figure 8, bottom): according to its absorption spectrum, this effect was higher under RL (IC 50 of 0.14 ± 0.06 µm) at the concentration range of 0-0.5 µm, while the IC 50 under BL was 2.20 ± 0.02 µm at the concentration range of 0-5.0 µm.
As noted previously, P1-C 5 has been shown to be photo-unstable at increasing red and blue light doses ( Figure S16), although its photodegradation seems to proceed without the formation of stable photoproducts. Considering the doses of BL and RL used for the experiments on cells, its instability under 25.0 J cm −2 of RL could be noteworthy for interpreting the in vitro experiments on cells. Therefore, to evaluate the potential role of any formed photoproduct in P1-C 5 cellular action, this porphyrin was dissolved in PBS, then preirradiated with RL at a total light dose of 25.0 J cm −2 (as this light dose induced a marked photodegradation), and subsequently incubated with cells at the same concentration range used for non-preirradiated porphyrin. The preirradiated P1-C 5 solution did not show any cytotoxic activity (see cell viability profile in Figure S17), confirming that the high photodynamic activity of the porphyrin P1-C 5 is not associated with any photodegradation products.

Detection of Singlet Oxygen and Other ROS in Cell Cultures
The production of singlet oxygen ( 1 O 2 ) by the porphyrins in T24 cell cultures was quantified under both BL and RL irradiations using Singlet Oxygen Sensor Green (SOSG) [34]. The 1 O 2 production was found to be proportional to the concentration of the compounds in the treated cells under RL ( Figure 9) and BL ( Figure S18). The amount of 1 O 2 correlates very well with their antiproliferative activity; in fact, P1-C 5 , which was able to produce more 1 O 2 under RL (Figure 9a), showed the lowest IC 50 compared to the other porphyrins. Even under BL, there was correspondence between the porphyrin with the lowest IC 50 (Zn-TMPyP) and the highest 1 O 2 production at all tested concentrations ( Figure S19a).
The production of singlet oxygen ( O2) by the porphyrins in T24 cell cultures wa quantified under both BL and RL irradiations using Singlet Oxygen Sensor Green (SOSG [34]. The 1 O2 production was found to be proportional to the concentration of the com pounds in the treated cells under RL ( Figure 9) and BL ( Figure S18). The amount of 1 O correlates very well with their antiproliferative activity; in fact, P1-C5, which was able to produce more 1 O2 under RL (Figure 9a), showed the lowest IC50 compared to the othe porphyrins. Even under BL, there was correspondence between the porphyrin with th lowest IC50 (Zn-TMPyP) and the highest 1 O2 production at all tested concentrations (Figur S19a).
2',7'-dichlorofluorescein diacetate (DCFDA) was used to detect the formation of tota ROS. When irradiated with RL, the production of ROS by TMPyP and Zn-TMPyP wa about 1.5 times higher than the controls (cells irradiated with RL alone or cells incubated with porphyrins without irradiation), as shown in Figure 9b. On the other hand, P1-C5 did not seem to generate significant ROS concerning the reference samples.
Overall, the photo-antiproliferative activity of non-immobilized porphyrins irradi ated with RL in T24 cells was not strictly correlated to the production of ROS. In fact, P1 C5, the most active porphyrin in killing T24 cells under RL, did not produce a significan amount of ROS, while TMPyP, the less active PS under RL, produced more ROS insid the cells than the other PS under study. Instead, BL ( Figure S18b) induced the production of more ROS by the compounds; in particular, P1-C5 demonstrated higher amounts than the other two porphyrins and four times higher than the controls. This result suggests tha such ROS are not the primary cause of photodamage, probably because the cells hav defense mechanisms to destroy them, which does not occur for 1 O2.

Flow Cytometric Analysis
To investigate the type of cell death involved in the treatments with non-immobilized porphyrins, flow cytometric analysis of cell death mechanism was performed. Before flow 2 ,7 -dichlorofluorescein diacetate (DCFDA) was used to detect the formation of total ROS. When irradiated with RL, the production of ROS by TMPyP and Zn-TMPyP was about 1.5 times higher than the controls (cells irradiated with RL alone or cells incubated with porphyrins without irradiation), as shown in Figure 9b. On the other hand, P1-C 5 did not seem to generate significant ROS concerning the reference samples.
Overall, the photo-antiproliferative activity of non-immobilized porphyrins irradiated with RL in T24 cells was not strictly correlated to the production of ROS. In fact, P1-C 5 , the most active porphyrin in killing T24 cells under RL, did not produce a significant amount of ROS, while TMPyP, the less active PS under RL, produced more ROS inside the cells than the other PS under study. Instead, BL ( Figure S18b) induced the production of more ROS by the compounds; in particular, P1-C 5 demonstrated higher amounts than the other two porphyrins and four times higher than the controls. This result suggests that such ROS are not the primary cause of photodamage, probably because the cells have defense mechanisms to destroy them, which does not occur for 1 O 2 .

Flow Cytometric Analysis
To investigate the type of cell death involved in the treatments with non-immobilized porphyrins, flow cytometric analysis of cell death mechanism was performed. Before flow cytometric analysis, T24 cells were previously treated with the three porphyrins at a concentration corresponding to their IC 50 , and then they were irradiated with RL at a light dose of 25 J cm −2 .
As can be seen from the right side of Figure 10a, for the sample treated with 1.76 µm TMPyP, the population found in the quadrant Q3 (corresponding to the living cells) was about 50% of the total population present, while the rest was mainly distributed in quadrants Q2 and Q4, thus indicating cell death due to apoptosis. dose of 25 J cm −2 .
As can be seen from the right side of Figure 10a, for the sample treated with 1.76 μm TMPyP, the population found in the quadrant Q3 (corresponding to the living cells) was about 50% of the total population present, while the rest was mainly distributed in quadrants Q2 and Q4, thus indicating cell death due to apoptosis.
In the case of the sample treated with 0.25 μm of Zn-TMPyP, about half of the tested population appeared to be alive, while the remaining part was divided into quadrants Q2 and Q4 as TMPyP, thus indicating an apoptotic type of death (Figure 10a, center).
For P1-C5, although the concentration corresponding to its IC50 0.14 μm was also used, the two-dimensional graph shows a slight reduction of the population in Q3, corresponding to the living cells; this result seems in contrast to the viability test with resazurin, but this could be explained by the greater sensitivity of the flow cytometric analysis concerning the viability test or to a late cellular damage due to a prolonged time between the marking and the reading at the flow cytometry. The remaining population is divided between quadrants Q2 and Q4, indicating death by apoptotic mechanism (Figure 10a, right).
This analysis clarifies the prevailing mechanism of apoptotic cell death induced by the three porphyrins under examination (Figure 10b).

Preparation of Porphyrin@GO and Porphyrin@GQD Hybrids for Cellular Experiments
To assess the performance of GO and GQDs as nanocarriers for the cationic porphyrins and to maximize their loading capability, an estimation of the saturation binding ratio between each porphyrin and the carbon nanomaterials was performed through spectro- In the case of the sample treated with 0.25 µm of Zn-TMPyP, about half of the tested population appeared to be alive, while the remaining part was divided into quadrants Q2 and Q4 as TMPyP, thus indicating an apoptotic type of death (Figure 10a, center).
For P1-C 5 , although the concentration corresponding to its IC 50 0.14 µm was also used, the two-dimensional graph shows a slight reduction of the population in Q3, corresponding to the living cells; this result seems in contrast to the viability test with resazurin, but this could be explained by the greater sensitivity of the flow cytometric analysis concerning the viability test or to a late cellular damage due to a prolonged time between the marking and the reading at the flow cytometry. The remaining population is divided between quadrants Q2 and Q4, indicating death by apoptotic mechanism (Figure 10a, right).
This analysis clarifies the prevailing mechanism of apoptotic cell death induced by the three porphyrins under examination (Figure 10b).

Photo-Antiproliferative Activity of Porphyrin@Carbon Nanomaterials on T24 Cells Preparation of Porphyrin@GO and Porphyrin@GQD Hybrids for Cellular Experiments
To assess the performance of GO and GQDs as nanocarriers for the cationic porphyrins and to maximize their loading capability, an estimation of the saturation binding ratio between each porphyrin and the carbon nanomaterials was performed through spectrofluorimetric titrations (Figure 11; see details in the Experimental section). The binding ratios of the three porphyrin@GO hybrids were determined graphically (Figure 11a-c) by interpolating the obtained curves with two broken lines; the intersection between the two lines represents the binding ratio (Table 2). A similar approach was performed to determine the binding ratio of porphyrins@GQDs (Figure 11d-f) but, as mentioned above, these titrations were performed oppositely. The porphyrin solutions were used as a titrating agent rather than the GQDs due to their stronger luminescence emission when compared to porphyrins fluorescence; the titration of a porphyrin solution with GQDs resulted in the immediate detector signal saturation. The determined binding ratios are summarized in Table 2; the values of the binding ratio of porphyrin/GO varied between 1/4.2 and 1/2, while the binding ratio of porphyrin/GQDs varied between 2/1 and 3/1.  (Figure 11; see details in the Experimental section). The binding ratios of the three porphyrin@GO hybrids were determined graphically (Figure 11a-c) by interpolating the obtained curves with two broken lines; the intersection between the two lines represents the binding ratio (Table 2). A similar approach was performed to determine the binding ratio of porphyrins@GQDs (Figure 11d-f) but, as mentioned above, these titrations were performed oppositely. The porphyrin solutions were used as a titrating agent rather than the GQDs due to their stronger luminescence emission when compared to porphyrins fluorescence; the titration of a porphyrin solution with GQDs resulted in the immediate detector signal saturation. The determined binding ratios are summarized in Table 2; the values of the binding ratio of porphyrin/GO varied between 1/4.2 and 1/2, while the binding ratio of porphyrin/GQDs varied between 2/1 and 3/1. Table 2. Binding ratios of the porphyrin@GO and porphyrin@GQDs hybrids in PBS.

Photo-Antiproliferative Activity of Porphyrin@GO Hybrids
The in vitro tests on cells carried out to verify the photo-antiproliferative activity of the porphyrin@GO hybrids after irradiation with RL ( Figure 12) followed the same procedure as with non-immobilized porphyrins, but the incubation time was increased up to 4 h to allow both the cell uptake of porphyrin@GO and/or porphyrin release from the hybrids.

Photo-Antiproliferative Activity of Porphyrin@GO Hybrids
The in vitro tests on cells carried out to verify the photo-antiproliferative activity of the porphyrin@GO hybrids after irradiation with RL ( Figure 12) followed the same procedure as with non-immobilized porphyrins, but the incubation time was increased up to 4 h to allow both the cell uptake of porphyrin@GO and/or porphyrin release from the hybrids. phyrin (0.25 μm).
When the cells were treated with increasing concentrations of P1-C5@GO hybrid, the obtained IC50 was 1.68 ± 0.03 μg mL −1 , corresponding to a concentration of the non-immobilized porphyrin of 0.31 μm (Figure 12c). This value was only slightly higher than the IC50 of the non-immobilized P1-C5 (0.14 μm), thus the photodynamic activity of both nonimmobilized and immobilized porphyrin is comparable.

Photo-Antiproliferative Activity of Porphyrin@GQDs Hybrids
A previous check of GQDs cell toxicity-biocompatibility was performed before evaluating the photo-antiproliferative activity of the porphyrin@GQDs hybrids ( Figure S19). The photocytotoxicity of non-functionalized GQDs against T24 cells was assessed in a wide range of concentrations (0-500.0 μg mL −1 ) as GQDs can produce singlet oxygen per se, without being charged with the PS, thus causing "independent" cytotoxic effects upon irradiation. The cells were incubated for 4 h with increasing concentrations of GQDs and were then irradiated with RL. GQDs were shown to be phototoxic at high concentrations, with an IC50 of 14.96 ± 0.94 μg mL −1 ; therefore, the concentration of GQDs present in the hybrids was lower than the IC50 identified by this test.
To evaluate the photo-antiproliferative activity of the porphyrin@GQDs on T24 cells, the same procedure used for the non-immobilized porphyrins was followed. By using a concentration range of 0-50 μg mL −1 , TMPyP@GQDs showed an IC50 of 10.88 ± 0.16 μg mL −1 after irradiation with RL (Figure 13a). At this IC50, the TMPyP concentration in the hybrid corresponded to 6.11 μm, again a value significantly higher than the IC50 before its immobilization (1.76 μm).
The Zn-TMPyP hybrid was tested in a concentration range of 0-50.0 µg mL −1 , ten times lower than that used for TMPyP@GO, which allowed identification of the IC 50 of 6.17 ± 0.16 µg mL −1 (Figure 12b). The molarity of Zn-TMPyP corresponding to this concentration was 0.94 µm, approximately 4 times higher than the IC 50 of non-immobilized porphyrin (0.25 µm).
When the cells were treated with increasing concentrations of P1-C 5 @GO hybrid, the obtained IC 50 was 1.68 ± 0.03 µg mL −1 , corresponding to a concentration of the nonimmobilized porphyrin of 0.31 µm (Figure 12c). This value was only slightly higher than the IC 50 of the non-immobilized P1-C 5 (0.14 µm), thus the photodynamic activity of both non-immobilized and immobilized porphyrin is comparable.

Photo-Antiproliferative Activity of Porphyrin@GQDs Hybrids
A previous check of GQDs cell toxicity-biocompatibility was performed before evaluating the photo-antiproliferative activity of the porphyrin@GQDs hybrids ( Figure S19). The photocytotoxicity of non-functionalized GQDs against T24 cells was assessed in a wide range of concentrations (0-500.0 µg mL −1 ) as GQDs can produce singlet oxygen per se, without being charged with the PS, thus causing "independent" cytotoxic effects upon irradiation. The cells were incubated for 4 h with increasing concentrations of GQDs and were then irradiated with RL. GQDs were shown to be phototoxic at high concentrations, with an IC 50 of 14.96 ± 0.94 µg mL −1 ; therefore, the concentration of GQDs present in the hybrids was lower than the IC 50 identified by this test.
To evaluate the photo-antiproliferative activity of the porphyrin@GQDs on T24 cells, the same procedure used for the non-immobilized porphyrins was followed. By using a concentration range of 0-50 µg mL −1 , TMPyP@GQDs showed an IC 50 of 10.88 ± 0.16 µg mL −1 after irradiation with RL (Figure 13a). At this IC 50 , the TMPyP concentration in the hybrid corresponded to 6.11 µm, again a value significantly higher than the IC 50 before its immobilization (1.76 µm).
The Zn-TMPyP@GQDs hybrid has been tested in the concentration range of 0-20 µg mL (Figure 13b), and the IC 50 was 0.37 ± 0.06 µg mL −1 , corresponding to a porphyrin concentration of 0.22 µm, was slightly lower than that identified for the non-immobilized Zn-TMPyP (0.25 µm).

Intracellular Localization
To determine the T24 intracellular distribution of the three porphyrins in the free form and in the GO and GQD hybrids, confocal microscopy was used to visualize the intrinsic fluorescence of the compounds (porphyrin channel) along with LysoTracker Green DND-26, a marker for acidic organelles such as early endosomes and lysosomes, both as individual channels and merged.
As shown in Figure 14, high co-localization (yellow) of the porphyrin red fluorescence with the green fluorescence of LysoTracker demonstrated that TMPyP and Zn-TMPyP in the free form are primarily transported to the lysosomes. The same was observed with the GO and GQD hybrids. It must be pointed out that large aggregates are present outside the cells for GO hybrids, indicating the difficulty of these hybrids in being fully internalized. However, the lysosomal localization of the two porphyrins indicates that they can be released from the GO platform in a high enough concentration to carry out the photosensitizing activity of the drugs. Conversely, the distribution of free P1-C5 was not confined to specific organelles, and its signal did not co-localize with organelle-specific probes ( Figure 15). Furthermore, such

Intracellular Localization
To determine the T24 intracellular distribution of the three porphyrins in the free form and in the GO and GQD hybrids, confocal microscopy was used to visualize the intrinsic fluorescence of the compounds (porphyrin channel) along with LysoTracker Green DND-26, a marker for acidic organelles such as early endosomes and lysosomes, both as individual channels and merged.
As shown in Figure 14, high co-localization (yellow) of the porphyrin red fluorescence with the green fluorescence of LysoTracker demonstrated that TMPyP and Zn-TMPyP in the free form are primarily transported to the lysosomes. The same was observed with the GO and GQD hybrids. It must be pointed out that large aggregates are present outside the cells for GO hybrids, indicating the difficulty of these hybrids in being fully internalized. However, the lysosomal localization of the two porphyrins indicates that they can be released from the GO platform in a high enough concentration to carry out the photosensitizing activity of the drugs.

Intracellular Localization
To determine the T24 intracellular distribution of the three porphyrins in the free form and in the GO and GQD hybrids, confocal microscopy was used to visualize the intrinsic fluorescence of the compounds (porphyrin channel) along with LysoTracker Green DND-26, a marker for acidic organelles such as early endosomes and lysosomes, both as individual channels and merged.
As shown in Figure 14, high co-localization (yellow) of the porphyrin red fluorescence with the green fluorescence of LysoTracker demonstrated that TMPyP and Zn-TMPyP in the free form are primarily transported to the lysosomes. The same was observed with the GO and GQD hybrids. It must be pointed out that large aggregates are present outside the cells for GO hybrids, indicating the difficulty of these hybrids in being fully internalized. However, the lysosomal localization of the two porphyrins indicates that they can be released from the GO platform in a high enough concentration to carry out the photosensitizing activity of the drugs. Conversely, the distribution of free P1-C5 was not confined to specific organelles, and its signal did not co-localize with organelle-specific probes ( Figure 15). Furthermore, such Conversely, the distribution of free P1-C 5 was not confined to specific organelles, and its signal did not co-localize with organelle-specific probes ( Figure 15). Furthermore, such behavior has also been observed in the corresponding hybrids. The corresponding antiproliferative activity showed their efficacy, even without specific cell localization and damage. behavior has also been observed in the corresponding hybrids. The corresponding antiproliferative activity showed their efficacy, even without specific cell localization and damage.

Conclusions
The evaluation of the photo-antiproliferative activity of the tetracationic porphyrins (TMPyP, Zn-TMPyP, and P1-C5) before and after non-covalent immobilization onto GO and GQDs proved their potential as photosensitizers towards T24 human BC cells under different light irradiations. The photodynamic efficacy of the non-immobilized porphyrins towards T24 human BC cells was dependent on the wavelengths used. TMPyP and Zn-TMPyP showed higher efficiency under BL (417 nm, 2.5 J cm −2 , IC50 values of 0.42 and 0.22 μm, respectively), while P1-C5 proved to be more active under RL (625 nm, 25 J cm −2 , IC50 value of 0.14 μm). The observed photocytotoxicity is mainly related to the specific high production of singlet oxygen (rather than ROS), which is mainly responsible for cancer cell death through apoptosis.
The antiproliferative photodynamic activity of the porphyrins in the hybrid materials after irradiation with RL was similar or only slightly reduced when compared to the nonimmobilized porphyrins. In general, porphyrin@GQDs demonstrated a higher PDT efficiency than porphyrin@GO hybrids, possibly due to their smaller size, which facilitates cell internalization. The most promising hybrid systems were Zn-TMPyP@GQDs, followed by P1-C5@GO and P1-C5@GQDs. Their low IC50 values prove that porphyrin functionalized carbon nanomaterials retain the high potential as PS for PDT, thus paving the way for future biomedical applications of such hybrids on cancer therapy, simplifying therapeutics formulation, increasing the stability of the compounds during distribution in the body after administration, and thus providing a greater chance to target the diseased tissue.
Results from the use of these carbon-based nanoplates suggest that these materials are useful for improving the stability of porphyrins in aqueous solutions. Moreover, these platforms present multiple functional groups that could be exploited for the conjugation of active-targeting molecules towards specific BC receptors (i.e., EGFR, HER-2, FGFR-3, etc.) [13]. A further advantage is their ability to absorb near-infrared light and convert it into heat, which is harmful to cells [48]. This phenomenon can be exploited to increase the

Conclusions
The evaluation of the photo-antiproliferative activity of the tetracationic porphyrins (TMPyP, Zn-TMPyP, and P1-C 5 ) before and after non-covalent immobilization onto GO and GQDs proved their potential as photosensitizers towards T24 human BC cells under different light irradiations. The photodynamic efficacy of the non-immobilized porphyrins towards T24 human BC cells was dependent on the wavelengths used. TMPyP and Zn-TMPyP showed higher efficiency under BL (417 nm, 2.5 J cm −2 , IC 50 values of 0.42 and 0.22 µm, respectively), while P1-C 5 proved to be more active under RL (625 nm, 25 J cm −2 , IC 50 value of 0.14 µm). The observed photocytotoxicity is mainly related to the specific high production of singlet oxygen (rather than ROS), which is mainly responsible for cancer cell death through apoptosis.
The antiproliferative photodynamic activity of the porphyrins in the hybrid materials after irradiation with RL was similar or only slightly reduced when compared to the non-immobilized porphyrins. In general, porphyrin@GQDs demonstrated a higher PDT efficiency than porphyrin@GO hybrids, possibly due to their smaller size, which facilitates cell internalization. The most promising hybrid systems were Zn-TMPyP@GQDs, followed by P1-C 5 @GO and P1-C 5 @GQDs. Their low IC 50 values prove that porphyrin functionalized carbon nanomaterials retain the high potential as PS for PDT, thus paving the way for future biomedical applications of such hybrids on cancer therapy, simplifying therapeutics formulation, increasing the stability of the compounds during distribution in the body after administration, and thus providing a greater chance to target the diseased tissue.
Results from the use of these carbon-based nanoplates suggest that these materials are useful for improving the stability of porphyrins in aqueous solutions. Moreover, these platforms present multiple functional groups that could be exploited for the conjugation of active-targeting molecules towards specific BC receptors (i.e., EGFR, HER-2, FGFR-3, etc.) [13]. A further advantage is their ability to absorb near-infrared light and convert it into heat, which is harmful to cells [48]. This phenomenon can be exploited to increase the antiproliferative activity of photosensitizers for their use in photodynamic-photothermal therapy, but further in-depth studies are needed.  Figure S10: (a) Fluorescence spectra of GO at the studied conditions used for porphyrin@GO hybrids: here, 50 µL of GO (1 mg mL −1 ) were added to 1 mL of PBS solution and excited at two maxima of TMPyP (λ max 423 nm) and Zn-TMPyP (λ max 438 nm). (b) Comparison of fluorescence of GQDs and GO at the same concentration under a 366 nm lamp; Figure S11: (a) High magnification AFM image of GO flakes deposited onto the Si substrate. The black line and the colored arrows represent the cross-section and the measurement points used to calculate the thickness, respectively. (b) AFM height profile of the cross-section analysis. STEM images (transmission mode) of (c) GO and (d) GQDs. The red dotted line represents the circle circumscribing the coalesced GQDs in the microscopy analysis; Figure S12: Raman spectra (532 nm excitation) of the sheet-edge areas of TMPyP@GO hybrid; Figure S13: UV-Vis absorbance spectrum of a GQDs aqueous solution (0.1 µg mL −1 ). The inset highlights that this solution is transparent under daylight; Figure S14: Fluorescence spectra of an aqueous solution of GQDs (0.1 µg mL −1 ), excited at 330 nm at each 10 min (during 100 min). At these conditions, a 26% fluorescence quenching was observed; Figure S15: (a) Raman spectra (532 nm excitation) of free GQDs (blue line), Zn-TMPyP (green line), and Zn-TMPyP@GQDs (red line). (b) Corresponding data for P1-C 5 @GQDs; Figure S16: UV-Vis spectra of P1-C 5 under increasing light irradiation doses of (a) RL (0-35 J cm −2 ) and (b) BL (0-20 J cm −2 ). BL had an irradiance of 17 mW cm −2 and RL had an irradiation of 12 mW cm −2 ; Figure S17: Antiproliferative photodynamic activity (25.0 J cm −2 of RL) of pre-irradiated P1-C 5 at different concentrations; Figure S18: (a) Singlet oxygen production of non-immobilized porphyrins at concentrations 0.1 µm, 1.0 µm, and 5.0 µm under BL. (b) ROS production by the non-immobilized porphyrins tested at their IC 50 under BL. Data are expressed as mean ± SD of at least two independent experiments carried out in triplicate. (* p < 0.0332, ** p < 0.0021, *** p < 0.0002, **** p < 0.0001; two-way ANOVA); Figure S19: Photo-proliferative activity of GQDs at different concentrations under RL.