Identification of Blood Transport Proteins to Carry Temoporfin: A Domino Approach from Virtual Screening to Synthesis and In Vitro PDT Testing

Temoporfin (mTHPC) is one of the most promising photosensitizers used in photodynamic therapy (PDT). Despite its clinical use, the lipophilic character of mTHPC still hampers the full exploitation of its potential. Low solubility in water, high tendency to aggregate, and low biocompatibility are the main limitations because they cause poor stability in physiological environments, dark toxicity, and ultimately reduce the generation of reactive oxygen species (ROS). Applying a reverse docking approach, here, we identified a number of blood transport proteins able to bind and disperse monomolecularly mTHPC, namely apohemoglobin, apomyoglobin, hemopexin, and afamin. We validated the computational results synthesizing the mTHPC-apomyoglobin complex (mTHPC@apoMb) and demonstrated that the protein monodisperses mTHPC in a physiological environment. The mTHPC@apoMb complex preserves the imaging properties of the molecule and improves its ability to produce ROS via both type I and type II mechanisms. The effectiveness of photodynamic treatment using the mTHPC@apoMb complex was then demonstrated in vitro. Blood transport proteins can be used as molecular “Trojan horses” in cancer cells by conferring mTHPC (i) water solubility, (ii) monodispersity, and (iii) biocompatibility, ultimately bypassing the current limitations of mTHPC.

Temoporfin (mTHPC, Foscan ® ) [6][7][8] is among the most promising photosensitizers used in PDT [9]. In Europe, since 2001, mTHPC has been approved for the photodynamic treatment of head and neck squamous cell carcinoma (HNSCC) [6]. Despite its excellent photophysical properties, the poor solubility of mTHPC still represents a challenge in clinical settings. In physiological environments, mTHPC forms aggregates, resulting in co-crystallized ligands were removed from the protein structures. This dataset was used for ensemble docking calculations.

Docking
Ensemble docking calculations were carried out using the mTHPC ligand and the blood transport proteins structural database. The docking poses were obtained using the PatchDock algorithm [55]. FireDock [56] was used to refine (rearranging the side-chains of amino acids close to the ligand and adjusting the relative orientation of the molecules) and rescore the obtained poses.

MD Simulations
Proteins were described by the Amber ff14SB force field; [57] the atomic charges of mTHPC were calculated with the Merz-Singh-Kollman scheme; whereas for the rest, mTHPC was parametrized using the GAFF force field [58]. The TIP3P water model was used for all simulations, and Na + or Cl − counterions were introduced to neutralize the system. The particle mesh Ewald summation and periodic boundary conditions (PBC) were applied throughout (with a cut-off radius of 10.0). A time step of 2 fs was chosen for all MD runs, and the SHAKE algorithm was used for H-atoms. The systems were minimized and equilibrated, and then 100 ns of MD simulations were carried out.

Molecular Mechanics-Generalized Born Surface Area (MM-GBSA) Analysis
Using CPPTRAJ [59], one frame every 0.1 ns was retrieved from the MD trajectories and used as input for the MM-GBSA analysis. MM-GBSA analysis was carried out with the MMPBSA.py [60] module, determining the binding affinity between mTHPC and the different proteins. Electrostatic and van der Waals (vdW) interactions were calculated considering an infinite cut-off. The polar solvation term was calculated using the Generalized Born (GB) model, whereas the non-polar solvation term was determined using solvent-accessible, surface-area-dependent terms. 3,3 ,3 ,3 -(7,8-dihydro-21H,23H-porphine-5,10,15,20-tetrayl)tetrakis-phenol (mTHPC) (Item No. 17333) was purchased from Cayman Chemical (Ann Arbor, MI, USA). All the reagents were used without further purification. Milli-Q water was used for the preparation of all the aqueous solutions.

Synthesis of the Reconstituted Myoglobin-Temoporfin Adduct
A standard procedure [41,61] was used to prepare apomyoglobin. HCl 1 M was added dropwise to 4 mL of water solution of myoglobin (90 µM), and the solution was kept under stirring at 0 • C into an ice bath. At pH 2.5, the interaction between the heme and the protein is strongly reduced due to the protonation of its carboxylic groups. The same volume of 2-butanone was added, and the extraction process was repeated three times with fresh 2-butanone. A solution of pure apo-myoglobin was obtained. Then, a solution of mTHPC in 2-butanone (a 2:1 stoichiometry of mTHPC/apoMb was used) was added to the aqueous phase of apoMb. NaOH 5 M was added to reach a pH of 12.5, maintaining the system under gentle stirring. Deprotonation of the phenolic moieties of mTHPC allowed its migration from the organic to the aqueous phase and its quantitative intercalation with the apoMb. The mTHPC@apoMb adduct was extensively dialyzed against PBS, using an MWCO 14 kDa dialysis tubing cellulose membrane to remove 2-butanone and mTHPC in excess. A 1:1 stoichiometry of mTHPC/apoMb was obtained after purification.
Fluorescence Spectroscopy. The fluorescence spectra of mTHPC and mTHPC@apoMb were acquired with an Edinburgh FLS920 equipped with a photomultiplier Hamamatsu R928P.

Detection of Reactive Oxygen Species
ABMDMA assay. To detect and estimate the amount of singlet oxygen ( 1 O 2 ) generated by mTHPC and mTHPC@apoMb upon irradiation, the ABMDMA assay was used. AB-MDMA reacts with singlet oxygen and this reaction can be followed by monitoring the decrease of the UV absorption bands of the ABMDMA [62][63][64][65].
Solutions of mTHPC and mTHPC@apoMb were exchanged with deuterated PBS. An amount of 97 µL of each sample (mTHPC and mTHPC@apoMb, 1 µM) and 3 µL of ABMDMA 5 mM in DMSO were loaded into the wells of a 96-multiwell plate. The plate was then irradiated using a cold white LED (Valex 30 W, 6500 K), (irradiance 24 mW cm −2 , energy fluence = 86 J cm −2 , measured with the photo-radiometer Delta Ohm LP 471 RAD on the plate surface). The absorbance before and after the irradiation was measured at 380 nm using an EnSpire ® Multimode Plate Reader (PerkinElmer, Waltham, MA, USA).
Amplex Red assay. The Amplex Red assay was used to quantify the peroxides generated by the irradiated samples. It works by monitoring the enzymatic reaction, catalyzed by HRP, that occurs between the nonfluorescent Amplex Red with peroxides, producing fluorescent resorufin [28,62,[65][66][67].
An amount of 90 µL of each sample (mTHPC and mTHPC@apoMb, 1 µM) was loaded into the wells of two 96-multiwell plates. One plate was irradiated under the same conditions used for the ABMDMA assay, whereas the other was kept in the dark. A working solution (WS) containing Amplex Red and HRP dissolved in phosphate buffer 50 mM was freshly prepared. An amount of 10 µL of the WS was then added to each well and both plates were kept in incubation for 30 min in dark conditions at room temperature.
The emission of the resorufin was recorded at 590 nm (λ ex 530 nm) using the EnSpire ® Multimode Plate Reader (PerkinElmer). The fluorescence values were converted to the H 2 O 2 concentration using a calibration curve created with standard solutions of H 2 O 2 . The H 2 O 2 concentration generated by irradiated samples was obtained by subtracting the contribution of the H 2 O 2 produced by the corresponding samples kept in the dark.

Cell Line
The HNSCC cells line, CAL27, derived from squamous carcinoma of the oral tongue, was kindly gifted by the laboratory of Experimental Radiotherapy, Leuven, Belgium. The cells were propagated in culture in Dulbecco's Modified Eagle Medium (DMEM) high glucose, supplemented with 10% heat-inactivated fetal bovine serum, 1% L-glutamine solution 200 mM, and 1% penicillin/streptomycin solution 100 U mL −1 (all supplied by Euroclone, Pero, Italy). The cells were cultured in incubator at 37 • C in a 5% CO 2 humidified atmosphere. To maintain exponential growth, the cells were trypsinized using the cell dissociation solution, Versene (Merck, Darmstadt, Germany), before reaching 80% confluence.

Cell Viability
CAL27 cells were treated in complete medium with increasing concentrations of mTHPC DMSO stock solution diluted in complete medium (mTHPC/CM) or mTHPC@apoMb (0.01-1.00 µM) for 45 min. When the incubation time was over, the cells were washed twice in PBS 1X and irradiated in PBS 1X with white light LED (24 mW cm −2 ) for 45 min. In order to verify possible dark toxicity, the cells were exposed to mTHPC/CM or mTHPC@apoMb while kept in the dark. After the cells were recovered for 24 h in complete medium, the cell viability was determined using a colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) assay (Merck) [68,69].

Identification of Blood Proteins as Carriers for mTHPC by Virtual Screening
We applied a reverse docking approach to identify blood transport proteins that are suitable to bind mTPHC. Following the classifications of the blood proteins proposed by Schaller [52], we first built a structural database of blood transport proteins, and then we carried out ensemble docking calculations to determine the affinity of mTHPC with the different proteins. To investigate in detail the binding between mTHPC and the most promising proteins, we carried out 100 ns of molecular dynamics (MD) simulation, followed by MM-GBSA analysis. Table 1 presents the proteins that bind the most to mTHPC, as ranked by the reverse docking program employed. We recently experimentally demonstrated the possibility to use HSA as a carrier for mTHPC [68]. Docking calculations are commonly used to identify the favourite binding site of a ligand in HSA [70,71]. HSA ranks fifth, which means that, potentially, hemoglobin, myoglobin, hemopexin, and afamin are even more suitable transport proteins for mTHPC. Very interestingly, hemoglobin and myoglobin were already used as a delivery system for PS [36][37][38][39][40][41][42][43][44][45]. In particular, the empty heme-binding pocket of apohemoglobin and hapomyoglobin, obtained via removal of the heme groups, can bind hydrophobic molecules such as PS, maintaining them in a monomeric state and providing water solubility via supramolecular interactions [36][37][38][39][40][41][42][43][44][45]. Docking calculations suggested that the same approach can be used for mTHPC. The docked structures show that mTHPC occupies the heme-binding sites, which are large enough to also host this molecule that is larger than the heme group ( Figure 1).
for PS [36][37][38][39][40][41][42][43][44][45]. In particular, the empty heme-binding pocket of apohemoglobin and hapomyoglobin, obtained via removal of the heme groups, can bind hydrophobic molecules such as PS, maintaining them in a monomeric state and providing water solubility via supramolecular interactions [36][37][38][39][40][41][42][43][44][45]. Docking calculations suggested that the same approach can be used for mTHPC. The docked structures show that mTHPC occupies the heme-binding sites, which are large enough to also host this molecule that is larger than the heme group ( Figure 1). Human hemoglobin (Hb), the main protein found in red blood cells (RBCs), is responsible for the transportation of oxygen (O2) and carbon dioxide (CO2) in blood. In physiological conditions, Hb is a tetramer (α2β2, 64 kDa) of noncovalently linked chains, generally two α-and two β-chains. In Hb, the two α subunits and the two β subunits are arranged into two dimeric halves (αβ) [52]. The folding of α and β chains is typical of globin proteins, characterized, respectively, by the presence of seven and eight α-helices. There are 141 amino-acid residues in α-globin and 146 in β-globin. The missing residues correspond to the D helix of β-globin [52]. The helices pack together into a small, tightlypacked globule, which heterodimerizes to create the usual quaternary shape of Hb. Each subunit is characterized by a central hydrophobic pocket, i.e., the heme-binding sites [52]. mTHPC occupies the heme-binding pocket, superimposing perfectly with the crystallographic structure of the heme (Figure 2A). The favourite binding site is located in the α-subunit, consistent with the fact that α chains in apohemoglobin dimers αβ have a higher affinity for heme than β chains [72]. The calculated ΔEbinding of mTHPC to Hb is −55.4 kcal mol −1 . The driving force of the binding are van der Waals interactions ( Figure  2B). Heme-binding pockets are hydrophobic environments, rich in aromatic and aliphatic amino acids. The presence of a hydrophobic pocket surrounding the heme is necessary for two reasons: (i) to provide a favorable environment to bind the hydrophobic heme, and (ii) to bind oxygen reversibly without going through oxidation or other undesired Human hemoglobin (Hb), the main protein found in red blood cells (RBCs), is responsible for the transportation of oxygen (O 2 ) and carbon dioxide (CO 2 ) in blood. In physiological conditions, Hb is a tetramer (α 2 β 2 , 64 kDa) of noncovalently linked chains, generally two αand two β-chains. In Hb, the two α subunits and the two β subunits are arranged into two dimeric halves (αβ) [52]. The folding of α and β chains is typical of globin proteins, characterized, respectively, by the presence of seven and eight α-helices. There are 141 amino-acid residues in α-globin and 146 in β-globin. The missing residues correspond to the D helix of β-globin [52]. The helices pack together into a small, tightlypacked globule, which heterodimerizes to create the usual quaternary shape of Hb. Each subunit is characterized by a central hydrophobic pocket, i.e., the heme-binding sites [52]. mTHPC occupies the heme-binding pocket, superimposing perfectly with the crystallographic structure of the heme (Figure 2A). The favourite binding site is located in the α-subunit, consistent with the fact that α chains in apohemoglobin dimers αβ have a higher affinity for heme than β chains [72]. The calculated ∆E binding of mTHPC to Hb is −55.4 kcal mol −1 . The driving force of the binding are van der Waals interactions ( Figure 2B). Heme-binding pockets are hydrophobic environments, rich in aromatic and aliphatic amino acids. The presence of a hydrophobic pocket surrounding the heme is necessary for two reasons: (i) to provide a favorable environment to bind the hydrophobic heme, and (ii) to bind oxygen reversibly without going through oxidation or other undesired processes. The aromatic ring of mTHPC is buried among the hydrophobic amino acids of the protein (Val62, Leu86, Leu91, Val93) that usually bind the hydrophobic portions of the heme ( Figure 2C,D). In Hb, anchoring of the heme is facilitated by a histidine nitrogen, His 58 (distal), that coordinates the iron. A second histidine is near the bound oxygen, His 87 (proximal). In mTHPC, the iron atom and the oxygen molecule are absent; the distal and proximal histidines are not involved in coordination and can interact with mTHPC via hydrogen bonds (electrostatic interactions in the MM-GBSA model).
Myoglobin (Mb) is an O 2 -binding hemoprotein, responsible for the storage of oxygen in muscles [52]. The stronger affinity for oxygen of myoglobin, with respect to hemoglobin, determines the diffusion of oxygen from the blood capillaries to the muscular tissues [52]. Mb is a single-chain monomeric protein composed of 153 amino acids (~17 kDa). Additionally, myoglobin shows the typical globin fold, consisting of eight α-helices [52]. Despite the difference in the kind and number of amino acids present, Mb shares almost identical secondary and tertiary structures with the α and β subunits of Hb. Similarities also includes the location and the chemical characteristics of the heme-binding pocket. In fact, both the total binding energy of mTHPC to Mb (∆E binding = −53.3 kcal mol −1 ) and the energy components of ∆E binding strongly resemble those of Hb ( Figure 3). The aromatic ring of mTHPC is accommodated in the same way of the heme (Figure 3), sandwiched by π-π interactions with Phe43 and Tyr103 and hydrophobic interactions with the aliphatic side chains of Val68, Ala71, Ile99, and Ile107 residues. The two distal (His64) and proximal (His93) histidine residues interact via hydrogen bonding with mTHPC ( Figure 3).
The rigid structure of apoHb and apoMb, characterized by a specific, narrow, and well-defined heme-binding site, represents a great potential in bio-applications, especially as a natural protein carrier. processes. The aromatic ring of mTHPC is buried among the hydrophobic amino acids of the protein (Val62, Leu86, Leu91, Val93) that usually bind the hydrophobic portions of the heme ( Figure 2C,D). In Hb, anchoring of the heme is facilitated by a histidine nitrogen, His 58 (distal), that coordinates the iron. A second histidine is near the bound oxygen, His 87 (proximal). In mTHPC, the iron atom and the oxygen molecule are absent; the distal and proximal histidines are not involved in coordination and can interact with mTHPC via hydrogen bonds (electrostatic interactions in the MM-GBSA model). Myoglobin (Mb) is an O2-binding hemoprotein, responsible for the storage of oxygen in muscles [52]. The stronger affinity for oxygen of myoglobin, with respect to hemoglobin, determines the diffusion of oxygen from the blood capillaries to the muscular tissues [52]. Mb is a single-chain monomeric protein composed of 153 amino acids (~17 kDa). Additionally, myoglobin shows the typical globin fold, consisting of eight α-helices [52]. interactions with Phe43 and Tyr103 and hydrophobic interactions with the aliphatic side chains of Val68, Ala71, Ile99, and Ile107 residues. The two distal (His64) and proximal (His93) histidine residues interact via hydrogen bonding with mTHPC ( Figure 3).
The rigid structure of apoHb and apoMb, characterized by a specific, narrow, and well-defined heme-binding site, represents a great potential in bio-applications, especially as a natural protein carrier. The third position of the rank is occupied by hemopexin (Hx). Human plasma hemopexin is a monomeric protein consisting of 439 amino acids (~60 kDa) [74]. The heme-binding site is located between two similar four-bladed β-propeller domains connected by an interdomain linker peptide [74]. Hemopexin is the protein with the The third position of the rank is occupied by hemopexin (Hx). Human plasma hemopexin is a monomeric protein consisting of 439 amino acids (~60 kDa) [74]. The heme-binding site is located between two similar four-bladed β-propeller domains connected by an interdomain linker peptide [74]. Hemopexin is the protein with the highest affinity to heme among known proteins (K d < 1 pM). Hx binds to the free hemes present in the blood, protecting against their cell-damaging effects and promoting their detoxification. After binding, hemopexin transports heme to the liver, which, after the interaction with specific receptors, releases it for breakdown and iron recovery [74]. The composition and location of the heme-binding pocket confers hemopexin high-affinity, yet reversible, binding to the heme. The high affinity for heme is obtained by the simultaneous iron coordination by two histidines (the binding site of Hx is rich in histidines), hydrophobic packing, and steric complementarity around the heme [74]. A notable characteristic of the heme-binding pocket in Hx is the preponderance of aromatic residues: of the eight hydrophobic residues forming the pocket, seven are aromatic [74]. Relative movements of the two domains and/or of the linker peptide connecting them break up the heme-binding pocket, causing its release [74]. It is the tridimensional organization of the two β-propeller Pharmaceutics 2023, 15, 919 9 of 17 domains, connected by an interdomain linker peptide, that defines a modular binding pocket, which contributes both to the high-affinity binding and/or release of the heme [74]. mTHPC binds in the heme-binding site of Hx ( Figure S1), with a ∆E binding of 45.7 kcal mol −1 , strongly interacting with aromatic residues, such as Tyr227, Phe228, and Tyr254, and histidines (His79, His236, His293) (Figure 4). notable characteristic of the heme-binding pocket in Hx is the preponderance of aromatic residues: of the eight hydrophobic residues forming the pocket, seven are aromatic [74]. Relative movements of the two domains and/or of the linker peptide connecting them break up the heme-binding pocket, causing its release [74]. It is the tridimensional organization of the two β-propeller domains, connected by an interdomain linker peptide, that defines a modular binding pocket, which contributes both to the high-affinity binding and/or release of the heme [74]. mTHPC binds in the heme-binding site of Hx ( Figure S1), with a ΔEbinding of 45.7 kcal mol −1 , strongly interacting with aromatic residues, such as Tyr227, Phe228, and Tyr254, and histidines (His79, His236, His293) ( Figure 4).
The binding of porphyrins and PS to hemopexin has already been observed, suggesting the feasibility to use this protein as an mTHPC carrier [75][76][77][78].  The binding of porphyrins and PS to hemopexin has already been observed, suggesting the feasibility to use this protein as an mTHPC carrier [75][76][77][78].
The fourth position of the rank is occupied by afamin. Afamin resembles serum albumin in both structure ( Figure S2) and functionality [52]. It is a carrier for hydrophobic molecules in body fluids, in particular for vitamin E [79]. It is a single-chain protein consisting of 578 amino acids. It has a molecular weight of 87 kDa, but in contrast to albumin, it is highly glycosylated, and treatment with N-glycanase reduces its MW to 65 kDa. The predicted binding site of mTHPC in afamin is the same identified for HSA, i.e., the cleft region ( Figure S2).
Additionally, the ∆E binding between mTHPC and either one of the two proteins is similar (∆E binding = 31.0 kcal mol −1 for afamin and ∆E binding = 36.4 kcal mol −1 for HSA).
Van der Waals interaction between the residues present in the cleft and mTHPC occur; in particular, mTHPC strongly interacts with Gln179, Ala183, Tyr187, Thr424, Thr428, and Leu442 ( Figure 5). The dispersion of mTHPC by human serum albumin (HSA) has already been discussed from a computational and experimental point of view [68]. Of course, the calculated binding energy of mTHPC with afamin and HSA is considerably lower than Hb, Mb, and Hx because these proteins possess a specific pocket for heme-binding. the cleft region ( Figure S2).
Additionally, the ΔEbinding between mTHPC and either one of the two proteins is similar (ΔEbinding = 31.0 kcal mol −1 for afamin and ΔEbinding = 36.4 kcal mol −1 for HSA).
Van der Waals interaction between the residues present in the cleft and mTHPC occur; in particular, mTHPC strongly interacts with Gln179, Ala183, Tyr187, Thr424, Thr428, and Leu442 ( Figure 5). The dispersion of mTHPC by human serum albumin (HSA) has already been discussed from a computational and experimental point of view [68]. Of course, the calculated binding energy of mTHPC with afamin and HSA is considerably lower than Hb, Mb, and Hx because these proteins possess a specific pocket for hemebinding.

Synthesis and Characterization of the mTHPC@apoMb Complex
To demonstrate the possibility of using the blood transport proteins as mTHPC carriers, we selected myoglobin for experimental studies. The calculated affinity of mTHPC to Hb and Mb, as well as the structure of the heme-binding proteins, is practically the same, but it is simpler to handle myoglobin due to its monomeric nature. The prosthetic heme groups of Hb and Mb may be synthetically removed [61,80], producing, respectively, apohemoglobin (apoHb) and apomyoglobin (apoMb), heme-binding apoproteins with high affinity to the heme. ApoHb and apoMb have already been used in biomedical applications as delivery systems [81], in catalysis [82,83], and in nanobiotechnology [84]. Here, we prepared apomyoglobin (apoMb) using a standard procedure of extraction of the heme group with 2-butanone in acid conditions [61]. The heme is non-covalently bound in myoglobin (b-type heme); after acidification, its interaction with the protein is strongly reduced due to the protonation of its carboxylic groups. The transfer of the neutral heme from the aqueous solution to the organic phase was observed upon the addition of 2-butanone. To reconstitute myoglobin, a solution of mTHPC in 2-butanone was added to the apoMb. Alkalization of the solution induces deprotonation of the phenolic group of mTHPC and migration of the negatively charged mTHPC from the organic phase to the aqueous solution, leading to the reconstitution of myoglobin with mTHPC as a guest (mTHPC@apoMb). The mTHPC@apoMb complex was then purified by dialysis. The UV-Vis spectrum of mTHPC@apoMb ( Figure 6) showed the success in the reconstitution of the Mb. In fact, it revealed the typical absorption band of the protein (280 nm). It also evidenced the diagnostic absorption bands of mTHPC. They are the Soret band at 422 nm and the four Q bands in the range of 500-680 nm.
prosthetic heme groups of Hb and Mb may be synthetically removed [61,80], producing, respectively, apohemoglobin (apoHb) and apomyoglobin (apoMb), heme-binding apoproteins with high affinity to the heme. ApoHb and apoMb have already been used in biomedical applications as delivery systems [81], in catalysis [82,83], and in nanobiotechnology [84]. Here, we prepared apomyoglobin (apoMb) using a standard procedure of extraction of the heme group with 2-butanone in acid conditions [61]. The heme is non-covalently bound in myoglobin (b-type heme); after acidification, its interaction with the protein is strongly reduced due to the protonation of its carboxylic groups. The transfer of the neutral heme from the aqueous solution to the organic phase was observed upon the addition of 2-butanone. To reconstitute myoglobin, a solution of mTHPC in 2-butanone was added to the apoMb. Alkalization of the solution induces deprotonation of the phenolic group of mTHPC and migration of the negatively charged mTHPC from the organic phase to the aqueous solution, leading to the reconstitution of myoglobin with mTHPC as a guest (mTHPC@apoMb). The mTHPC@apoMb complex was then purified by dialysis. The UV-Vis spectrum of mTHPC@apoMb ( Figure 6) showed the success in the reconstitution of the Mb. In fact, it revealed the typical absorption band of the protein (280 nm). It also evidenced the diagnostic absorption bands of mTHPC. They are the Soret band at 422 nm and the four Q bands in the range of 500-680 nm. mTHPC@apoMb in PBS displays the typical UV-Vis spectrum of mTHPC monomers ( Figure 7A, solid red line), keeping the well-defined shape of the Soret band (422 nm) and the four Q-bands (500-680 nm) [85].
On the other hand, the broad and red-shifted Soret band, centered at 430 nm, which appears in the spectrum of the dispersion of mTHPC in PBS from a stock solution in  The photosensitizing and imaging capabilities of mTHPC are significantly influenced by the aggregation state of mTHPC. The aggregated state dramatically reduces the fluorescence of mTHPC and its ability to produce ROS, due to quenching phenomena that occur in the aggregates [86][87][88]. Conversely, mTHPC monomers exhibit a significantly higher fluorescence emission and are unaffected by quenching [86][87][88].
The typical emission of mTHPC at 655 nm is observed for the mTHPC@Mb complex ( Figure 7B), where mTHPC is monomolecularly dispersed, whereas an isoabsorbing solution of mTHPC in PBS does not show any emission signal due to aggregation (Figure On the other hand, the broad and red-shifted Soret band, centered at 430 nm, which appears in the spectrum of the dispersion of mTHPC in PBS from a stock solution in DMSO ( Figure 7A, dashed blue line), revealed that in this formulation, the majority of mTHPC exists in aggregated form. The photosensitizing and imaging capabilities of mTHPC are significantly influenced by the aggregation state of mTHPC. The aggregated state dramatically reduces the fluorescence of mTHPC and its ability to produce ROS, due to quenching phenomena that occur in the aggregates [86][87][88]. Conversely, mTHPC monomers exhibit a significantly higher fluorescence emission and are unaffected by quenching [86][87][88].
The typical emission of mTHPC at 655 nm is observed for the mTHPC@Mb complex ( Figure 7B), where mTHPC is monomolecularly dispersed, whereas an isoabsorbing solution of mTHPC in PBS does not show any emission signal due to aggregation ( Figure 7B). The electrophoretic analysis of mTHPC@apoMb on agarose gel in native conditions clearly shows that the protein contains encasulated mTHPC. Whereas mTHPC alone does not show any emission signal (line A, Figure S3) due to aggregation, in the mTHPC@apoMb complex, the fluorescent spot of mTHPC superimposes with the spot of the protein (line B, Figure S3) following Coomassie staining, demonstrating the simultaneous presence of mTHPC and protein.

Evaluation of the PDT Performances of the mTHPC@apoMb Complex
The effect of the encapsulation inside myoglobin on the photosensitizing properties of mTHPC was evaluated using 9,10-Anthracenediyl-bis(methylene)dimalonic acid, AB-MDMA, and Amplex Red assays ( Figure 8). These tests determine, respectively, the generation of singlet oxygen (type II mechanism) and peroxides (type I mechanism) upon irradiation. The photosensitizing and imaging capabilities of mTHPC are significantly influenced by the aggregation state of mTHPC. The aggregated state dramatically reduces the fluorescence of mTHPC and its ability to produce ROS, due to quenching phenomena that occur in the aggregates [86][87][88]. Conversely, mTHPC monomers exhibit a significantly higher fluorescence emission and are unaffected by quenching [86][87][88].
The typical emission of mTHPC at 655 nm is observed for the mTHPC@Mb complex ( Figure 7B), where mTHPC is monomolecularly dispersed, whereas an isoabsorbing solution of mTHPC in PBS does not show any emission signal due to aggregation ( Figure  7B). The electrophoretic analysis of mTHPC@apoMb on agarose gel in native conditions clearly shows that the protein contains encasulated mTHPC. Whereas mTHPC alone does not show any emission signal (line A, Figure S3) due to aggregation, in the mTHPC@apoMb complex, the fluorescent spot of mTHPC superimposes with the spot of the protein (line B, Figure S3) following Coomassie staining, demonstrating the simultaneous presence of mTHPC and protein.

Evaluation of the PDT Performances of the mTHPC@apoMb Complex
The effect of the encapsulation inside myoglobin on the photosensitizing properties of mTHPC was evaluated using 9,10-Anthracenediyl-bis(methylene)dimalonic acid, ABMDMA, and Amplex Red assays ( Figure 8). These tests determine, respectively, the generation of singlet oxygen (type II mechanism) and peroxides (type I mechanism) upon irradiation. Figure 8. Quantification of ROS generated through (A) type I mechanism using Amplex Red assay for H2O2 detection and (B) type II mechanism using ABMDMA assay for 1 O2 detection. Figure 8. Quantification of ROS generated through (A) type I mechanism using Amplex Red assay for H 2 O 2 detection and (B) type II mechanism using ABMDMA assay for 1 O 2 detection.
The performances of the mTHPC@apoMb are significantly higher than mTHPC. The encapsulation of mTHPC improves both type I and type II ROS generation mechanisms.
In the case of type II mechanism, (i) the monomolecular dispersion of mTHPC inside the protein eliminates the well-known phenomenon of aggregation-caused quenching (ACQ), typical of many photosensitizers [22]; and (ii) the rapid quenching of the singlet oxygen produced by water molecules is reduced by the hydrophobic environment [89] of the heme-binding pocket, improving the production of 1 O 2 by~300%.
In the case of the type I mechanism, an additional factor boosts the generation of ROS, which is the presence of electron-rich protein residues in Mb that take part in the electron transfer process necessary to generate oxygen radicals. The type I mechanism is generally activated by sacrificial electron donors; in the case of the mTHPC@apoMb complex, the process is self-activated by the protein itself [22,33,[62][63][64][66][67][68].
We compared the cytotoxic and phototoxic potential of mTHPC@apoMb complex with mTHPC/CM, i.e., mTHPC dissolved in DMSO and diluted in complete medium (mTHPC/CM) (Figure 9). CAL27 cells were selected because they are one of the most frequently used cell lines in the field of head and neck cancer, for which mTHPC is approved. monodispersed formulation of mTHPC@apoMb, whereas we recorded a significant decrease in cell viability after treatment with the mTHPC/CM at all tested concentrations ( Figure 9). These cytotoxic effects may be due to the formation of aggregates that impaired cell viability. Upon irradiation, the phototoxicity of mTHPC/CM is negligible, whereas the phototoxic effects of mTHPC@myo significantly increase in a concentration-dependent manner (Figure 9). The calculated IC50 (inhibitory concentration causing 50% of cell death) of the irradiated mTHPC@apoMb is 0.24 μM. In conclusion, we used virtual screening to identify blood transport proteins able to bind mTHPC. Apohemeproteins, such as apohemoglobin and apomyoglobin; the heme "scavenger" protein hemopexin; and two of the proteins responsible for the transport of hydrophobic molecules in the blood, i.e., afamin and human serum albumin, were identified. We developed a synthetic procedure able to replace the heme group with mTHPC in myoglobin. The protein confers to mTHPC (i) water solubility, (ii) monodispersity, (iii) and stability in physiological environments. We then validated the computational results using apomyoglobin as a true "Trojan horse" inside cancer cells. Figure 9. Cell viability of CAL27 cells after 24 h from 45 min treatment with increasing concentrations of mTHPC/CM (black bars) or mTHPC@apoMb (red bars), and in the presence or absence of 45 min of light irradiation. Statistical significance was calculated by one-way ANOVA followed by Dunnett's multiple comparisons test or t-test. * p < 0.05, ** p < 0.001; **** p < 0.0001 compared to untreated cells; ## p < 0.01 compared to the same concentration of treatment with mTHPC/CM after light exposure.
In dark conditions, we observed the absence of cytotoxic effects for the monodispersed formulation of mTHPC@apoMb, whereas we recorded a significant decrease in cell viability after treatment with the mTHPC/CM at all tested concentrations (Figure 9). These cytotoxic effects may be due to the formation of aggregates that impaired cell viability. Upon irradiation, the phototoxicity of mTHPC/CM is negligible, whereas the phototoxic effects of mTHPC@myo significantly increase in a concentration-dependent manner (Figure 9). The calculated IC 50 (inhibitory concentration causing 50% of cell death) of the irradiated mTHPC@apoMb is 0.24 µM.
In conclusion, we used virtual screening to identify blood transport proteins able to bind mTHPC. Apohemeproteins, such as apohemoglobin and apomyoglobin; the heme "scavenger" protein hemopexin; and two of the proteins responsible for the transport of hydrophobic molecules in the blood, i.e., afamin and human serum albumin, were identified. We developed a synthetic procedure able to replace the heme group with mTHPC in myoglobin. The protein confers to mTHPC (i) water solubility, (ii) monodispersity, (iii) and stability in physiological environments. We then validated the computational results using apomyoglobin as a true "Trojan horse" inside cancer cells. The mTHPC@apoMb complex represents a novel protein-based phototheranostic platform that preserves the imaging properties of mTHPC and, at the same time, improves its ability to generate ROS via both type I and type II mechanisms. In vitro experiments demonstrated that the new mTHPC@apoMb formulation is biocompatible and more effective than mTHPC in PDT.

Supplementary Materials:
The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/pharmaceutics15030919/s1, Figure S1: Superposition of the crystallographic structure of heme and docked structure of mTHPC in Hx; Figure S2: mTHPC-binding pocket in afamin and in HSA; Figure S3: The agarose gel electrophoresis of mTHPC and mTHPC@apoMb.