**4. Discussion**

The increased polyamine metabolism and its correlation with cancer have prompted many efforts at developing drugs that could inhibit polyamine biosynthesis or block the active, carrier-mediated transport of polyamine into the cells as a therapeutic approach in cancer treatment.

More recently, attention has been paid to molecules that, on the contrary, upregulate polyamine catabolism in cells as a means to generate toxic compounds that can induce cancer cell death [7]. In fact, it has been demonstrated that the production of H2O<sup>2</sup> through polyamine catabolism is implicated in the cytotoxic responses of several types of tumors to different polyamine analogs [8].

Among the molecules currently used to upregulate polyamine catabolism, polyamine analogs, such as N1,N11-diethylnorspermine (DENSpm) and N1-ethyl-N11-[(cyclopropyl)m ethyl]-4,8- diazaundecane (CPENSpm), have demonstrated activity in several types of tumors [9,11,12]. However, these molecules showed some drawbacks, such as poor specificity for cancer cells, induction of epithelial-mesenchymal transition-dedifferentiation in noncancerous cells, and limited activity in clinical trials [11,12].

Therefore, a new approach to stimulate polyamine catabolism might rely on increasing intracellular polyamine concentrations above their physiological values to trigger the production of oxidation metabolites at levels exceeding cell tolerance. However, exogenous administration of free polyamines cannot raise their intracellular concentrations over physiological levels because intracellular penetration is strictly regulated by import-export mechanisms based on specific membrane carriers, such as SLC22A16 and SLC3A2, that import or export polyamines depending on the cell necessity [13]. In addition, the carriers' activity is regulated by a feedback mechanism based on the immediate synthesis of antizyme, a regulatory protein that blocks the polyamine uptake in the presence of small increases in free cellular polyamine levels [14].

Other mechanisms of cell penetration, such as passive diffusion or endocytosis, are ineffective compared to the carrier-mediated active transport. Passive diffusion, in particular, is strongly hindered by the high intracellular polyamine concentrations in the mM range that sustain a negative outside-inside concentration gradient, being the physiological concentration of extracellular polyamines in the µM range. In addition, the extensive protonation of spermidine (pKa 10.9) at physiological pH reduces the concentration of the diffusible, unprotonated molecules to about 0.03% of the total extracellular concentration [14–16].

Furthermore, the extracellular concentration of polyamines is physiologically regulated to prevent accidental increases that could trigger an uncontrolled influx of polyamines into the cells driven by passive diffusion.

The control of extracellular polyamine concentration starts in the intestine where polyamines from the diet or produced by luminal bacteria are absorbed by a saturation regulated mechanism and are metabolized in the enterocytes before reaching the systemic circulation [56].

With the aim to overcome these limitations and obtain supra-physiological levels of spermidine in tumor cells, we prepared nanospermidine by encapsulation of spermidine in nanomicelles that we had previously demonstrated to be suitable carriers for fenretinide in tumor cells [21–24]. Next, we evaluated the effect of nanospermidine in two neuroblastoma cell lines: NLF (with MYCN amplification) and BR6 (without MYCN amplification, with TrKA expression).

A significant decrease in cell viability was obtained in both cell lines with nanospermidine concentrations of 0.15 mg/mL corresponding to 150 µM spermidine. With free spermidine, on the contrary, no cytotoxic effects were observed up to 1200 µM, and a slowing of cell proliferation in NLF and a complete inhibition in BR6 were obtained at 1500 µM.

The results were in agreement with the ability of the nanomicelles to transport the encapsulated spermidine within the cells, previously demonstrated for fenretinide encapsulated in the same nanomicelles [21,22] and supported by confocal laser-scanning fluorescence microscopy showing increased fluorescence in cells treated with Red Nile stained nanospermidine compared to controls treated with the free dye.

Cytotoxicity of empty nanomicelles was excluded by comparative experiments, indicating no cytotoxic effects at nanomicelle concentrations higher than those used in this study. Moreover, the cytotoxicity of nanospermidine was higher in tumor cells than in normal cells, such as WS1 fibroblasts, in accordance with the increased sensitivity of tumor cells to ROS increase compared to their normal counterparts. In fact, it has been demonstrated that cancer cells have ROS levels higher than normal cells, and their redox system is easily overwhelmed by additional ROS production that induces oxidative stress leading to cell death [57,58].

Based on this feature, many strategies are being developed to raise ROS levels and realize targeted therapies by induction of cell death in tumors without relevant toxicity to normal cells [59].

The tenfold higher cytotoxicity of nanospermidine than free spermidine suggested the suitability of the nanomicelles as polyamine transporters and revealed the efficiency of this mechanism to provide the cell with high amounts of spermidine owing to the favorable spermidine loading (12.92% *w*/*w*) in the nanomicelles. On the contrary, the much higher concentration of free spermidine required to trigger cytotoxicity correlates with the difficulty for the free molecule to enter the cells in amounts exceeding physiological values, evading cell transport regulation. These results denote the unsuitability of free spermidine as an antitumor agent because the high concentrations required at the tumor site would be hardly achieved even by the administration of large doses.

Moreover, the results obtained indicate that nanospermidine can be regarded as a new tool to increase intracellular polyamines levels and induce tumor cell death by using spermidine at biocompatible doses. In addition, the in-vivo administration of nanospermidine

could exploit the mechanism of nanomedicines accumulation at the tumor site by extravasation through the discontinuities of the capillaries generated by angiogenesis, providing on-target activity. The ability of nanospermidine to increase its size at high dilutions could represent an additional favorable feature for targeted localization. Indeed, in the presence of high dilutions, such as those provided by the tumor matrix, the size expansion of the extravasated nanoparticles could prevent their retro-diffusion to the venous circulation that, at present, is recognized as a drawback in passive drug targeting of nanomedicines [60].

We found a correlation between nanospermidine cytotoxicity and increased intracellular ROS levels. This prompted us to evaluate the combination of nanospermidine with nanofenretinide, as we had previously demonstrated that nanofenretinide was endowed with high antitumor activity towards these same cell lines [21,22] due to ROS increase.

Nanofenretinide improved the effect of nanospermidine at both cytotoxic and subcytotoxic nanospermidine concentrations. It increased cell death at the cytotoxic concentrations while, at the sub-cytotoxic concentrations, it decreased cell proliferation, as indicated by prolonged duplication times and decreased cell motility.

Furthermore, cotreatment with nanofenretinide at the sub-cytotoxic nanospermidine concentrations slightly increased sphericity and decreased biomass. These slight but important modifications were disclosed by time-lapse QPI techniques that measured several parameters at the same time and provided a multifaced image of the cells. They indicated that, while treatment with nanospermidine at sub-cytotoxic concentrations did not influence cell morphology, cotreatment with nanofenretinide may induce apoptosis, even at such low nanospermidine concentrations, as suggested by the reduced biomass and the increased sphericity indicative of decreased cell attachment. This information could not have been obtained by other techniques such as MTT assay.
