Lignin Nanoparticles Deliver Novel Thymine Biomimetic Photo-Adducts with Antimelanoma Activity

We report here the synthesis of novel thymine biomimetic photo-adducts bearing an alkane spacer between nucleobases and characterized by antimelanoma activity against two mutated cancer cell lines overexpressing human Topoisomerase 1 (TOP1), namely SKMEL28 and RPMI7951. Among them, Dewar Valence photo-adducts showed a selectivity index higher than the corresponding pyrimidine-(6-4)-pyrimidone and cyclobutane counterpart and were characterized by the highest affinity towards TOP1/DNA complex as evaluated by molecular docking analysis. The antimelanoma activity of novel photo-adducts was retained after loading into UV photo-protective lignin nanoparticles as stabilizing agent and efficient drug delivery system. Overall, these results support a combined antimelanoma and UV sunscreen strategy involving the use of photo-protective lignin nanoparticles for the controlled release of thymine dimers on the skin followed by their sacrificial transformation into photo-adducts and successive inhibition of melanoma and alert of cellular UV machinery repair pathways.


Introduction
Exposure to UV radiation represents a risk factor for photo-aging, epigenetic changes, suppression of the immune system, angiogenesis, nucleobase mutation, and emergence of melanoma [1]. UV radiation excites DNA pyrimidine bases, and in particular thymine [2], to corresponding singlet and triplet energy states yielding cyclobutane pyrimidine (CPD), pyrimidine-(6-4)-pyrimidone , and Dewar Valence (DV) photo-adducts [3]. CPDs are three to four times more frequent in comparison with other photo-adducts [4]. UV-B is absorbed directly by DNA, while UV-A induces the formation of reactive radical species responsible for the oxidation of nucleobases and activation of pheomelanin electron transfer damage pathways [5,6]. To prevent genome mutations, the cell activates a complex network of repair pathways, including photo-damage recognition [7], cell cycle arrest [8], photoadduct excision, and apoptosis [9,10]. The insurgence of melanoma can occur when these pathways are not effective [11]. In this context, the amount of photo-adducts produced during UV damage plays a key role [12]. The catabolic way of photo-adducts is still a matter of debate [9]. Recent reports suggest that they are released inside short oligonucleotides (30 mer) in a tight complex with the Transcription Factor II Human (TFIIH) and Replication Protein A (RPA) [9,13]. These oligonucleotides undergo a limited degree of degradation and can activate Mitogen-Activated Protein Kinase (MAPK) and checkpoint pathways involving tumor protein 53 (p53), Cyclin Dependent Kinases (CDK) inhibitor factor and p21, respectively, for the prevention of UV-damage, or in alternative, apoptosis [14]. In addition, they stabilize human Topoisomerase 1 (TOP1), affecting transcription bubble and replication fork [15]. Stable Topoisomerase Cleavage Complex (TOP1cc) forms near UV lesions preventing re-ligation processes [16,17]. In particular, TOP1 is trapped by oligonucleotide sequences containing photo-adducts with generation of Single-Strand Breaks (SSBs) [18]. Melanoma is one of the most aggressive forms of skin cancer, with limited therapeutic options. Since its incidence has been rapidly rising in recent years, the study of selective therapy has grown of interest [19]. The implication of nanoscience in the development of targeted therapies for melanoma shows multiple benefits and could significantly improve the outcome of melanoma patients [19,20]. Nanoparticles from natural sources are extensively investigated as drug delivery systems thanks to several benign factors, such as controlled release properties, good biocompatibility, biodegradability, and multifunctional and UV shielding properties [21]. May a combined strategy including the controlled release of thymine dimers and photo-adducts from nanoparticles be effective in the protection of the cell from UV damage and insurgence of melanoma?

Summary and Concise Description of the Study
We synthesized CPD, (6-4)PP, and DV photo-adducts from biomimetic thymine dimers characterized by a simple carbon side-chain spacer between the nucleobases. Pyrimidine dimers resembling thymidine residues have been previously reported as simple mimetics of DNA sequences [22,23], associated with (in a preliminary test) sunscreen protection against UV damage [24] (Scheme 1, panel A). Photo-adducts have been loaded inside lignin nanoparticles (LNPs) with high loading efficiency and capacity values and successively characterized for their stability and releasing properties, followed by the evaluation of anti-melanoma activity (Scheme 1, panel B). As a result, we observed an unprecedented antimelanoma activity of thymine photo-adducts against two mutated cell lines overexpressing TOP1, namely SKMEL-28 and RPMI7951. DV photo-adducts were more selective than PP and CPD counterparts. The antimelanoma activity of these compounds was retained after loading into LNPs as a stabilizing and photo-protective drug delivery system (Scheme 1, panel C). In addition, docking computational studies showed the high stability of the interaction between DV photo-adducts and DNA/TOP1 complex (Scheme 1, panel D).
We performed the photo-addition of 4a in a Pyrex immersion well reactor (furnished of a refrigerator) with Haerus source 250 Watt (5.0 cm distance from the reactor; emission spectrum in Figure S1) in degassed deionized water (MilliQ)/acetone mixture (8:2 v/v, 100 mL) under argon atmosphere for 2.0 h at 25 • C and 0 • C. The irradiation of 4a (0.17 mmol) at 25 • C quantitatively converted the substrate to afford CPD 6a, (6-4)PP 7a, and DV 8a in low yield (Scheme 3; Table 2, entry 1) besides polar products not recovered under our experimental conditions. CPD 6a was deprived of the adsorption band at 260 nm (C=C double bond), showing a λ max value of 191 nm. This value was in accordance with data previously reported for CPD bearing 2 -deoxy-ribose/phosphate [38], or in alternative, a phosphoramidide-like backbone [39].  As revealed by the 1 H-NMR analysis, 6a was isolated as a mixture of trans-sin isomer (TS) (H-6 multiplet signals at 4.40 ppm and 4.27 ppm, respectively) and cis-sin isomer (CS) (H-6 singlet signal at 4.36 ppm) in 4.0:1.0 TS/CS ratio (Scheme 3, Table 2). The antiisomer was not detected, probably due to steric effects of the spacer. Compounds 7a and 8a showed the expected 1 H NMR signals at 7.48 ppm (H-6, singlet) and 5.32 ppm (H-6 , singlet), and 5.31 ppm (H-6, singlet) and 6.92 ppm (β-lactame bridge hydrogen, singlet), respectively. The mechanism of formation of (6-4)PP and DV photo-adducts is reported, focusing on the initial formation of an oxetane intermediate and successive intramolecular 4π electro-cyclization of the pyrimidone ring [40]. In order to improve the yield of photo-adducts, the reaction was repeated at 0 • C [27,28] to afford 6a, 7a, and 8a in appreciable yield and quantitative conversion of substrate (Table 2, entry 2). Better results were finally obtained increasing the amount of acetone (water/acetone 6:4 v/v, 100 mL) at 0 • C, in which case photo-adducts were obtained in 76% total yield and quantitative conversion of substrate (Table 2, entry 3). These latter conditions were applied in the transformation of thymine dimer homologues 4b-d, to afford CPDs 6b-d, DVs 8b-d, and (6-4)PPs 7b-c from low to acceptable yield ( Table 2, entries 4-6). Spectroscopical data of novel compounds were consistent with the expected structures, as well as full comparable with data previously reported for photo-adducts deriving from N-1 methyl thymine and thymidine derivatives [27][28][29][30][31][32][33][34][35][36][37]. Again, CPD-TS isomers prevailed with respect to the CS counterpart ( Table 2, entries 4-6) [41][42][43]. As a general trend, the total yield of photo-adducts increased by increasing the carbon length of the spacer. In accordance with data reported for irradiation of DNA [44,45], CPDs were generally obtained as the major reaction products.

Lignin Nanoparticles as a Drug Delivery System for Biomimetic Thymine Dimers and Photo-Adducts
We evaluated the use of lignin nanoparticles (LNPs) as natural carriers for the drug delivery of biomimetic thymine dimers and photo-adducts [46]. LNPs meet the challenge of biocompatibility and photo-stability in drug-delivery systems, increasing the permeability across the skin barrier [47,48]. The capacity of LNPs to adsorb UV-A and UV-B is reported, and the role played by π-interactions between the aromatic sub-units of the polymer in the adsorption process has been deeply investigated [46,49]. In addition, LNPs are potent radical scavengers [21,46,49]. LNPs were prepared starting from Kraft Lignin (KL) by the nanoprecipitation technology [50] that requires the dissolution of the polymer in a primary solvent followed by addition of deionized water (secondary solvent) [51]. After the addition of water, lignin quickly aggregates in order to balance repulsive electrostatic forces, thus yielding stable colloidal nanoparticles that can entrap bioactive substances in their internal cavity [47,52].
Initially, the optimal conditions to solubilize KL and 4a in deionized water and dimethyl-isosorbide (DMI) were evaluated. DMI was selected as a primary solvent due to its green properties, being eco-certified by the European Chemical Agency (ECHA) [53]. In addition, DMI showed beneficial interactions with the epidermis, favoring the diffusion of bioactive compounds inside the corneous layer [54]. Compound 4a (3.4 mM, 6.8 mM, 8.0 mM and 10 mM) and KL (1:5 w/w, 1:10 w/w, and 1:20 w/w, with respect to 4a) were dissolved in DMI/water mixtures (1:1 v/v, 2:1 v/v, 3:1 v/v; total volume 3.0 mL) at 25 • C for 24 h. Irrespective from the concentration of 4a and KL, the optimal DMI/water ratio was 2:1 v/v. Nanoprecipitation was then performed by adding different amounts of deionized water (from 3.0 mL to 24 mL, respectively). Results are reported in Table 3 and Figure 1 and depended on the amount of water we obtained: (i) stable colloidal LNPs (30 days at 25 • C; Figure 1A); (ii) low stable colloidal LNPs (1-2 days at 25 • C; Figure 1B); (iii) aggregates of LNPs ( Figure 1C); (iv) formation of precipitate ( Figure 1D); and (v) absence of LNPs ( Figure 1E). ×14 Absence of LNPs Absence of LNPs Absence of LNPs a General condition for the preparation of LNPs loaded with compound 4a include 3.0 mL of DMI/water 2:1 v/v as primary solvent containing 4a and KL at 25 • C. b The volume of secondary solvent (deionized water) expressed as multiple of the volume of the primary solvent. c 4a/KL w/w ratio (1:5 w/w, 1:10 w/w, and 1:20 w/w values, respectively). In particular, the addition of 3.0 mL of deionized water always produced a precipitate ( Table 3, entry 1), while stable colloidal LNPs were obtained (1:5 w/w and 1:10 w/w 4a/KL ratio, respectively) using both 6.0 mL and 12 mL of deionized water, respectively (Table 3, entry 2 and entry 3).
A further increase in the secondary solvent afforded low stable colloidal LNPs (18 mL; Table 3, entry 4) and aggregates (24 mL; Table 3, entry 5). Finally, LNPs were not formed after the addition of high volumes of deionized water (Table 3, entries 6-8). The optimal conditions affording stable colloidal LNPs (1:5 w/w and 1:10 w/w 4a/KL ratio, 3.0 mL of DMI/water 2:1 v/v, and 6.0 mL of secondary solvent) were successively applied for the encapsulation of 4b-d, and of the most abundant photo-adducts 6a-d and 8a-d. The loading capacity (LC) and loading efficiency (LE) Equations (1) and (2) were measured as a function of the concentration of the encapsulated compound [55][56][57][58].
LNPs were separated by centrifugation, and the supernatant was analyzed by HPLC at λ max value of the encapsulated compound (271 nm for 4a-d, 191 nm for 6a-d, and 192 nm for 8a-d, respectively). LC and LE are reported in Table 4. Irrespective from the experimental conditions, the highest LE and LC were obtained in the presence of the highest value of the compound/KL ratio (compound/KL ratio of 1:5) [59]. These results are of the same order of magnitude than that previously reported for the encapsulation of other bioactive natural substances inside LNPs [60]. As a general trend, 4a-d generally showed LE and LC values higher than that of the corresponding photo-adducts 6a-d and 8a-d (Table 4, entries 1-4 versus entries 5-12). In addition, LE and LC increased by increasing the number of carbon atoms in the spacer (see for example Table 4, entry 4 versus entries 1-3), highlighting the role played by hydrophobicity in the nucleation of lignin and bioactive compounds during the loading process [61,62]. The role of hydrophobicity in the encapsulation of drugs inside lignin nanoparticles is well reported, and the role played by Van der Walls, hydrogen bonding, and π-π stacking interactions in this process have been deeply investigated [63,64].
Loaded and unloaded LNPs showed a regular spherical shape with the surface characterized by a rough aggregate of small clumps (Figure 2, Panel A). Crystals were not present on the surface of loaded particles [65]. In addition, some LNPs/6a and LNPs/8a showed holes in their structure. The presence of these holes may be due to jet collision-like processes involving nanoclusters of lignin and DMI [66] and successive rush-out effects of the solvent during drying [67]. Dynamic light scattering analysis showed monodisperse nanoparticles. Different values of the average diameter were found depending on the presence or absence of loaded compounds. Empty LNPs showed an average particle diameter of 188 nm lower than LNPs/4a (268 nm), LNPs/6a (275 nm), and LNPs/8a (281 nm), respectively (Figure 2, Panel B). In addition, the diameter of LNPs/6a and LNPs/8a was larger than that of the thymine dimer counterpart (compound 4a), in accordance with the expected higher steric hindrance of photo-adducts.

UV Shielding Capacity of Loaded LNPs
The UV shielding capacity of LNPs allows them to protect bioactive compounds from premature deactivation [21]. To evaluate the photo-protective role of LNPs we studied the irradiation of LNPs/4a, LNPs/6a, LNPs/7a, and LNPs/8a as a selected representative example of thymine dimers. The samples were suspended in DMI/water mixture (2:1 v/v, 1.0 mL) and treated with UV under previously reported experimental conditions. At scheduled times (5, 10, and 15 min), aliquots (0.2 mL) of the suspension were withdrawn and analyzed. An excess of DMI (1.0 mL) was added to disaggregate LNPs, followed by filtration of residual lignin. The residual concentrations of 4a, 6a, 7a, and 8a (expressed as % with respect to the starting material) were evaluated at λ max 271 nm. Data were compared with 4a, 6a, 7a, and 8a after irradiation under similar experimental conditions ( Figure 3). Irrespective from the experimental conditions, LNPs showed an excellent photo-protective capacity after 15 min of irradiation, reaching in the case of compound 7a a photo-protection capacity of almost 40%.

Releasing Properties
The cumulative releasing capacity of LNPs was evaluated as previously reported in the literature [68] and normalized with respect to the total amount of entrapped compound. In a typical procedure, the appropriate LNPs (7.5 mg) were suspended in water (3.0 mL) and dialyzed against deionized water (60.0 mL) at two selected pH values (pH 7.4 and 5.5, respectively) at 37.5 • C for 24 h [69]. The amount of released compound was measured by UV-vis spectrophotometry at the optimal λ max value for each compound by sampling the dialysis solution at appropriate time intervals (Figure 4). The withdrawn volume was replaced by corresponding deionized water in order to maintain constant the volume of the solution (Section 3). Firstly, we studied the kinetic release of 4a at λ max 271 nm. LNPs/4a exhibited a fast release of 4a at both pH values during the first 4 h (57% at pH 7.4, and 68% at pH 5.5, respectively), reaching a plateau within 24 h (60% at pH 7.4, and 70% at pH 5.5, respectively) ( Figure 4, Panel A). The overall efficacy of the process was favored at pH 5.5 in accordance with the effect exerted by acidic conditions in the opening of the pores of LNPs [70]. The cumulative releasing capacity of 4b-d was studied at the optimal pH value of 5.5. LNPs/4b-d showed a kinetic release similar to 4a, compound 4d being released in the highest amount (71% within 4 h) (Figure 4, Panel A). The cumulative releasing effect of LNPs/6a-d (Figure 4, Panel B) and LNPs/8a-d (Figure 4, Panel C) at pH 5.5 (measured at optimal λ max 191 nm, and λ max 192 nm, respectively) was similar to that of LNPs/4a-d and higher than LNPs/8a-d.

Biological Activity
Preliminary sunscreen data of 4a-d have been reported focusing on SKH1-E cell line, human abdominal skin explants, and in vivo hairless mouse models [23]. Data for photoadducts 6a-d, 7a-d, and 8a-d are not available. Thymine photo-adducts are well recognized as responsible for the activation of different molecular pathways involved in the cellular response to UV damage [71], including DNA repair system, melanogenesis [72,73], and Nucleotide Excision Repair (NER) system [74][75][76][77][78]. In addition, TOP1ccs trapped near photo-adducts activates BER and remove (6-4)PPs in NER-deficient cells [4]. TOP1 is highly expressed in malignant tumors, including carcinomas of the colon, prostate, ovary, lung, and melanoma. These data associated with the active role played by photo-adducts in melanogenesis [79] open the possibility for these compounds to possess antimelanoma activity. For this reason, we decided to study the antimelanoma activity of 4a-d, 6a-d, and 8a-d against two melanoma cancer cell lines, namely SK-MEL28 and RPMI7951 bearing a different genetic background [80]. The antimelanoma activity was determined using the MTT cell viability assay, following the absorbance change (λ max 570 nm and 630 nm). The human FB789 cell line was used as reference. The CC 50 (cytotoxic concentration that causes death of 50% of viable cells), IC 50 (minimum concentration inhibiting 50% of the melanoma cells), and Selective Index (SI) (ratio between IC 50 of reference cell line versus melanoma cell line) values are reported in Table 5. Compounds were not toxic after 24 h treatment in the FB789 cell line (IC 50 > 100 µg/mL; Table 5, entries 1-11), with the only exception of 8d (Table 5, entry 12). In the case of 4a-d this result was in accordance with data previously reported [23]. Compounds 4a-d were ineffective against melanoma cell lines. Instead, photo-adducts showed appreciable antimelanoma activity and selectivity ( Table 5, entries 1-4). In the case of the CPDs family, 6b showed the highest SI value (Table 5, entry 6 versus entries 5 and 7-8). DVs 8b and 8d showed the highest activity against SK-Mel 28 ( Table 5, entries 10 and 12, respectively), while 8a and 8c were effective against RPMI7951 (Table 5, entries 9 and 11, respectively). In addition, high values of SI were obtained for 8a and 8b, compound 8b being the most selective one with SI value of c.a. 234. Next, the antimelanoma activity of compounds 8a and 8b was evaluated after loading inside LNPs.
The kinetic release of LNPs/8b in the culture medium (Dulbecco's modified eagle medium) is reported in Figure S2 as a selected sample and compared with data referring to the buffer medium at pH 5.5 (Section 2.4). Compound 8b was released faster in DMEM than in buffer, accordingly to data previously reported [81]. LNPs alone showed no toxicity against the FB789 cell line (Table S1). In this latter case, the antimelanoma activity was measured at three different times (2, 4, and 24 h) in order to evaluate the effect of the kinetic release of the compound on the biological activity. LNPs/8a and LNPs/8b retained the antimelanoma activity against both SK-Mel 28 and RPMI7951 cell lines, respectively ( Table 5, entries 13 and 14). Note that the antimelanoma activity of LNPs/8a and LNPs/8b increased during the time. In addition, LNPs/8a and LNPs/8b showed values of SI of the same order of magnitude than 8a and 8b, confirming the efficacy of LNPs in the drug delivery process. Note that the recorded SI values were higher than those, or of the same order of magnitude as, the approved drugs such as camptotechin, cisplatin, and doxorubicin [82,83].

In Silico Molecular Docking Analysis
In silico molecular docking analysis was performed to evaluate binding affinity, binding conformation, and non-covalent interactions between DVs 8a and 8b and TOP1 in complex with 22 base pairs DNA duplex (PDB ID: 1T8I) [84]. Compounds 4a and 4b were used as references. The active site of the TOP1/DNA duplex was characterized by amino acids Arg364 and Asp533, and nucleotides DC112, DA113, and DT10, as the most relevant residues [85]. In order to obtain well-balanced and reliable intermolecular interactions, we performed an energy minimization run of the system using the Amber99 force field [84] provided by Gromacs 2020.3 (www.gromacs.org).
DVs 8a and 8b showed higher TOP1/DNA duplex binding affinity than corresponding dimers 4a and 4b (Table 6, entries 1, 2 versus entries 3, 4), 8b having the highest binding affinity ( Table 6, entry 2). These data are in accordance with the trend of the antimelanoma activity previously reported. The conformations of ligands with the best binding affinity towards the TOP1/DNA duplex are in Figure S3. The non-covalent interactions between compounds 8a-b and 4a-b, and TOP1/DNA duplex, were calculated through the protein-ligand interaction profiler PLIP 2021 [86], by using AutoDock Vina for input poses. Compound 8b ( Table 6, entry 2) showed one H-bonding interaction with DC111, DC112, Asp533, and Thr718 and two H-bonding interactions with Arg364. This compound also formed a salt bridge with Asp533 ( Figure 5, Panel  A). Similar H-bonding interactions were observed for 8a ( Table 6, entry 1). In this latter case, only one H-bonding interaction was observed with Arg364, and no salt interaction was operative with Asp533 ( Figure 5, Panel B). The higher number of non-covalent interactions of 8b with Arg364 and Asp533 may be responsible for the highest antimelanoma observed for this compound. This hypothesis is in accordance with the AutoDock Vina's binding affinity rank. Regarding thymine dimers, 4b (Table 6, entry 4) formed H-bonding interactions with DC112, DA113, and Arg364. Instead, 4a (Table 6, entry 3) interacted with DG12, DC112, Arg364, and Thr718. Hydrophobic interactions were observed only for 4a. Note that neither of the two dimers 4a and 4b (Table 6, entries 3 and 4) formed non-covalent interactions with Asp533, suggesting a lower efficacy of dimers in the interaction with TOP1/DNA duplex with respect to 8a-b ( Figure 5, Panels C and D, respectively).

Preparation and Loading of Lignin Nanoparticles
LNPs were produced and loaded by the nanoprecipitation technique. The appropriate compound (3 mg) was dissolved in DMI/water (3.0 mL) at 50 • C under gentle stirring conditions followed by the addition of KL (15 mg). The fast addition of deionized water (6.0 mL) instantaneously produced LNPs. LNPs were then recovered by centrifugation (7500 rpm), washed with deionized water (10 mL) for 3 times, and lyophilized. The loading capacity (LC) and loading efficiency (LE) were evaluated by UHPLC analysis.

UV Photo-Protective Effect of Loaded LNPs
LNPs/4a were prepared following the standard procedure previously described. Typically, LNPs/4a (1.0 mg) suspended in DMI/water mixture (2:1 ratio, 1.0 mL) were deposited on a glass lens and irradiated with Haerus source (250 Watt, the lamp was at 20.0 cm distance). At scheduled time intervals (5, 10, and 15 min), 0.2 mL of irradiated colloidal suspension were withdrawn and treated with DMI (0.2 mL) to disaggregate LNP structures, allowing the recovery of residual 4a. The photo-protection capacity of LNPs was determined by measuring the decrease in absorbance (λ max 271 nm) of recovered 4a compared with free reference (Varian UVCary-50). The irradiation of 4a (1.0 mg) alone, under similar experimental conditions, was performed as a reference.

Kinetic Release of Lignin Nanoparticles
LNPs (7.5 mg, compound/KL 1:5 ratio) were suspended in water (3.0 mL), placed in a dialysis membrane, and immersed in PBS buffer (60 mL) at two selected pH values (7.4 and 5.5, respectively) at 37.5 • C for 24 h. A similar procedure was applied solubilizing LNPs/8b (7.5 mg) water (3.0 mL), followed by dialysis against the culture medium DMEM (60 mL). At scheduled time intervals (2,5,10,15,20,30,45,60, 120, 180 min and 24 h) aliquots of the solution (0.2 mL) were withdrawn for UV-visible analysis, the same volume of fresh medium being continuously replaced to maintain constant the total volume of the system. The quantification of released compound was conducted by UV-Vis spectrophotometry at the λ max characteristic for each type of compound (271 nm for 4a-d, and 191 and 192 nm for 6a-d and 8a-d, respectively) using semi-micro cuvettes. Water (1.5 mL) was used as a baseline reference.

Biological Assay
FB789 was grown in DMEM/F10, RPMI7951 melanoma cancer cells were grown in Minimal Essential Medium (MEM), while SK-Mel 28 melanoma cancer cells were grown in Dulbecco's Modified Eagle Medium (DMEM). All culture media were supplemented with 10% Fetal Bovine Serum (FBS) plus 1.0 mM glutamine (40 µg/mL). The antimelanoma activity of 4a-d, 6a-d, and 8a-d was evaluated by MTT viability assay according to reference [87]. Briefly, aliquots of the appropriate compound (1.0 mmol) were solubilized in dimethyl sulfoxide (one drops, DMSO) and administered to the selected cell line. After incubation for 3 h at 37 • C with MTT (0.5 mg/mL) the supernatant was removed, and 100 mL of lysis solution (10% SDS and 0.6% acetic acid) was added to dissolve the formazan crystals. Optical density detection was performed by DTX880 Multimode Detector (Beckman Coulter) with 630 nm (background) and 570 nm filters. The percentage of cell viability was calculated as follows: % cell viability = 100-% cell cytotoxicity. The experiments for the evaluation of the antimelanoma activity of LNPs/8a and LNPs/8b were performed in a similar way using the appropriate amount of nanoparticles containing 1.0 mmol of the active compound and analyzing the antimelanoma activity at three different release times (2, 4, and 24 h).

In Silico Molecular Docking Analysis
The crystal structure of TOP1 and 22 base pair DNA duplex (PDB ID: 1T8I) was used in order to perform molecular docking analysis. The energy minimization run of TOP1/DNA duplex was performed by the steepest descent algorithm with a maximum number of minimization steps of 50,000. The complex was centered in a dodecahedron box with a minimal distance of 1.0 nm to the edge of the box, and the TIP3P water model was used to solvate the system by adding 18 sodium ions [88]. Both the receptor and the ligands were prepared for the docking analysis using Autodocktools v. 1.5.6 [89]. The receptor was prepared by removing crystal ligands, adding polar hydrogens and Kollman charges as partial charges. 3D coordinates of 8a, 8b, 4a, and 4b, provided in smile format, were generated by using Open Babel software v. 2.3.2 [90]. Polar hydrogens were added to each ligand, and Gasteiger charges were added as partial charges.
AutoDock Vina software [91] was used to perform docking experiments between the selected compound and TOP1/DNA duplex. A grid box with size 10 × 10 × 10 and centered at X = 94.906 Y = 95.914 Z = 32.500 was used as search space. AutoDock Vina provided the best binding affinity rank for each ligand.

Conclusions
A panel of biomimetic thymine dimers and photo-adducts, differing in the number of carbon atoms in the spacer between nucleobases, was synthesized under photochemical conditions. The highest yield of photo-adducts was obtained in the presence of acetone as photosensitizer at 0 • C. The complete panel of expected products was isolated, including CPD, (6-4)PP and DV photo-adducts. In the case of CPD, the steric constrains of the linker favored the formation of the trans-syn stereoisomer with respect to the cis-syn counterpart. CPD were synthesized in a yield higher than the other products. Due to the reported affinity between pyrimidine photo-adducts and TOP1, the antimelanoma activity of novel derivatives was tested against two mutated cancer cell lines overexpressing TOP1, namely SKMEL-28 and RPMI7951. Among photo-adducts, DV showed higher antimelanoma activity than corresponding (6-4)PP and CPD counterparts, the highest value of SI being associated with compounds 8a and 8b. The SI value of these compounds was higher, or of the same order of magnitude, than commercial antitumor agents, such as doxorubicin (SI 26.0), 5-fluro uracil 5-FU (SI 55.6), and camptothecin (SI 88.3) [92]. The antimelanoma activity of 8a and 8b was retained after loading inside LNPs, which were able to release gradually the bioactive compound in the culture medium, preserving it from photo-degradation. The encapsulation of compounds 8a and 8b was obtained by a sustainable nanoprecipitation procedure affording high loading capacity and efficiency values, as well as optimal kinetic release. Molecular docking analysis suggested the antimelanoma activity of 8a and 8b is correlated to the formation of a stable interaction with TOP1/DNA complex at the cleavage site of the system. Overall, these results open a new entry for the design of a sustainable drug-delivery system combining the photoprotective effect of LNP to the antimelanoma activity of thymine photo-adducts produced after the release of thymine dimers on the skin. The possibility that thymine photo-adducts may also activate specific alert pathways for the cell against UV damage represent a further frontier to be investigated.