A Supramolecular Nanoassembly of Lenvatinib and a Green Light-Activatable NO Releaser for Combined Chemo-Phototherapy

The chemotherapeutic Lenvatinib (LVB) and a nitric oxide (NO) photodonor based on a rhodamine antenna (RD-NO) activatable by the highly compatible green light are supramolecularly assembled by a β-cyclodextrin branched polymer (PolyCD). The poorly water-soluble LVB and RD-NO solubilize very well within the polymeric host leading to a ternary supramolecular nanoassembly with a diameter of ~55 nm. The efficiency of the NO photorelease and the typical red fluorescence of RD-NO significantly enhance within the polymer due to its active role in the photochemical and photophysical deactivation pathways. The co-presence of LVB within the same host does not affect either the nature or the efficiency of the photoinduced processes of RD-NO. Besides, irradiation of RD-NO does not lead to the decomposition of LVB, ruling out any intermolecular photoinduced process between the two guests despite sharing the same host. Ad-hoc devised Förster Resonance Energy Transfer experiments demonstrate this to be the result of the not close proximity of the two guests, which are confined in different compartments of the same polymeric host. The supramolecular complex is stable in a culture medium, and its biological activity has been evaluated against HEP-G2 hepatocarcinoma cell lines in the dark and under irradiation with visible green light, using LVB at a concentration well below the IC50. Comparative experiments performed using the polymeric host encapsulating the individual LVB and RD-NO components under the same experimental conditions show that the moderate cell mortality induced by the ternary complex in the dark increases significantly upon irradiation with visible green light, more likely as the result of synergism between the NO photogenerated and the chemotherapeutic.


Introduction
Lenvatinib (LVB) is a multikinase inhibitor (MKI) exerting anti-malignant effects by inhibiting many tyrosine kinase receptors such as vascular endothelial growth factor (VEGF) and fibroblast growth factor receptors [1][2][3][4]. LVB has proven to be more efficacious than sorafenib in unresectable hepatocellular carcinoma (HCC) [5,6]. However, the poor aqueous solubility under physiological conditions [7], the low bioavailability [8], and the serious side effects, including hypertension, nausea, vomiting, weight loss, edema, and rash [9], represent some of the main limitations of this drug. Furthermore, analogously to many chemotherapeutics, multidrug resistance (MDR) remains a major hurdle for its long-term anti-cancer activity [10]. Therefore, the development of new approaches is very challenging for both improving the therapeutic efficacy of LVB and reducing its side effects.
The combination of chemotherapy with light-triggered treatment modalities as a suitable strategy has been receiving increasing attention, especially over the last decade.

Materials
LVB mesylate was purchased from MedChemExpress (HY-10981, Madrid, Spain). All other chemicals were purchased by Sigma-Aldrich and used as received. PolyCD was prepared by cross-linking β-CD with epichlorohydrin, under strongly alkaline conditions,

Materials
LVB mesylate was purchased from MedChemExpress (HY-10981, Madrid, Spain). All other chemicals were purchased by Sigma-Aldrich and used as received. PolyCD was prepared by cross-linking β-CD with epichlorohydrin, under strongly alkaline conditions, following a described method [35]. A mixture of different molecular weight compounds was obtained in the polycondensation reaction. Separation was performed with size exclusion chromatography (SEC) with water-compatible high-performance columns using pullulan standards. The β-CD content in PolyCD was~70% w/w, as determined based on NMR spectra. RD-NO and NBF were synthesized according to our previously reported procedures [34,36]. The concentration of LVB was determined by UV-Vis spectroscopy, using a molar extinction coefficient of 72.350 M -1 cm -1 in MeOH at λ = 241 nm and 62.100 M -1 cm -1 at λ = 245 nm when complexed within PolyCD. All solvents used for the spectrophotometric studies were at spectrophotometric grade. MilliQ water was used for the solubilization of the polymer and for the chemical and photochemical experiments.

Instrumentation
UV-Vis absorption and fluorescence emission spectra were recorded with a JascoV-560 spectrophotometer and a Spex Fluorolog-2 (mod. F-111) spectrofluorimeter, respectively, in air-equilibrated solutions, using either quartz cells with a path length of 0.1 cm or 1 cm. Fluorescence lifetimes were recorded with the same fluorimeter equipped with a TCSPC Triple Illuminator. The samples were irradiated by pulsed diode excitation sources (LED) at 455 nm, and the decays were monitored at 590 nm. The system allowed the measurement of fluorescence lifetimes with a time resolution of 200 ps. The multiexponential fit of the fluorescence decay was obtained using the following equation: I(t) = Σα i exp(−t/τi) Absorption spectral changes were monitored by irradiating the sample in a thermostated quartz cell (1 cm path length, 3 mL capacity) under gentle stirring, using a continuum laser with λ exc = 532 nm,~100 mW and having a beam diameter of~1.5 mm.
Direct monitoring of NO release in solution was performed by amperometric detection (World Precision Instruments, Sarasota, Fl, USA), with an ISO-NO meter equipped with a data acquisition system, and based on direct amperometric detection of NO with short response time (<5 s) and sensitivity range 1 nM-20 µM. The analog signal was digitalized with a four-channel recording system and transferred to a PC. The sensor was accurately calibrated by mixing standard solutions of NaNO 2 with 0.1 M H 2 SO 4 and 0.1 M KI according to the reaction: Irradiation was performed in a thermostated quartz cell (1 cm path length, 3 mL capacity) using the continuum laser with λ exc = 532 nm. NO measurements were carried out under stirring with the electrode positioned outside the light path in order to avoid NO signal artifacts due to photoelectric interference on the ISO-NO electrode.
The size distribution was determined on dynamic light scattering Horiba LS 550 apparatus equipped with a diode laser with a wavelength of 650 nm.

NO Photorelease and Fluorescence Quantum Yields
NO photorelease quantum yield, Φ NO , was determined at λ exc = 532 nm within the 20% transformation of RD-NO by using the following equation: where [tr-RD-NO] is the concentration of phototransformed RD-NO, V is the volume of the sample, t is the irradiation time, A is the absorbance of the sample at the excitation wavelength, and I the intensity of the excitation light source. The concentration of the phototransformed RD-NO was determined by UV-Vis absorption, monitoring the absorption changes at λ = 380 nm and using a ∆ε 380 = 7100 M -1 cm -1 . I was calculated by potassium ferrioxalate actinometry.
The value of fluorescence quantum yield, Φ f , was determined using optically matched solutions at the excitation wavelength of RD-NO within the PolyCD and Rhodamine B in EtOH as a standard (Φ f = 0.79) [37] and calculated through the following equation: where Φ f (s) is the fluorescence quantum yield of the standard; I and I(s) are the areas of the fluorescence spectra of RD-NO and the standard, respectively; n and n(s) is the refraction index of the solvents used for compound and standard. Absorbance at the excitation wavelength was less than 0.1 in all cases.

Preparation of the Supramolecular Complexes
A stock solution of LVB, RD-NO, or both components in MeOH was utilized, and the solvent was evaporated under reduced pressure at 25 • C to form a thin film. The resulting film was rehydrated with 3 mL of phosphate buffer solution (PBS) 10 −2 M, pH 7.4 of PolyCD (2 mg mL −1 ) by stirring for 5 h at room temperature to obtain the binary complexes, PolyCD/LVB and PolyCD/RD-NO, and the ternary complex PolyCD/LVB/RD-NO. For the biological assays and some of the chemical evaluations, the complexes were prepared by rehydrating the films with a solution of PolyCD (2 mg mL −1 ) in DMEM high glucose (without phenol red), as detailed in the biological experiment section.

Determination of IC 50
For the determination of the IC 50 , a stock solution of LVB (10 mM) was prepared in DMSO, and it was diluted in an FBS-free medium in order to have scalar dilutions for in vitro experiments (0.06; 0.13; 0.78; 1.56; 3.125; 6.25; 12.5; 25; 50; 100 µM). Control culture received an equal amount of vehicles (DMSO). The incubation was carried out for 24 or 4 h. The short-term incubation of 4 h was followed by an additional 24 h of cell growth in a drugfree medium with serum (4 h +24 h) and then exposed to 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) salt for the determination of cell viability.

Cell Viability
Cells were detached from the bottom of flasks by trypsin-EDTA solution (Sigma-Aldrich, Milan, Italy), counted in a Bürker chamber, and plated into 96-well plates (Euro-Clone, Milan, Italy) at a density of 1 × 10 4 for 48 h until the monolayer reached a confluency of approximately 80%. For in vitro evaluation, PolyCD was dissolved in an FBS-free culture medium (2 mg mL −1 ). LVB was loaded into the PolyCD by dissolving the thin film obtained after drying a methanol solution with a solution of PolyCD (2 mg mL −1 in FBS-free culture medium) under 5-h magnetic stirring in the dark, reaching a final LVB concentration of 50 µM. To obtain PolyCD loaded only with RD-NO or with both LVB and RD-NO, films of RD-NO were dissolved in the proper volume of PolyCD alone or PolyCD loaded with LVB in order to obtain a final concentration of RD-NO of 6 µM. Cells were exposed to PolyCD alone or in combination with LVB in the presence and absence of RD-NO for 4 h prior to irradiation with light. Cells were then washed, and the phenol red-free and FBS-free medium were added. The experimental plates were divided into two experimental groups: one was maintained in the dark, while the other was exposed to a 150 W Xe lamp (the measured power at 30 cm away from the lamp where the samples were placed was~50-60 mW) through a 500 nm cut-off filter for 10, 20, or 30 min. In each group, untreated cells were maintained as a control. After the irradiation, the FBS-free medium was removed, and the complete medium was added. All the plates were placed back into the incubator. After 24 h, the cellular metabolic activity was assessed by MTT assay. Briefly, after the established time, the medium was replaced with 180 µL of medium and 20 µL of MTT solution (stock solution at 5 mg mL −1 ), and the wells were left incubated for 3 h at 37 • C. Under this condition, mitochondrial dehydrogenases of viable cells conversed the tetrazolium ring of MTT into formazan crystals that were solubilized in 100 µL of DMSO. The absorbance was measured at λ = 570 nm in a microplate reader (AMR-100, Allsheng, Zhejian, China). The cytotoxicity index was calculated using the control cells as 100% viability, and the IC 50 was extrapolated from the dose-response graph by nonlinear regression analysis using GraphPad Prism Software version 7.00 (GraphPad Software, La Jolla California USA). Cell viability after light exposure was calculated as percentage ± SD with respect to untreated cells kept in the dark.

Results and Discussion
CD branched polymers represent an intriguing class of scaffolds for the supramolecular assembling of multiple guests since they offer the possibility of interactions with diverse binding sites, such as the 3D macromolecular network and the hydrophobic CD cavities [38][39][40][41]. This feature not only enhances the solubility of hydrophobic guests but also avoids their self-association propensity [38][39][40][41]. This aspect is crucial in the case of photoresponsive compounds since it may preclude any response to light due to selfquenching phenomena. Furthermore, the modular character of supramolecular assembling does not require any additional synthetic efforts, permits the easy regulation of the relative concentrations of multiple guests, and facilitates the delivery process in a biological environment [42].
On these bases, we chose the branched polymer PolyCD as a suitable host and carrier system for LVB and RD-NO. This polymer consists of β-CD units interconnected by epichlorohydrin spacers to form glyceryl cross-linked β-CD polymer. It is highly soluble in a water medium, well-tolerated in vivo, and able to entrap a variety of guests with enhanced stability constants and payloads when compared with the unmodified β-CD. Moreover, it has been extensively used in our group to integrate a number of phototherapeutics and effectively deliver them into cancer cells [43][44][45][46].
LVB is poorly soluble in PBS solution with a solubility limit of~4.1 µM (~1.7 µg mL −1 ). Solubility studies were performed by using a concentration of PolyCD of 2 mg mL −1 (see experimental). Figure 1 shows a significant enhancement of the drug solubility in the presence of the polymer, reaching a limiting value of~60 µM (~25.5 µg mL −1 ), which is more than one order of magnitude than that observed in the absence of the polymer. Interestingly, apart from a slight red shift of the absorption maxima due to the localization of the drug in an environment of different polarity, the absorption profile of the encapsulated LVB is very similar to that observed in the methanol solution (f in Figure 1A). This finding rules out any intra-aggregation of the drug, which is instead present mainly in its monomeric form. The complex is stable for several days under ambient conditions, as demonstrated by the unchanged position and absorbance values of the absorption spectrum.
LVB is poorly soluble in PBS solution with a solubility limit of ~4.1 µM (~1.7 µg mL −1 ). Solubility studies were performed by using a concentration of PolyCD of 2 mg mL −1 (see experimental). Figure 1 shows a significant enhancement of the drug solubility in the presence of the polymer, reaching a limiting value of ~60 µM (~25.5 µg mL −1 ), which is more than one order of magnitude than that observed in the absence of the polymer. Interestingly, apart from a slight red shift of the absorption maxima due to the localization of the drug in an environment of different polarity, the absorption profile of the encapsulated LVB is very similar to that observed in the methanol solution (f in Figure  1A). This finding rules out any intra-aggregation of the drug, which is instead present mainly in its monomeric form. The complex is stable for several days under ambient conditions, as demonstrated by the unchanged position and absorbance values of the absorption spectrum. In the frame of our interest in photoactivatable NO releasers, we have recently developed the compound RD-NO (see Scheme 1) in which a rhodamine antenna is covalently linked to an N-nitroso appendage [34]. Excitation of the rhodamine chromophore with the highly biocompatible green light triggers the NO release through an intramolecular electron transfer without precluding the typical red fluorescence In the frame of our interest in photoactivatable NO releasers, we have recently developed the compound RD-NO (see Scheme 1) in which a rhodamine antenna is covalently linked to an N-nitroso appendage [34]. Excitation of the rhodamine chromophore with the highly biocompatible green light triggers the NO release through an intramolecular electron transfer without precluding the typical red fluorescence emission of rhodamine, which is useful for imaging [34]. This compound is not soluble in a water medium but is soluble in the presence of the PolyCD polymer both without and with LVB. Figure 2A shows the absorption spectra of the polymeric host loaded with both RD-NO and LVB and, for comparison, those of the polymer loaded with the single guest components at the same concentration. The spectrum of the ternary complex (a in Figure 2A) matches very well the spectral profile obtained by summing the spectra of the binary complexes obtained by loading the polymeric host with the single components (b and c in Figure 2A). This accounts for the absence of relevant interactions between RD-NO and LVB in the ground state. Furthermore, the unchanged position of all the absorption maxima indicates that the presence of RD-NO does not induce any rearrangement of LVB within the polymer and vice versa. Dynamic light scattering measurements indicate an average hydrodynamic diameter for the supramolecular assembly of about 50 nm in the presence of both guests ( Figure 2B), a value in excellent agreement with those already observed for the same polymer co-encapsulating different guests [43][44][45][46]. LVB in the ground state. Furthermore, the unchanged position of all the absorption maxima indicates that the presence of RD-NO does not induce any rearrangement of LVB within the polymer and vice versa. Dynamic light scattering measurements indicate an average hydrodynamic diameter for the supramolecular assembly of about 50 nm in the presence of both guests ( Figure 2B), a value in excellent agreement with those already observed for the same polymer co-encapsulating different guests [43][44][45][46]. As shown in Figure 2, LVB absorption falls in the UV spectral region and does not interfere with the typical absorption band of RD-NO in the visible region centered at 570 nm. This spectral condition permits the selective excitation of RD-NO with the highly biocompatible green light. Figure 3A shows the spectral absorption changes observed upon irradiation of the PolyCD/LVB/RD-NO ternary complex at 532 nm. They show the formation of a new absorption at 397 nm and bleaching at 291 nm, accompanied by the presence of a clear isosbestic point at 348 nm, indicative of a clean photochemical reaction. This photochemical profile is very similar to that previously observed for the free RD-NO in methanol/water [34] and is in agreement with the light-triggered release of NO. Note that both the efficiency and nature of the photochemical process are not influenced by the co-presence of LVB, as demonstrated by the identical absorbance changes observed for the RD-NO-loaded PolyCD in the absence of LVB (see inset Figure 3A) and the lack of any spectral modification in the correspondence of the absorption band of LVB at 245 nm. As shown in Figure 2, LVB absorption falls in the UV spectral region and does not interfere with the typical absorption band of RD-NO in the visible region centered at 570 nm. This spectral condition permits the selective excitation of RD-NO with the highly biocompatible green light. Figure 3A shows the spectral absorption changes observed upon irradiation of the PolyCD/LVB/RD-NO ternary complex at 532 nm. They show the formation of a new absorption at 397 nm and bleaching at 291 nm, accompanied by the presence of a clear isosbestic point at 348 nm, indicative of a clean photochemical reaction. This photochemical profile is very similar to that previously observed for the free RD-NO in methanol/water [34] and is in agreement with the light-triggered release of NO. Note that both the efficiency and nature of the photochemical process are not influenced by the co-presence of LVB, as demonstrated by the identical absorbance changes observed for the RD-NO-loaded PolyCD in the absence of LVB (see inset Figure 3A) and the lack of any spectral modification in the correspondence of the absorption band of LVB at 245 nm.  NO photorelease from the ternary and, for comparison, from the binary supramolecular assemblies was demonstrated by the direct detection of NO. Figure 3B shows the NO release profile observed through amperometric detection, which clearly indicates that NO is produced under green light excitation, stops in the dark, and restarts as the light is turned on again. According to the identical photochemical kinetics (inset Figure 3A), the NO photorelease efficiency was not affected by the presence of LVB. Quantitative analysis of the photolysis process revealed that the polymeric nanocarrier has a profound effect on the quantum yield of NO generation. A value of Φ NO = 4 ± 0.5 × 10 −3 was, in fact, observed under these conditions, which is~4-fold larger than that observed for the free RD-NO [34]. This larger photoreactivity is not uncommon for radical-mediated photodecomposition occurring within polymer environments. In our case, it may be the result of the active role the polymeric network plays in providing easily abstractable hydrogens close to the anilinyl radical intermediate involved in the mechanism of the NO photorelease. This has already been reported for a similar nitroso-derivative NOPD encapsulated in polymeric hosts [36].
RD-NO combines NO photorelease regulated by green light to remarkable fluorescence emission, which is useful for cell tracking by fluorescence imaging. Figure 4A shows the typical fluorescence emission spectrum of RD-NO observed after its encapsulation within PolyCD in the presence of LVB. The fluorescence quantum yield was Φ f = 0.40, which is the same as that observed in the absence of LVB but almost two-fold larger than that observed for the free guest RD-NO (Φ f = 0.22) [34]. This indicates a reduction of the non-radiative decay pathways, probably due to a restriction of the free rotation motion of the fluorophore embedded within the polymeric network. Accordingly, the fluorescence dynamic Figure 4B) showed a bi-exponential decay with a shorter lifetime τ 1 = 1.4 ns, the same as what was observed for the free guest [34], and a longer component with τ 1 = 3.6 ns, with the relative weight of~40% and 60%, respectively.  The preservation of the photochemical properties of a photoactivatable component, once co-encapsulated with another guest in the same host, is not a trivial result and is a fundamental requisite for combining photo-with chemo-therapeutic effects. Potential photoinduced energy/electron transfer between the two guests in close proximity can indeed lead to undesired processes, which preclude the final goal. In our case, the photochemical and photophysical behaviors of RD-NO are even improved after its encapsulation in the host nanocarrier and are not affected by the co-presence of LVB. This finding rules out any quenching by the chemodrug on the excited states of RD-NO, accounting for not close proximity of the two guests, which are probably confined in different locations of the polymeric host. To gain more insight into this important aspect, The preservation of the photochemical properties of a photoactivatable component, once co-encapsulated with another guest in the same host, is not a trivial result and is a fundamental requisite for combining photo-with chemo-therapeutic effects. Potential photoinduced energy/electron transfer between the two guests in close proximity can indeed lead to undesired processes, which preclude the final goal. In our case, the photochemical and photophysical behaviors of RD-NO are even improved after its encapsulation in the host nanocarrier and are not affected by the co-presence of LVB. This finding rules out any quenching by the chemodrug on the excited states of RD-NO, accounting for not close prox-imity of the two guests, which are probably confined in different locations of the polymeric host. To gain more insight into this important aspect, we carried out ad-hoc devised Förster Resonance Energy Transfer (FRET) experiments, which permit to shed light on the spatial proximity of two guests. LVB shows a fluorescence emission at~380 nm (a in Figure 5). However, this emission does not show a significant overlap with the absorption spectrum of RD-NO (b in Figure 5), which is a necessary condition for the occurrence of FRET between these two components. Accordingly, we did not observe any FRET upon excitation of LVB in the presence of RD-NO. This spectral limitation, thus, precludes obtaining information into the relative location of the guests in the host. To overcome this limitation, we used the nitrobenzofurazan derivative NBF ( Figure 5) as a suitable additional guest mediator. This compound is highly hydrophobic, dissolves well within the PolyCD, and exhibits spectral properties that make it an ideal energy acceptor and energy donor towards LVB and RD-NO, respectively. In fact, LVB emission significantly overlaps the NBF absorption (see spectra a and c in Figure 5), and the NBF fluorescence emission significantly overlaps with the RD-NO absorption (see spectra d and b in Figure 5).  Figure 6A). This reflects w absorption spectra of every single component, ruling out any sign between them in the ground state. In the first experiment, we recorde emission spectrum of the sample loaded with all three guests by exci energy donor) at 240 nm, and we compared it with that of a similar sam of NBF (potential energy acceptor). As shown in Figure 6B, no quenchi was observed, suggesting the absence of FRET between LVB an accounting for their not close proximity in the host. In a second experi LVB, NBF, and RD-NO (all at the same concentration) can be simultaneously co-loaded in the PolyCD (2 mg mL −1 ), as demonstrated by the absorption spectrum showing the typical spectral features of all guests ( Figure 6A). This reflects well the sum of the absorption spectra of every single component, ruling out any significant interaction between them in the ground state. In the first experiment, we recorded the fluorescence emission spectrum of the sample loaded with all three guests by exciting LVB (potential energy donor) at 240 nm, and we compared it with that of a similar sample in the absence of NBF (potential energy acceptor). As shown in Figure 6B, no quenching of LVB emission was observed, suggesting the absence of FRET between LVB and NBF and thus accounting for their not close proximity in the host. In a second experiment, we recorded the fluorescence emission spectrum of the sample loaded with all three guests by exciting NBF (potential energy donor) at 445 nm (a in Figure 6C), and we compared it with that of similar samples loaded with either NBF (potential energy donor) alone (b in Figure 6C) or RD-NO (potential energy acceptor) alone (c in Figure 6C). The results obtained show a significant quenching of the NBF fluorescence in the presence of RD-NO accompanied by a corresponding increase in the RD-NO emission, in accordance with an effective FRET process between these two guests and accounting for their close proximity. Therefore, taken together, all these results will confirm that the therapeutic components LVB and RD-NO are confined in different compartments of the polymeric host and very distant from each other. This feature allows the photochemical and photophysical properties of RD-NO within the nanocarrier to be preserved even in the presence of LVB, which, in turn, is not decomposed upon light excitation of RD-NO (i.e., by undesired photoinduced bimolecular processes), thus preserving its chemotherapeutic action. Note that the FRET experiments reported in Figure 6C were well reproduced in the presence of a culture medium (5% in serum) for several days, confirming the stability of the supramolecular complex even under these experimental conditions. The excellent emissive properties of RD-NO in the host nanocarrier and their preservation, even after co-encapsulation with LVB, allowed for exploring the cellular uptake of the supramolecular assembly by fluorescence microscopy. Figure 7 shows representative fluorescence images of HEP-G2 hepatocarcinoma cell lines obtained after 4 h incubation with PolyCD loaded with RD-NO, LVB, or both components. It is evident that in all cases, the typical red fluorescence originating from RD-NO accounts for a localization mainly at the cytoplasmatic level, and it is not affected by the presence of LVB. The excellent emissive properties of RD-NO in the host nanocarrier and their preservation, even after co-encapsulation with LVB, allowed for exploring the cellular uptake of the supramolecular assembly by fluorescence microscopy. Figure 7 shows representative fluorescence images of HEP-G2 hepatocarcinoma cell lines obtained after 4 h incubation with PolyCD loaded with RD-NO, LVB, or both components. It is evident that in all cases, the typical red fluorescence originating from RD-NO accounts for a localization mainly at the cytoplasmatic level, and it is not affected by the presence of LVB.
preservation, even after co-encapsulation with LVB, allowed for exploring the uptake of the supramolecular assembly by fluorescence microscopy. Figure 7 representative fluorescence images of HEP-G2 hepatocarcinoma cell lines obtain 4 h incubation with PolyCD loaded with RD-NO, LVB, or both components. It is that in all cases, the typical red fluorescence originating from RD-NO accoun localization mainly at the cytoplasmatic level, and it is not affected by the presence Cell mortality experiments were carried out with the above-mentioned cell a preliminary experiment, we determined the IC50 value of the chemod investigating the cell viability as a function of the free LVB concentration in a ser Cell mortality experiments were carried out with the above-mentioned cell lines. In a preliminary experiment, we determined the IC 50 value of the chemodrug by investigating the cell viability as a function of the free LVB concentration in a serum-free medium. Under our experimental condition, the 4 h incubation with LVB did not alter cell viability, and 24 h treatment had a limited role in inducing mitochondrial dysfunctions. According to these results, the IC 50 was 132.6 ± 26.8 µM at 24 h and 572.1 ± 235.6 µM at 4 h + 24 h (Figure 8).
Pharmaceutics 2023, 15, x FOR PEER REVIEW medium. Under our experimental condition, the 4 h incubation with LVB did not a viability, and 24 h treatment had a limited role in inducing mitochondrial dysfu According to these results, the IC50 was 132.6 ± 26.8 µM at 24 h and 572.1 ± 235.6 h + 24 h (Figure 8). The biological activity of the ternary supramolecular assembly was tested visible light excitation and in the dark and, for the sake of comparison, was comp that of the binary complexes containing the individual either LVB or RD-NO identical experimental conditions. These experiments were performed with LV concentration of 50 µM, which is well below the IC50, and RD-NO at a concentrat µM with irradiation not longer than 30 min, which allows generating NO moderately cytotoxic doses. The overall results are illustrated in Figure 9.  The biological activity of the ternary supramolecular assembly was tested under visible light excitation and in the dark and, for the sake of comparison, was compared to that of the binary complexes containing the individual either LVB or RD-NO under identical experimental conditions. These experiments were performed with LVB at a concentration of 50 µM, which is well below the IC 50 , and RD-NO at a concentration of 6 µM with irradiation not longer than 30 min, which allows generating NO only at moderately cytotoxic doses. The overall results are illustrated in Figure 9.
The biological activity of the ternary supramolecular assembly was tested under visible light excitation and in the dark and, for the sake of comparison, was compared to that of the binary complexes containing the individual either LVB or RD-NO under identical experimental conditions. These experiments were performed with LVB at a concentration of 50 µM, which is well below the IC50, and RD-NO at a concentration of 6 µM with irradiation not longer than 30 min, which allows generating NO only at moderately cytotoxic doses. The overall results are illustrated in Figure 9. The PolyCD polymeric host is well tolerated both in the dark and under light irradiation. Besides, under the chosen experimental conditions, LVB-only loaded PolyCD showed negligible toxicity in the dark and a moderate reduction in cell viability only with longer irradiation time. In contrast, PolyCD loaded with the photoactive RD-NO induced levels of toxicity depending on the irradiation time according to the production of moderate NO doses. Interestingly, a significant reduction in cell viability was observed in the case of the ternary complex containing both LVB and RD-NO. The potentiation indexes (% viability free LVB/% viability polyCD/LVB/RD-NO) were~3, 5, and 8 at 10, 20, and 30 min of irradiation, respectively. It should be emphasized that since the presence of LVB in the supramolecular assembly does not affect the NO photorelease efficiency (see Figure 3B), the enhancement of cell mortality cannot be due to higher NO concentration photogenerated by the complex. Therefore, the occurrence of a notable additive/synergistic action between the chemotherapeutic and the NO appears clear. A possible explanation for this synergism can be related to recent studies that have demonstrated that MKI, including LVB, induces the generation of reactive oxygen species (ROS) in human HCC cell lines in vivo and in vitro [47,48]. In our case, the formation of ROS, such as superoxide anion, may lead to a diffusion-controlled reaction with NO, generating the highly cytotoxic peroxynitrite [49]. Investigations addressed to gain insights into the mechanisms at the basis of such synergistic action definitely deserve attention.

Conclusions
Our results have demonstrated that the anti-cancer activity of LVB enhances by its combination with an NO photoreleaser through a supramolecular complex based on a biocompatible polymeric host nanocarrier. This latter (i) increases the solubility of LVB by more than one order of magnitude with respect to that observed for the free drug in water, (ii) enhances the photochemical and photophysical performances of the NO photoreleaser and (iii) represents an ideal container, which permits the confinement of the two guests in different compartments, avoiding any intermolecular communication between them and, thus, preserving their individual properties.
The synergistic action between LVB used at a concentration well below its IC 50 and the NO generated at moderated cytotoxic doses through the appropriate light tuning validate the combination of photoregulated NO with MKI as a suitable approach to enhance their anti-cancer action. Compared with our previous work on sorafenib, the present work represents a step forward in terms of biocompatibility of excitation light (from blue to green) to activate the NO delivery, with a gain of more than 100 nm towards lower energy. Furthermore, the modular character of the supramolecular assembling of both therapeutic components does not require any additional synthetic efforts, permitting the easy regulation of the relative concentrations of the two guests and facilitating the delivery process in a biological environment.
Finally, the use of spontaneous NO releasers in combination with MKI has also been recently proposed as a suitable strategy to exploit the role of NO as a vasodilator able to reduce hypertension [50,51], which is one of the main side effects of this class of chemotherapeutics [9]. In this view, and considering the well-known critical dichotomous biological effects of NO strictly depending on its concentration [23][24][25][26][27], our findings may also open up interesting avenues in the perspective of in vivo applications not only to enhance the anti-cancer activity of MKI but also to reduce one of their major side effects.