*2.5. The E*ff*ect of Glutathione Responsive* β*-cyclodextrin-Based Nanosponges on Reactive Oxygen Species Production*

After exposure to the respective GSH-NS B and GSH-NS D IC<sup>50</sup> values at 24 h, the dichlorofluoresceindiacetate (DCFH-DA) assay did not indicate a significant intracellular ROS increase at 1, 12 (data not shown), and 24 h (Figure 5) in all cell lines considered, except for HCT116 cells, where a significant increase in intracellular ROS was observed at 24 h (Figure 5A) after both GSH-NS formulations incubation (Figure 5A).

(Figure 6B,C,H,I).

GSH-NS formulations incubation (Figure 5A).

*Production*

*2.5. The Effect of Glutathione Responsive β-cyclodextrin-Based Nanosponges on Reactive Oxygen Species* 

After exposure to the respective GSH-NS B and GSH-NS D IC50 values at 24 h, the dichlorofluorescein-diacetate (DCFH-DA) assay did not indicate a significant intracellular ROS increase at 1, 12 (data not shown), and 24 h (Figure 5) in all cell lines considered, except for HCT116

**Figure 5.** Glutathione responsive β-cyclodextrin-based nanosponge reactive oxygen species (ROS) production. HCT116 (**A**), HT-29 (**B**), DU145 (**C**), and PC-3 (**D**) cells were treated with GSH-NS B or GSH-NS D at the respective IC50 for 24 h. ROS levels, detected by dichlorofluorescein-diacetate (DCFH-DA) assay by flow cytometry, were expressed as the integrated median fluorescence intensity (iMFI) ratio and nanosponge-induced ROS levels were compared to those of control, that is, untreated cells, fixed at 1 and shown by the dotted line. Statistically significant difference versus untreated cells: \*\* *p* < 0.01; \*\*\* *p* < 0.001. **Figure 5.** Glutathione responsive β-cyclodextrin-based nanosponge reactive oxygen species (ROS) production. HCT116 (**A**), HT-29 (**B**), DU145 (**C**), and PC-3 (**D**) cells were treated with GSH-NS B or GSH-NS D at the respective IC<sup>50</sup> for 24 h. ROS levels, detected by dichlorofluorescein-diacetate (DCFH-DA) assay by flow cytometry, were expressed as the integrated median fluorescence intensity (iMFI) ratio and nanosponge-induced ROS levels were compared to those of control, that is, untreated cells, fixed at 1 and shown by the dotted line. Statistically significant difference versus untreated cells: \*\* *p* < 0.01; \*\*\* *p* < 0.001.

#### *2.6. Glutathione Responsive β-cyclodextrin-Based Nanosponge Cytotoxicity in Three-Dimensional Cell 2.6. Glutathione Responsive* β*-cyclodextrin-Based Nanosponge Cytotoxicity in Three-Dimensional Cell Cultures*

*Cultures* The next step was the analysis of GSH-NS cytotoxicity in the colorectal and prostatic cancer cell lines with the highest GSH basal level, HCT116 and DU145, which were three-dimensionally cultured as multicellular spheroids (MCSs), cellular aggregates organized in a specific cell-to-cell and cell– matrix interaction, closer to in vivo features [41]. Dose response curves were obtained by exposing HCT116 and DU145 on 3D model to different concentrations (0.5, 2.0, 4.0, and 6.0 mg/mL) of GSH-NS B and GSH-NS D, for 24, 48, and 72 h to obtain the respective IC1 and IC50 values (Table 3 and Figure 6A,G). The MCS uptake of non-cytotoxic (IC1) and cytotoxic (IC50) concentrations of The next step was the analysis of GSH-NS cytotoxicity in the colorectal and prostatic cancer cell lines with the highest GSH basal level, HCT116 and DU145, which were three-dimensionally cultured as multicellular spheroids (MCSs), cellular aggregates organized in a specific cell-to-cell and cell–matrix interaction, closer to in vivo features [41]. Dose response curves were obtained by exposing HCT116 and DU145 on 3D model to different concentrations (0.5, 2.0, 4.0, and 6.0 mg/mL) of GSH-NS B and GSH-NS D, for 24, 48, and 72 h to obtain the respective IC<sup>1</sup> and IC<sup>50</sup> values (Table 3 and Figure 6A,G). The MCS uptake of non-cytotoxic (IC1) and cytotoxic (IC50) concentrations of fluorescent GSH-NS was confirmed by fluorescence microscope imaging after 24 h of exposure (Figure 6B,C,H,I).


fluorescent GSH-NS was confirmed by fluorescence microscope imaging after 24 h of exposure **Table 3.** GSH-NS IC values in the 3D cell cultures.

bond content, GSH-NS D, was more cytotoxic in DU145 cells than in HCT116 (Figure 2A,C). Figure Statistically significant difference between the two nanosponge formulations: ns.

6 shows images of 72 h untreated HCT116 MCS (10.51 ± 1.74 µm3,panel D) and DU145 MCS (11.66 ± 0.92 µm3, panel J), which are compared to GSH-NS B and GSH-NS D treated HCT116 MCS (4.03 ± 0.09 µm3, panel E and 4.55 ± 0.77 µm3, panel F, respectively) (*p* < 0.05) and DU145 MCS (4.63 ± 0.27 µm3, panel K and 5.12 ± 0.60 µm3, panel L, respectively) (*p* < 0.05). GSH-NS B and D gave similar IC<sup>50</sup> values in both cell lines (Table 3, Figure 6A,G) and across all incubation times, whereas GSH-NS B gave a higher cytotoxic effect in HCT116 cells than GSH-NS D in the monolayer cultures (Table 2 and Figure 2A). Interestingly, the nanosponge cytotoxic effect was time dependent in both cell lines (Figure 6A,G). This also differs from the results obtained in the monolayer cultures (Figure 2A,C). Moreover, significantly higher GSH-NS B and GSH-NS D IC<sup>50</sup> values at 24 h were observed in DU145 cells than in HCT116 cells (Figure 6G), suggesting a lower cytotoxic effect of the GSH-NS at 24 h in DU145 cells than in HCT116 cells. These results differ from those observed in monolayer cultures, where the nanosponge formulation with the higher disulfide-bond content, GSH-NS D, was more cytotoxic in DU145 cells than in HCT116 (Figure 2A,C). Figure 6 shows images of 72 h untreated HCT116 MCS (10.51 ± 1.74 µm<sup>3</sup> , panel D) and DU145 MCS (11.66 ± 0.92 µm<sup>3</sup> , panel J), which are compared to GSH-NS B and GSH-NS D treated HCT116 MCS (4.03 ± 0.09 µm<sup>3</sup> , panel E

and 4.55 ± 0.77 µm<sup>3</sup> , panel F, respectively) (*p* < 0.05) and DU145 MCS (4.63 ± 0.27 µm<sup>3</sup> , panel K and 5.12 ± 0.60 µm<sup>3</sup> , panel L, respectively) (*p* < 0.05). IC1 0.02 ± 0.00 0.04 ± 0.00 IC1 0.03 ± 0.00 0.02 ± 0.00 IC1 0.07 ± 0.01 0.02 ± 0.00 IC50 5.38 ± 1.15 5.80 ± 1.78 IC50 2.68 ± 0.25 3.24 ± 0.34 IC50 2.01 ± 0.24 2.26 ± 0.23 Statistically significant difference between the two nanosponge formulations: ns.

**Table 3.** GSH-NS IC values in the 3D cell cultures. **HCT116 IC Values (mg/mL ± St.Dev) on 3D Cultures 24 h GSH-NS B GSH-NS D 48 h GSH-NS B GSH-NS D 72 h GSH-NS B GSH-NS D** IC1 0.01 ± 0.00 0.02 ± 0.00 IC1 0.01 ± 0.00 0.01 ± 0.00 IC1 0.01 ± 0.00 0.01 ± 0.00 IC50 3.92 ± 0.95 4.19 ± 0.98 IC50 2.75 ± 0.18 3.05 ± 0.31 IC50 1.98 ± 0.15 2.22 ± 0.21 **DU145 IC Values (mg/mL ± St.Dev) on 3D Cultures 24 h GSH-NS B GSH-NS D 48 h GSH-NS B GSH-NS D 72 h GSH-NS B GSH-NS D**

**Figure 6.** Glutathione responsive β-cyclodextrin-based nanosponge cytotoxicity and uptake on 3D culture. HCT116 (**A**) and DU145 (**G**) multicellular spheroids (MCSs) were incubated with GSH-NS B and GSH-NS D at different concentrations (0.5, 2.0, 4.0, and 6.0 mg/mL) and MCS volume was measured after 24, 48, and 72 h. Representative images at 24 h of HCT116 and DU145 MCS uptake of 6-coumarin loaded GSH-NS B IC50 (**B** and **H**, respectively) and of 6-coumarin loaded GSH-NS D IC50 (**C** and **I**, respectively). Representative phase contrast images at 72 h of HCT116 MCS: untreated (**D**), treated with GSH-NS B IC50 (**E**) and GSH-NS D IC50 (**F**). Representative phase contrast images at 72 h of DU145 MCS: untreated (**J**), treated with GSH-NS B IC50 (**K**), and GSH-NS B IC50 (**L**). Images are at **Figure 6.** Glutathione responsive β-cyclodextrin-based nanosponge cytotoxicity and uptake on 3D culture. HCT116 (**A**) and DU145 (**G**) multicellular spheroids (MCSs) were incubated with GSH-NS B and GSH-NS D at different concentrations (0.5, 2.0, 4.0, and 6.0 mg/mL) and MCS volume was measured after 24, 48, and 72 h. Representative images at 24 h of HCT116 and DU145 MCS uptake of 6-coumarin loaded GSH-NS B IC<sup>50</sup> (**B**,**H**, respectively) and of 6-coumarin loaded GSH-NS D IC<sup>50</sup> (**C**,**I**, respectively). Representative phase contrast images at 72 h of HCT116 MCS: untreated (**D**), treated with GSH-NS B IC<sup>50</sup> (**E**) and GSH-NS D IC<sup>50</sup> (**F**). Representative phase contrast images at 72 h of DU145 MCS: untreated (**J**), treated with GSH-NS B IC<sup>50</sup> (**K**), and GSH-NS B IC<sup>50</sup> (**L**). Images are at 10× magnification.

#### 10× magnification. **3. Discussion**

**3. Discussion** Understanding the effect that nanoparticles have on cells is crucial to predict theirin vivo toxicity and avoid any undesirable nanoparticle activities. Although there are numerous in vitro cytotoxicity assays that can be applied for the general screening of nanoparticles [42,43], it is of vital importance that the research covers nanoparticle cytotoxicity itself. In this contest the use of strictly controlled in vitro experimental conditions can ensure that the measured effect is the result of nanoparticle toxicity and not unstable culturing conditions [44]. Moreover, up to nowadays, there has been no single analysis able to provide sufficient information to correlate the biomaterial chemistry and surface with biological response [45]. Herein, we investigated the in vitro biological effects of a stimuli-responsive Understanding the effect that nanoparticles have on cells is crucial to predict their in vivo toxicity and avoid any undesirable nanoparticle activities. Although there are numerous in vitro cytotoxicity assays that can be applied for the general screening of nanoparticles [42,43], it is of vital importance that the research covers nanoparticle cytotoxicity itself. In this contest the use of strictly controlled in vitro experimental conditions can ensure that the measured effect is the result of nanoparticle toxicity and not unstable culturing conditions [44]. Moreover, up to nowadays, there has been no single analysis able to provide sufficient information to correlate the biomaterial chemistry and surface with biological response [45]. Herein, we investigated the in vitro biological effects of a stimuli-responsive nanosystem, that is, glutathione responsive β-cyclodextrin-based nanosponges (GSH-NS), in various cancer cell lines, characterized by their GSH basal content, as this nanosystem is designed to be a GSH responsive anticancer drug carrier.

GSH plays a key role in cellular defense against oxidative stress [46] and its increased redox capacity in cancer cells is well-known [34,47]. Consequently, GSH has been recognized to be an ideal intracellular trigger for selective drug delivery by responsive nanocarriers, as many compounds exert their therapeutic effects only inside cells. As disulfide chemistry is particularly versatile, a wide range of GSH-responsive nano-vehicles, such as micelles, nanoparticles, and nanogels, have been recently developed [28]. Among them, glutathione responsive β-cyclodextrin-based nanosponges incorporate high payload and provide controlled drug release over time, with the further advantage of triggered intracellular drug delivery in response to cell GSH content. In addition, GSH-NSs are able to protect degradable drugs from the external environment. It is foreseen that β cyclodextrin-based nanosponges

will have a significant positive impact on anticancer therapeutic scenarios [13,24,37,38]. Taking into account the promising results concerning the efficacy of GSH-NSs as an anticancer drug delivery system [37,38], the biological safety of the nanosponge itself is a critical parameter for their future clinical application.

β-cyclodextrin toxicology has been evaluated in in vitro and in vivo studies that have reported it as non-toxic and well tolerated even at very high doses [48]. Previous in vitro studies showed no signs of cytotoxicity after cell exposure to unloaded nanosponges in the 10–100 µg/mL concentration range used for the delivery of therapeutic drugs [23,49,50]. In addition, in vivo experiments have shown that β-cyclodextrin-based nanosponges prepared with pyromellitic dianhydride as a cross-linking agent have been orally administered to rats without showing any toxic side effects at selected doses in an acute and repeated dose toxicity study [51]. Previously, GSH-NSs have been investigated as doxorubicin carrier. No acute cardiotoxic effects were observed in mice after the in vivo administration of doxorubicin-loaded GSH-NS [37]. Recently, the hepatotoxicity of this nanoformulation was investigated either in vitro on human HepG2 cell line or ex vivo on rat precision-cut liver slices (PCLSs), where a good nanosponge safety profile was demonstrated, showing a comparable hepatotoxicity to that of free doxorubicin [52].

As no reports have been published on the effects at a cellular level of GSH-NS as such, it was decided to study the effect of GSH-NS per se on HCT116, HT-29, DU145, and PC-3 cancer cell lines with various GSH content in a concentration range that is about fifty times higher than that used in the above mentioned studies to ensure the use of cytotoxic concentrations. HCT116 and DU145 cells showed the highest GSH values in colorectal and prostatic cancer cell lines, respectively; previous research studies have shown that DU145 cells have the highest GSH content [53]. Non-toxic (IC1) and cytotoxic (IC50) GSH-NS concentrations were determined by a 2D cell assay, which measured mitochondrial activity. A decrease in cell growth with significantly different IC<sup>1</sup> or IC<sup>50</sup> values was observed when the two nanosponge formulations were compared in all cell lines, except in DU145 cell line, where no statically significant difference was observed.

DU145 cell line was the most sensitive to the GSH-NS D cytotoxic effect among all cell lines tested. Notably, DU145 cells are more resistant to electrophilic toxicity than other cells owing to their high levels of redox-sensitive transcription factor, nuclear factor erythroid 2-related factor-2 (Nrf2), which activates cytoprotective pathways against oxidative injury, such as GSH synthesis [54,55]. As this nanosystem has the ability to disrupt itself in the presence of GSH, we can hypothesize that it is the high GSH content in DU145 cells that allows GSH-NS to exert their cytotoxic effect, whatever the disulfide-bond concentration. Further studies are needed to investigate whether agents able to modulate intracellular GSH, such as *N*-acetyl cysteine or buthionine sulfoximine [56] could affect nanosponge intracellular drug release and cytotoxicity.

Our study shows that colorectal cancer cells, in particular HCT116 cells, have the most pronounced GSH-NS B and D cellular uptake. This difference in nanosponge cellular uptake in this cell line may be owing to differing uptake mechanisms, as cell surface thiols have been reported to affect disulfide-conjugated peptide cell entry [57]. Indeed, disulfide bridge cleavage may start at the cell surface via thiol/disulfide exchange reactions catalyzed by redox proteins such as thioredoxines [58]. Therefore, the 2D data on IC<sup>50</sup> would appear to indicate that prostatic cancer cell lines are more sensitive to GSH-NS cytotoxic effects.

Worthy of note is that cell cycle analyses revealed a significant cell cycle arrest in the G0/G<sup>1</sup> phase in all cell lines at 24 h IC<sup>50</sup> values. Thus, to further investigate this cell cycle arrest, we analyzed a panel of genes that are involved in cell cycle regulation. Notably, the results show significant mRNA over-expression in the cell cycle progression regulators at G1, *CDKN1A*, and *CDKN2A*, which code for p21 and p16 that inhibit the cyclin-CDK2 and -CDK4 complexes, respectively. Apart from this, the mRNA expression of *CDC25A*, *CDK1*, *CDK2*, and *CDK4* was either unaffected or down-regulated in all cell lines. These data demonstrate that GSH-NS inhibition of cell proliferation is essentially owing to G<sup>1</sup> cell cycle arrest, in agreement with previous reports by Choi et al. [59]. Interestingly, only HTC116 cells showed significant ROS production after GSH-NS exposure, which is most likely owing to their high GSH-NS cellular uptake.

Lastly, the investigation of nanosponges effects on MCS growth was carried out. The results were interesting as differences in the 2D study were observed. There were no significant differences between the two GSH-NS formulations in HCT116 and DU145 MCS, whereas there was a significant difference in the 2D HCT-116 culture. Indeed, IC<sup>50</sup> values were significantly lower in the 2D cultures than in the 3D cultures, especially after 24 h incubation, where similar values were reached only after 72 h of incubation. For example, IC<sup>50</sup> was twofold higher after 24 h in 3D cultures for HCT116 and three-fold higher in DU145 than in their respective monolayers. On the other hand, IC<sup>1</sup> concentrations were significantly lower in HCT116 MCS than in HCT116 cell monolayers, whereas IC<sup>1</sup> was quite similar both in DU145 spheroids and monolayers.

GSH-NS cytotoxicity might appear to be linked to disulfide-bond content in 2D cell monolayers as the formulation with the higher disulfide-bond content, GSH-NS D, had the lowest cytotoxic effect in all cell lines, except for the DU145 cell line. On the other hand, GSH-NS cytotoxicity was not influenced by the disulfide-bond content in MCS and the most pronounced cell growth decrease was observed in the colorectal cancer cell line, HCT116, after 24 h of exposure to GSH-NS. Tissue-like morphology and phenotypic change may be identified as the major factors in diminishing toxicity on MCS. This means that in vitro 3D cell culture models could act as an intermediate stage and bridge the gap between in vitro 2D and in vivo studies, which would extend current cellular level cytotoxicity to the tissue level and improve the predictive power of in vitro nanoparticle toxicology [60]. Finally, GSH-NSs showed a limited toxicity, leading to G<sup>1</sup> cell cycle arrest, without membrane damage or oxidative stress generation at significantly higher concentrations about fifty times those used for the delivery of anticancer drugs.
