Preclinical In Vitro Studies with 3D Spheroids to Evaluate Cu(DDC)2 Containing Liposomes for the Treatment of Neuroblastoma

Preclinical in vitro studies of drug candidates for anticancer therapy are generally conducted on well-established 2D cell models. Unfortunately, these models are unable to mimic the properties of in vivo tumors. However, in vitro 3D models (spheroids) have been proven to be superior in reflecting the tumor microenvironment. Diethyldithiocarbamate (DDC−) is the active metabolite of Disulfiram, an approved drug for alcoholism and repurposed for cancer treatment. DDC− binds copper in a molar ratio of 2:1 resulting in a water-insoluble Cu(DDC)2 complex exhibiting anticancer activities. Delivery of the Cu(DDC)2 complex using nanoparticulate carriers provides decisive advantages for a parental application. In this study, an injectable liposomal Cu(DDC)2 formulation was developed and the toxicity was compared with a 2D neuroblastoma and a 3D neuroblastoma cell model. Our results indicate that Cu(DDC)2 liposomes complied with the size requirements of nanoparticles for intravenous injection and demonstrated high drug to lipid ratios as well as colloidal stability upon storage. Furthermore, an efficient cytotoxic effect on neuroblastoma 2D cell cultures and a very promising and even more pronounced effect on 3D cell cultures in terms of neuroblastoma monoculture and neuroblastoma co-culture with primary cell lines was proven, highly encouraging the use of Cu(DDC)2 liposomes for anticancer therapy.


Introduction
The development of drug candidates for anticancer therapy needs proper preclinical in vitro and in vivo studies to prevent clinical trial failure. The assessment of drug potency and selectiveness is of particular importance. The drug candidate should be able to reach the target tissue and eliminate cancer cells without significantly affecting non-malignant cells [1]. Regarding in vitro assays, 2D cell monoculture models are most frequently used. These models are easily established, generate reproducible results, and are performable at low costs. They can be applied as a primary testing model using cancer cells, as well as nonmalignant cells to assess efficiency, toxicity, and selectiveness of anticancer drugs. However, 2D cell monoculture models are unable to reflect the properties of in vivo tumors and their resistance to drugs [2]. Performing in vitro tests with 2D models unfortunately often leads to false positive testing results and, when being further examined in preclinical steps, to ineffective treatment of tumor xenografts. As an alternative, in vitro multicellular 3D models more accurately represent the microenvironment of in vivo tumors, e.g., allowing cells to secrete extracellular matrix (ECM) components. Cells can interact with the ECM and among each other, affecting cell proliferation, differentiation, gene and protein expression, and resistance to anticancer drugs [3]. 3D cultures can be categorized as scaffold-free or scaffold-based culture systems, where scaffolds can be produced with natural or synthetic 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-distearoyl-sn-glycero-3phosphoethanolamine-N-[methoxy(polyethylene glycerol)-2000] (DSPE-mPEG 2000 ) were generously donated by Lipoid GmbH (Ludwigshafen, Germany). Cholesterol and 3,3dioctadecyloxacarbocyanin perchlorate (DiO, catalog number 42364) were purchased from Sigma Aldrich (Darmstadt, Germany). RPMI 1640 with (cat. no. P04-18500) and without (catalog number P04-16520) phenol red, DMEM (catalog number P04-04500), sodium pyruvate 10 nM, PBS w/o Ca 2+ and Mg 2+ , Trypsin/EDTA 0.05%/0.02% and 0.25%/0.02% were obtained from PAN Biotech (Aidenbach, Germany). Fetal calf serum (FCS) was purchased from Sigma Aldrich. Sephadex G-50 was obtained from GE Healthcare Life Science (Marlborough, MA, USA), CellTiter Glo ® 2D and 3D assays (catalog number G7572 and G9683) from Promega (Walldorf, Germany), calcein-AM (catalog number C3099) from Molecular Probes (Eugene, OR, USA), and propidium iodide (PI, catalog number CN74.1) from Carl Roth (Karlsruhe, Germany). Copper sulphate, HEPES, sodium diethyldithiocarbamate trihydrate, sucrose, and all other chemicals were purchased from Carl Roth (Germany). Vivaspin Turbo 4, filtration unit, 100,000 MWCO (catalog number VSO4T42) was purchased from Sartorius and sterile cellulose acetate filter, 0.2 µm and 0.45 µm from VWR (International, Radnor, PA, USA).

Liposomal Preparation
The preparation of Cu(DDC) 2 liposomes followed the idea of Wehbe et al. (2016) and was modified as indicated ( Figure 1) [8]. Liposomes composed of DSPC:Chol [55:45 molar ratio] and DSPC:Chol:DSPE-mPEG 2000 [50:45:5 molar ratio] were prepared by the thin film hydration method [22] with subsequent hand extrusion [23]. Briefly, organic stock solutions were transferred at the indicated ratios into a round bottom flask and the solvent was removed via rotary evaporation. The flask was placed under vacuum for at least 2 h to remove residual solvent. The lipid film was hydrated with 1 mL of an aqueous CuSO 4 solution (Cu 2+ , 150 mM in aqua destillata, pH 3.5) resulting in a lipid concentration of 40 mM. After 30 min of swelling and agitation, the formed liposomes were extruded for 41 passages through a polycarbonate membrane with a pore diameter of 80 nm (Nuclepore ® , GE Healthcare Life Science, Chicago, IL, USA). Solvent removal, film hydration, and extrusion were performed at 65 • C which is 10 • C above the phase transition temperature of DSPC. Prior to the addition of DDC − , Cu 2+ liposomes were separated from non-encapsulated Cu 2+ via size exclusion chromatography (SEC). The SEC column, prepared with Sephadex G-50, was equilibrated with an EDTA containing sucrose buffer (SHE, 300 mM sucrose, 20 mM HEPES, 30 mM EDTA, pH 7.4). Hereafter, buffer exchange with an EDTA-free sucrose buffer (SH, 300 mM sucrose, 20 mM HEPES, pH 7.4) was performed via three centrifugation steps at room temperature (RT) and 3000× g for 1 h using Vivaspin ® Turbo 4 filtration units (100 kDa MWCO, Sartorius, Germany). To allow the diffusion of DDC − into the liposomal interior, followed by complexation of DDC − with the liposomal encapsulated Cu 2+ , the preparation was left at 25 • C for 10 min at 300 rpm in a shaker. Free DDC − was removed via filtration units as described above with the final SH formulation buffer. The resulting Cu(DDC) 2 liposomes were prefiltered via a cellulose acetate (CA) 0.45 µm filter (VWR International, USA) to remove non-incorporated and precipitated Cu(DDC) 2 . For cell experiments, the liposomes were sterile filtered under aseptic conditions through a 0.2-µm CA filter. Fluorescent-labeled liposomes (with 0.5% DiO, related to the total lipid amount) were used to perform flow cytometry experiments. For the DiO liposome preparation, DiO was added from an organic stock solution to the residual liposomal components in a round bottom flask. The lipid film was treated as described above. In this case, the lipid film was hydrated with 1 mL HEPES buffered saline (HBS, 10 mM HEPES, 140 mM NaCl, pH 7.4), resulting in a final lipid concentration of 20 mM. The liposomal dispersion was extruded through an 80-nm polycarbonate membrane (Nuclepore ® , GE Healthcare Life Science, USA) after 30 min of swelling and agitation and stored at 4-6 • C protected from light until use. Healthcare Life Science, USA) after 30 min of swelling and agitation and stored at 4-6 °C protected from light until use.

Surface Area Filtration
The usage of surface area filtration had to be validated to remove non-incorporated and precipitated Cu(DDC)2 from liposomal Cu(DDC)2. For this purpose, freshly prepared Cu(DDC)2 liposomes were prefiltered through a 0.45-µm CA filter and sterile-filtered through a 0.2-µm CA filter. These filtration steps are part of the liposomal production process and are referred to as the 1st + 2nd filtrations. To validate the influence of filtration steps on liposomal stability, two additional filtration steps were performed referred to as the 3rd and 4th filtrations. Filtered samples were taken after each filtration step for analyzing lipid and Cu(DDC)2 content.

Surface Area Filtration
The usage of surface area filtration had to be validated to remove non-incorporated and precipitated Cu(DDC) 2 from liposomal Cu(DDC) 2 . For this purpose, freshly prepared Cu(DDC) 2 liposomes were prefiltered through a 0.45-µm CA filter and sterile-filtered through a 0.2-µm CA filter. These filtration steps are part of the liposomal production process and are referred to as the 1st + 2nd filtrations. To validate the influence of filtration steps on liposomal stability, two additional filtration steps were performed referred to as the 3rd and 4th filtrations. Filtered samples were taken after each filtration step for analyzing lipid and Cu(DDC) 2 content.

Quantification of Lipids and Cu(DDC) 2
The Cu(DDC) 2 liposomes were analyzed for total lipid and Cu(DDC) 2 content. Phospholipid content was determined via Bartlett assay [24] and Cu(DDC) 2 content via a spectrophotometric method. Quantification of DDC − was determined by measuring the absorbance of complexed Cu 2+ at a wavelength of λ max = 435 nm with a GENESYS 10S UV-Vis spectrophotometer (Thermo Scientific™, Waltham, MA, USA). Standards were prepared by mixing Cu 2+ solutions (0.015-0.1 µmol Cu 2+ in aqua dest., pH 3.5) with a 70-µmol methanolic DDC − stock solution resulting in the formation of solubilized Cu(DDC) 2 . The calibration curve was obtained by plotting the absorbance versus the amount of substance of Cu 2+ . DDC − chelates Cu 2+ in a 2:1 molar ratio and, therefore, the amount of complexed DDC − can be calculated. For the analysis of Cu(DDC) 2 liposomes, the use of methanol has the advantage of disrupting the liposomal membrane and solubilizing the liposomal trapped Cu(DDC) 2 . Subsequently, the absorbance of complex Cu 2+ and accordingly the amount of Cu(DDC) 2 can be determined via UV-Vis spectrometry. Linearity was verified by linear least square regression analysis, where y is the absorbance, x the amount of substance of sample, m the slope, and b the y intercept: In this study, the Cu(DDC) 2 concentrations of Cu(DDC) 2 liposomes refer to the amount of intraliposomal complexed DDC − , calculated as described above. Accordingly, the drug concentration should be indicated with DDC − /Cu 2+ , whereby the molar ratio of DDC − + 0.5 mol Cu 2+ should be noted. For simplicity, DDC − /Cu 2+ liposomes were entitled as Cu(DDC) 2 liposomes. This is justified by the fact, that Cu(DDC) 2 is formed by DDC − and Cu 2+ in the aqueous core of liposomes and that Cu(DDC) 2 is responsible for the anticancer ctivity. The size expressed as the hydrodynamic diameter (d h ) and the polydispersity index (PDI) were measured via dynamic light scattering (DLS, ZetaPals, Brookhaven Instruments Corporation, Holtsville, NY, USA). Liposomal samples were diluted to 1-5 mM (total lipid) in filtered SH buffer for size and polydispersity analyses.

Cu(DDC) 2 to Lipid Ratio
The Cu(DDC) 2 to lipid ratio (drug to lipid ratio, D/L) was calculated with Equation (2), where n entrapped drug is the amount of substance of complexed DDC − with Cu 2+ and n total lipid the amount of substance of lipid:

Liposomal Stability Studies
For storage stability studies the liposomes were kept at 4-6 • C for 6 months (183 days) and at RT (22 ± 2 • C) and −20 • C for 2 months (61 days). Therefore, aliquots of non-PEGylated and PEGylated Cu(DDC) 2 liposomes were prepared and characterized via DLS at the indicated time points. Stability studies included the determination of the release rate of Cu(DDC) 2 from liposomes after storage of samples at 4-6 • C for 6 months using surface area filtration (Section 2.3).
Next, 3D cell cultures were generated by the liquid overlay method [25]. Briefly, each well of a 96-well microtiter plate was coated under aseptic conditions with 50 µL of an autoclaved and therefore liquefied 1.5% (w/v) agarose-PBS solution. After cool down (to RT) and solidification of the agarose, 200 µL of a 20,000 cell/mL suspension was added per well. Following a centrifugation step at 600× g, for 5 min at 20 • C, the plate was incubated for 72 h in an incubator (37 • C and 5% CO 2 , humidified) until reaching a spheroid diameter of approximately 400 µm to start the liposomal treatment. Spheroid formation and growth were documented with an Axiovert 40 CFL microscope, an Axiocam 305 color camera, and a ZEN core v3.0 software via extended depth of focus (Carl Zeiss, Oberkochen, Germany). For co-culture spheroids, the cell number of a primary and a cancer cell dispersion was adjusted to a mixed cellular dispersion at a 1:1 ratio. LS-hDFa co-culture spheroids were cultivated in RPMI supplemented with 10% FCS and LS-hSkMC co-culture spheroids were cultivated in DMEM supplemented with 5% sodium pyruvate and 10% FCS.

Cell Viability Experiments
For 2D monolayer experiments neuroblastoma cells and primary cells were seeded in white opaque 96-well microtiter plates (Greiner Bio, Frickenhausen, Germany) in a density of 2000-7000 cells per well and cultured overnight. Cells were treated with Cu(DDC) 2 liposomes, free Cu(DDC) 2 and Cu 2+ . After 24, 48, or 72 h of treatment, CellTiter-Glo ® for 2D cell cultures was added according to the manufacturer's protocol. For 3D spheroid experiments, 3D cell cultures were generated as described in Section 2.6. Following 3 days of incubation in an incubator, spheroids were treated with Cu(DDC) 2 liposomes. After 72 h, the viability was determined with the CellTiter-Glo ® assay for 3D cell cultures. EC 50 (half maximal effective concentration) values were calculated by Excel solver and non-linear least squares fitting [26] with the following equation.
The cell viability (y) is calculated as concentration (x) in µM. The term y 100 represents the viability of control cells, y 0 the lowest viability, and p the hill coefficient [27]. Cytotoxicity analyses were evaluated with at least three different data sets. The EC 50 value is calculated as the mean of the generated EC 50 values.

Epifluorescent Staining of 3D Spheroids
LS spheroids were cultured in RPMI 1640 medium w/o phenol red and treated with non-PEGylated or PEGylated Cu(DDC) 2 liposomes. Spheroids were stained with 1 µM calcein-AM and 5 µM PI for the last 2 h of a 72 h incubation period of Cu(DDC) 2 liposomes. Living and dead cells were analyzed with an Axiocam 305 color camera, a ZEN core v3.0 software (Carl Zeiss, Oberkochen, Germany), a 10× objective and a corresponding filter set for calcein-AM and propidium iodide. a 488-nm blue laser for excitation and a 530/30-nm filter set. There were 10,000 events per sample recorded and cells were gated using forward and sideward scattering to determine the live cell population. Afterwards, the extracellular DiO fluorescence was quenched by the addition of 0.08% trypan blue [28] and the samples were repeatedly analyzed to distinguish cellular association and cellular uptake of DiO-labelled liposomes. The fluorescence intensity and the percentage of fluorescent cells (living gate) were evaluated.

Statistical Analysis
All data are expressed as mean ± SD. GraphPad Prism 8.01 software was used for statistical analysis. One-way ANOVA with Tukey's multiple comparison test was used to evaluate differences between unpaired data sets. One-way ANOVA with Dunnett's multiple comparison test was used to evaluate differences between the mean of each sample with the mean of the control (unpaired). Differences were considered significant when p < 0.05 (* < 0.05, ** < 0.01, *** < 0.001) in all analyses.

Validation of Non-Encapsulated Cu(DDC) 2 Precipitate Removal via Surface Area Filtration
The preparation process of Cu(DDC) 2 liposomes and the experimental set up for stability analyses was validated. Initially, liposomal agglomeration occurring after the preparation process of Cu(DDC) 2 liposomes was noticed via DLS and Cryo-TEM (data not shown). This was mainly due to the presence of extraliposomal (non-encapsulated) precipitated Cu(DDC) 2 which agglomerates with liposomes due to the lipophilicity of the Cu(DDC) 2 complex. Therefore, surface area filtration was introduced as the final step in the preparation process to retain extraliposomal Cu(DDC) 2 from Cu(DDC) 2 liposomes. This approach had to be validated to ensure extraliposomal Cu(DDC) 2 retention without lipid loss. Figure (Figure 2b). A reduction of the D/L ratio was noted before and after 1 st + 2 nd filtration for non-PEGylated and PEGylated Cu(DDC) 2 liposomes, whereas the lipid amount remained constant. No significant differences in D/L ratios and lipid amounts were detected for both formulations between the 1 st + 2 nd filtration and 3 rd filtration. Non-PEGylated liposomes revealed a slight significant difference among the D/L ratios (p = 0.0104) between the 1 st + 2 nd filtration and 4 th filtration. Overall, the lipid amount did not change for neither non-PEGylated nor PEGylated Cu(DDC) 2 liposomes during consecutive filtration through 0.2-µm CA filters and no significant change of the D/L ratio was observed after the 3 rd filtration step.

In Vitro Colloidal Stability and Drug Release Analysis of Cu(DDC) 2 Liposomes
As described in Section 2.2, non-PEGylated and PEGylated Cu(DDC) 2 liposomes were prepared via the thin film hydration method with subsequent extrusion. Non-PEGylated Cu(DDC) 2 liposomes showed a d h of 156 ± 7 nm, a PDI of 0.16 ± 0.02, and a D/L ratio of 0.15 ± 0.03 mol:mol. PEGylated Cu(DDC) 2 liposomes exhibited a d h of 161 ± 7 nm, a PDI of 0.14 ± 0.01, and a D/L ratio of 0.30 ± 0.04 mol:mol (Table 1).

In Vitro Colloidal Stability and Drug Release Analysis of Cu(DDC)2 Liposomes
As described in Section 2.2, non-PEGylated and PEGylated Cu(DDC)2 liposomes were prepared via the thin film hydration method with subsequent extrusion. Non-PEGylated Cu(DDC)2 liposomes showed a dh of 156 ± 7 nm, a PDI of 0.16 ± 0.02, and a D/L ratio of 0.15 ± 0.03 mol:mol. PEGylated Cu(DDC)2 liposomes exhibited a dh of 161 ± 7 nm, a PDI of 0.14 ± 0.01, and a D/L ratio of 0.30 ± 0.04 mol:mol (Table 1).  The D/L ratio and lipid amount of non-PEGylated and PEGylated liposomes were analyzed after consecutive filtration steps and evaluated statistically via one-way ANOVA and Tukey's multiple comparison test (* p = 0.0104). Data are expressed as the mean ± SD (n = 3-4). Particle size, size distribution, and drug retention are critical parameters during storage. Therefore, aliquots of Cu(DDC) 2 liposomes were stored (in SH buffer) for 6 months (183 days) at 4-6 • C ( Figure 3a) and for 2 months (61 days) at RT (22 ± 2 • C; Figure 3b) and −20 • C ( Figure 3c) and analyzed for size and size distribution. Data was evaluated statistically using one-way ANOVA and Dunnett's post-hoc test by comparing the mean Pharmaceutics 2021, 13, 894 9 of 21 of each d h -column with the mean of the control d h -column (t = 0 days). For Cu(DDC) 2 liposomes, stored at 4-6 • C, a significant difference in d h was observed after 3 months (91 days) for non-PEGylated (p < 0.001) and after 6 months for PEGylated Cu(DDC) 2 liposomes (p = 0.006) compared to freshly prepared preparations. Figure 3b shows initial instabilities after 14 days for non-PEGylated (p = 0.02) and after 28 days for PEGylated Cu(DDC) 2 liposomes (p = 0.01) stored at RT. After one day storage at −20 • C non-PEGylated (p < 0.001) and PEGylated (p < 0.001) Cu(DDC) 2 liposomes exhibited increasing instability. The storage temperature of 4-6 • C resulted in constant size and size distribution for non-PEGylated and PEGylated Cu(DDC) 2 liposomes.
Pharmaceutics 2021, 13, x 9 of 22 Particle size, size distribution, and drug retention are critical parameters during storage. Therefore, aliquots of Cu(DDC)2 liposomes were stored (in SH buffer) for 6 months (183 days) at 4-6 °C ( Figure 3a) and for 2 months (61 days) at RT (22 ± 2 °C; Figure 3b) and −20 °C (Figure 3c) and analyzed for size and size distribution. Data was evaluated statistically using one-way ANOVA and Dunnett's post-hoc test by comparing the mean of each dh-column with the mean of the control dh-column (t = 0 days). For Cu(DDC)2 liposomes, stored at 4-6 °C, a significant difference in dh was observed after 3 months (91 days) for non-PEGylated (p < 0.001) and after 6 months for PEGylated Cu(DDC)2 liposomes (p = 0.006) compared to freshly prepared preparations. Figure 3b shows initial instabilities after 14 days for non-PEGylated (p = 0.02) and after 28 days for PEGylated Cu(DDC)2 liposomes (p = 0.01) stored at RT. After one day storage at −20 °C non-PEGylated (p < 0.001) and PEGylated (p < 0.001) Cu(DDC)2 liposomes exhibited increasing instability. The storage temperature of 4-6 °C resulted in constant size and size distribution for non-PEGylated and PEGylated Cu(DDC)2 liposomes. After the determination of a suitable storage temperature, it had to be proven that Cu(DDC) 2 does not leak out of liposomes in the timeframe between preparation, storage, and actual administration. Therefore, aliquots of non-PEGylated and PEGylated Cu(DDC) 2 liposomes were adjusted to equal D/L ratios and stored for 6 months at 4-6 • C. The liposomal aliquots were refiltrated through a 0.2-µm CA filter after specific storage periods and the D/L ratio of the filtrates was determined. A decrease in D/L ratio indicated Cu(DDC) 2 release. The mean D/L ratio of t = 0 days was defined as control (100%), so the collected data is expressed as % of control and plotted against the time (t = 0-183 days; Figure 4). PEGylated Cu(DDC) 2 liposomes displayed 90-100% drug retention until 3 months of storage, which decreased to 74% after 6 months of storage. For non-PEGylated Cu(DDC) 2 liposomes the drug retention decreased to 54% after 7 days and furthermore to 15% after 6 months.
After the determination of a suitable storage temperature, it had to be proven that Cu(DDC)2 does not leak out of liposomes in the timeframe between preparation, storage, and actual administration. Therefore, aliquots of non-PEGylated and PEGylated Cu(DDC)2 liposomes were adjusted to equal D/L ratios and stored for 6 months at 4-6 °C. The liposomal aliquots were refiltrated through a 0.2-µm CA filter after specific storage periods and the D/L ratio of the filtrates was determined. A decrease in D/L ratio indicated Cu(DDC)2 release. The mean D/L ratio of t = 0 days was defined as control (100%), so the collected data is expressed as % of control and plotted against the time (t = 0-183 days; Figure 4). PEGylated Cu(DDC)2 liposomes displayed 90-100% drug retention until 3 months of storage, which decreased to 74% after 6 months of storage. For non-PEGylated Cu(DDC)2 liposomes the drug retention decreased to 54% after 7 days and furthermore to 15% after 6 months
Furthermore, in vitro cellular uptake experiments were conducted on LS monolayers with 0.75 mM non-PEGylated and PEGylated DiO liposomes and determined via flow cytometry ( Figure 6). Cells were treated for 15 min, 1, 2, 4, and 6 h with DiO liposomes. The cellular uptake was rather low in the first hours and reached 24.8 ± 5.2% for non-PEGylated DiO liposomes and 17.8 ± 3.0% for PEGylated DiO liposomes after 6 h of incubation. A similar cellular fluorescence intensity was measured for both formulations.    Furthermore, in vitro cellular uptake experiments were conducted on LS monolayers with 0.75 mM non-PEGylated and PEGylated DiO liposomes and determined via flow cytometry ( Figure 6). Cells were treated for 15 min, 1, 2, 4, and 6 h with DiO liposomes. The cellular uptake was rather low in the first hours and reached 24.8 ± 5.2% for non-PEGylated DiO liposomes and 17.8 ± 3.0% for PEGylated DiO liposomes after 6 h of incubation. A similar cellular fluorescence intensity was measured for both formulations.

Cytotoxicity and Selectivity of Cu(DDC) 2 Liposomes on 2D and 3D Cell Cultures
Drug potency and selectivity of Cu(DDC) 2 liposomes were evaluated with LS neuroblastoma and primary hSkMC and hDFa cells using 2D monolayers and 3D spheroids. Initially, the cytotoxic potency was analyzed in 2D LS monolayers after 24, 48, and 72 h of incubation with non-PEGylated and PEGylated Cu(DDC) 2 liposomes (Figure 7). The cell viability was quantified via the 2D CellTiter-Glo ® assay. Statistical analyzes were performed with one-way ANOVA and Tukey's multiple comparison test. The resulting EC 50 values for 24, 48, and 72 h were in a concentration range of 0.16 ± 0.01 µM, 0.06 ± 0.01 µM, and 0.05 ± 0.02 µM for non-PEGylated and 0.13 ± 0.01 µM, 0.04 µM, and 0.07 ± 0.02 µM for PEGylated Cu(DDC) 2 liposomes, respectively. Cu(DDC) 2 liposomes affected the cell viability in a concentration and time dependent manner (Figure 7a,b), whereby the calculated EC 50

Cytotoxicity and Selectivity of Cu(DDC)2 Liposomes on 2D and 3D Cell Cultures
Drug potency and selectivity of Cu(DDC)2 liposomes were evaluated with LS neuroblastoma and primary hSkMC and hDFa cells using 2D monolayers and 3D spheroids. Initially, the cytotoxic potency was analyzed in 2D LS monolayers after 24, 48, and 72 h of incubation with non-PEGylated and PEGylated Cu(DDC)2 liposomes (Figure 7). The cell viability was quantified via the 2D CellTiter-Glo ® assay. Statistical analyzes were performed with one-way ANOVA and Tukey's multiple comparison test. The resulting EC50 values for 24, 48, and 72 h were in a concentration range of 0.16 ± 0.01 µM, 0.06 ± 0.01 µM, and 0.05 ± 0.02 µM for non-PEGylated and 0.13 ± 0.01 µM, 0.04 µM, and 0.07 ± 0.02 µM for PEGylated Cu(DDC)2 liposomes, respectively. Cu(DDC)2 liposomes affected the cell viability in a concentration and time dependent manner (Figure 7a,  Furthermore, cell viability of 3D LS spheroids was analyzed 72 h after treatment with Cu(DDC)2 liposomes via the 3D CellTiter-Glo ® assay. 3D spheroids were generated as described in Section 2.6 via the liquid overlay technique resulting in dense spheroids with a highly reproducible diameter and an inter plate coefficient of variation less than 3% [27]. Cu(DDC)2 liposomes affected the cell viability of 3D LS spheroids dose-dependently (Figure 8a,b). Non-PEGylated Cu(DDC)2 liposomes revealed an EC50 value of 4.09 ± 0.01 µM and PEGylated Cu(DDC)2 liposomes of 3.95 ± 0.80 µM when used on 3D LS spheroids. Furthermore, cell viability of 3D LS spheroids was analyzed 72 h after treatment with Cu(DDC) 2 liposomes via the 3D CellTiter-Glo ® assay. 3D spheroids were generated as described in Section 2.6 via the liquid overlay technique resulting in dense spheroids with a highly reproducible diameter and an inter plate coefficient of variation less than 3% [27]. Cu(DDC) 2 liposomes affected the cell viability of 3D LS spheroids dosedependently (Figure 8a,b). Non-PEGylated Cu(DDC) 2 liposomes revealed an EC 50 value of 4.09 ± 0.01 µM and PEGylated Cu(DDC) 2 liposomes of 3.95 ± 0.80 µM when used on 3D LS spheroids. Hence, the resulting EC 50 values were far higher in comparison to those of 2D LS monolayers (Figure 8c,d). Microscopy images of 3D LS spheroids were taken over time during treatment with Cu(DDC) 2 liposomes (Figure 8e,f). Surprisingly, treatment with non-PEGylated liposomes led to complete disintegration of the spheroids, whereby treatment with PEGylated liposomes resulted in maintenance of the spherical shape.
Hence, the resulting EC50 values were far higher in comparison to those of 2D LS monolayers (Figure 8c,d). Microscopy images of 3D LS spheroids were taken over time during treatment with Cu(DDC)2 liposomes (Figure 8e,f). Surprisingly, treatment with non-PEGylated liposomes led to complete disintegration of the spheroids, whereby treatment with PEGylated liposomes resulted in maintenance of the spherical shape.  Live/dead staining of 3D LS spheroids allowed to distinguish between proliferating and dead cells to further evaluate the influence of Cu(DDC) 2 liposomes (Figure 9). Therefore, 3D LS spheroids were stained with calcein-AM and PI for the last 2 h of a 72 h incubation period with Cu(DDC) 2 liposomes and compared with untreated spheroids being incubated with cell medium for the same duration of time (72 h total). The untreated control revealed a typical architecture of a dense spheroid containing a necrotic core (red) surrounded by an outer rim consisting of proliferating cells (green; Figure 9). This cellular arrangement changed dose dependently after treatment with Cu(DDC) 2 liposomes, result-ing in an increasing portion of dead cells, according to the viability analyses ( Figure 8). Single detached dead cells were visible after administration of low liposomal Cu(DDC) 2 dosages. With increasing concentration of Cu(DDC) 2 liposomes the PI signal intensified in the outer proliferating cell layer until it reached the necrotic core (Figure 9). and dead cells to further evaluate the influence of Cu(DDC)2 liposomes (Figure 9). Therefore, 3D LS spheroids were stained with calcein-AM and PI for the last 2 h of a 72 h incubation period with Cu(DDC)2 liposomes and compared with untreated spheroids being incubated with cell medium for the same duration of time (72 h total). The untreated control revealed a typical architecture of a dense spheroid containing a necrotic core (red) surrounded by an outer rim consisting of proliferating cells (green; Figure 9). This cellular arrangement changed dose dependently after treatment with Cu(DDC)2 liposomes, resulting in an increasing portion of dead cells, according to the viability analyses ( Figure 8). Single detached dead cells were visible after administration of low liposomal Cu(DDC)2 dosages. With increasing concentration of Cu(DDC)2 liposomes the PI signal intensified in the outer proliferating cell layer until it reached the necrotic core ( Figure 9). In a next step, the cytotoxicities of non-PEGylated and PEGylated Cu(DDC)2 liposomes were examined by using non-malignant cells. 2D monolayers of hSkMC as well as hDFa were incubated for 24 h with Cu(DDC)2 liposomes. The cell viability was analyzed via the 2D CellTiter-Glo ® assay (Figure 10a  In a next step, the cytotoxicities of non-PEGylated and PEGylated Cu(DDC) 2 liposomes were examined by using non-malignant cells. 2D monolayers of hSkMC as well as hDFa were incubated for 24 h with Cu(DDC) 2 liposomes. The cell viability was analyzed via the 2D CellTiter-Glo ® assay (Figure 10a  3D LS spheroids were also generated as co-culture with hSkMC as well as hDFa to analyze the cytotoxic effectiveness of Cu(DDC)2 liposomes. Consequently, 3D co-culture spheroids were also treated for 72 h with non-PEGylated and PEGylated Cu(DDC)2 liposomes. The viability was determined via the 3D CellTiter Glo ® assay, resulting in similar cytotoxic responses as determined in the tested cell lines (Figure 11a  3D LS spheroids were also generated as co-culture with hSkMC as well as hDFa to analyze the cytotoxic effectiveness of Cu(DDC) 2 liposomes. Consequently, 3D coculture spheroids were also treated for 72 h with non-PEGylated and PEGylated Cu(DDC) 2 liposomes. The viability was determined via the 3D CellTiter Glo ® assay, resulting in similar cytotoxic responses as determined in the tested cell lines (Figure 11a

Discussion and Outlook
In this study, liposomes were used as a drug delivery system for the water-insoluble Cu(DDC) 2 complex which exhibits cytotoxic effects on cancer cells. A reproducible preparation process of Cu(DDC) 2 liposomes was developed on the basis of the publication of Wehbe et al. (2016) [8]. The first part of this paper describes the additional integration of surface area filtration into the liposomal preparation process as a suitable method to remove extraliposomal Cu(DDC) 2 precipitates from Cu(DDC) 2 liposomes without lipid loss. It was proven that further consecutive filtration steps had no impact on the D/L ratio of Cu(DDC) 2 liposomes. Therefore, surface area filtration might also be considered as an in-process control step to evaluate and retain eventually released Cu(DDC) 2 from Cu(DDC) 2 liposomes after stress exposure (e.g., heat or storage), without affecting the D/L ratio by itself. Furthermore, the colloidal stability and drug release upon storage of non-PEGylated (DSPC:Chol [55:45 molar ratio]) and PEGylated (DSPC:Chol:DSPE-mPEG 2000 [50:45:5 molar ratio]) Cu(DDC) 2 liposomes were compared. In general, the physical stability of a liposomal dispersion is affected by changes in liposomal size due to aggregation and fusion and the leakage of entrapped drug. These physical instabilities originate from oxidation and hydrolysis of the liposomal constituents. Hence, the determination of the liposomal size and size distribution is a sensitive indicator of the physical liposomal stability [29]. The colloidal stability of Cu(DDC) 2 liposomes were evaluated at different storage temperatures (4-6 • C, RT, and −20 • C) in the presence of an SH buffer. An ideal storage temperature of 4-6 • C for liposomes was confirmed in various studies [30,31]. Furthermore, in this study, it was shown that a constant colloidal stability of non-PEGylated and PEGylated Cu(DDC) 2 liposomes were achieved at a storage temperature of 4-6 • C. Increasing instabilities were observed while storing at RT and −20 • C, despite the presence of a cryoprotective disaccharide (sucrose) in the liposomal buffer. In particular, storage at −20 • C could lead to mechanical stresses due to the formation of ice crystals during freezing. Buffers with glycerol and carbohydrates (e.g., trehalose) are reported to be more suitable for freezing liposomal dispersions in a temperature range between −20 and −30 • C [32]. However, it is recommended to store Cu(DDC) 2 liposomes at 4-6 • C, since this condition already reduces temperature-dependent hydrolysis and further cooling down of the preparation did not improve stability [29]. Moreover, the integration of hydrophilic PEG polymers on the surface of Cu(DDC) 2 liposomes not only resulted in higher D/L ratios, it also improved the colloidal stability during storage at 4-6 • C, probable through steric [33] and electrostatic repulsion [34]. Also, PEGylated Cu(DDC) 2 liposomes displayed a higher drug retention upon storage at 4-6 • C in comparison to non-PEGylated Cu(DDC) 2 liposomes. Thus, the usage of PEGylated Cu(DDC) 2 liposomes is recommended due to higher membrane stability. Overall, the presence of DSPE-mPEG 2000 can significantly modify characteristics of a liposomal formulation (e.g., increased entrapment efficiencies and higher drug retentions) [35]. In future studies, further impacts (e.g., pH, oxygen, and light) have to be examined to fully analyze the stability and degradation profile of Cu(DDC) 2 liposomes during storage. In addition, it is recommended to support stability analysis (conducted via DLS) with Cryo-TEM images. For long term storage of Cu(DDC) 2 liposomes, lyophilization can be taken into consideration. Furthermore, serum stability should be tested to further evaluate Cu(DDC) 2 liposomes for future in vivo experiments.
In the second part of this study, the in vitro cytotoxicity of Cu(DDC) 2 liposomes on 2D and 3D neuroblastoma cell models was examined. Initially it was shown, that liposomal Cu(DDC) 2 was more cytotoxic for 2D neuroblastoma cell cultures than free Cu(DDC) 2 , which highlights the usage of a liposomal Cu(DDC) 2 formulation. Furthermore, the viability reduction of neuroblastoma cells could be attributed to the Cu(DDC) 2 complex, since the corresponding concentration of free Cu 2+ did not affect the cell viability by itself. Beyond that, non-PEGylated and PEGylated Cu(DDC) 2 liposomes efficiently affected the viability of 2D LS monolayers in a time-and concentration-dependent manner. However, the resulting EC 50 values achieved for both formulations after 48 and 72 h of liposomal incubation did not differ significantly. Tawari et al. (2015) claimed, that Cu(DDC) 2 induces the cell death in a direct phase and in a delayed phase [36]. In the direct phase, cell viability is reduced via the formation of reactive oxygen species (ROS) which damages cells [18,34]. In the delayed phase, Cu(DDC) 2 interferes with essential molecular signaling pathways inducing apoptosis [36]. Consequently, the cellular ATP level rises, since the induction of apoptosis requires energy [37]. Therefore, it is possible, that the cell viability (measured via an ATP-dependent assay) is not significantly reduced after 72 h of incubation compared to 48 h of incubation with Cu(DDC) 2 liposomes, due to the delayed phase of the cell death induction triggered by Cu(DDC) 2 .
The cytotoxic effectiveness of Cu(DDC) 2 liposomes were further determined in more complex 3D LS monoculture spheroids. Surprisingly, treatment with non-PEGylated and PEGylated Cu(DDC) 2 liposomes lead to a distinct change in spheroid morphology. During treatment with non-PEGylated Cu(DDC) 2 liposomes a disintegration of the spheroid structure was observed where dead cells detached from the spheroid. No disintegration of the spheroid morphology was observed during the treatment of PEGylated Cu(DDC) 2 liposomes. However, the treatment with non-PEGylated Cu(DDC) 2 liposomes leads to the same extent of a complete viability reduction throughout different cell layers of the spheroid as with PEGylated Cu(DDC) 2 liposomes. This outcome implies high drug effectivity of the Cu(DDC) 2 formulations. Anyhow, a question arises about how the treatment with both Cu(DDC) 2 formulations can result in different spheroid morphologies. On the one hand, it is reported, that PEGylated liposomes could not interact with the spheroid due to PEG chains hindering an intratumoral transport [38]. This supports the suggestion that the different interactions of non-PEGylated and PEGylated Cu(DDC) 2 liposomes with cells or the spheroid could lead to distinct spheroid morphologies after treatment. On the other hand, Niora et al. (2020) described that PEGylated liposomes penetrate better into deeper cell layers of a spheroid due to poor interactions with cells compared to non-PEGylated liposomes [39]. Therefore, it is not clear whether non-PEGylated and PEGylated Cu(DDC) 2 liposomes accumulate at the spheroid surface or whether Cu(DDC) 2 liposomes are able to diffuse into deeper cell layers. Situation-dependent distinct cell deaths might be induced which could influence cell-cell-contacts and therefore the spheroid morphology. Consequently, differences in the induction of apoptosis and necrosis might be analyzed after spheroid treatment with non-PEGylated and PEGylated Cu(DDC) 2 liposomes. Particularly apoptosis leads to a dissolution of cellular binding properties [40]. The usage of light sheet fluorescence microscopy could give further information on the penetration behavior of fluorescence-labeled Cu(DDC) 2 liposomes in 3D spheroids [6,7]. Nonetheless, far higher EC 50 values were obtained as expected after Cu(DDC) 2 liposome treatment of 3D LS monoculture spheroids compared to those of 2D LS monolayers. Reasons for the drug resistance of 3D LS spheroids are obviously attributed to specific cellular interactions of the spheroid, to the three-dimensional arrangement of heterogeneous cell populations and the limited diffusion of mass transport which hinders the penetration of drugs and nanoparticles [6,7]. Moreover, cells of a spheroid are able to secrete components of an extracellular matrix (ECM) [3]. Desoize et al. (1998) described how cellular interactions with the microenvironment (cell-cell-contact and cell-ECM-contact) play a crucial role in resistance development against anticancer agents [41]. Presumably, a combination of the before mentioned factors leads to the resistance of 3D LS spheroids against Cu(DDC) 2 liposomes. Still, Cu(DDC) 2 liposomes effectively reduced the cell viability throughout different cell layers of the spheroid.
An effective drug candidate for cancer therapy should be able to reach the target tissue and to eliminate cancer cells without affecting non-malignant cells. Therefore, in the third part of this study the in vitro cytotoxicity of Cu(DDC) 2 liposomes on 2D primary cell models (hSkMC and hDFa) and 3D co-culture cell models (composed of neuroblastoma and primary cells) was analyzed. Here it was shown, that 2D hSkMC and 2D hDFa monolayers were less sensitive towards Cu(DDC) 2 liposomes compared to 2D LS monolayers. Likewise, Wehbe et al. (2016) observed in vitro a less cytotoxic effect of Cu(DDC) 2 liposomes on a primary cell line compared to cancer cells [8]. A possible explanation for the outcome could be a cell line-specific internalization rate for liposomes, since the uptake rate of nanoparticles was reported to be lower for non-malignant cells than for cancer cells [42,43]. Thus, the internalization rates of hSkMC and hDFa should be analyzed via uptake experiments similar to Section 3.2. Nevertheless, Cu(DDC) 2 liposomes turned out to reduce the cell viability of primary cell lines, which means that healthy tissue could also be affected during efficient treatment of an in vivo tumor. This problem might be solved via surface modifications of liposomes with antibodies or antibody fragments to specifically target cancer cells [44]. For a more reliable evaluation of the cytotoxic efficiency of Cu(DDC) 2 liposomes and heterotypic 3D-LS-hSkMC and 3D-LS-hDFa co-culture spheroids were used as cell models. The achieved EC 50 values of LS co-culture spheroids did not differ from those of LS monoculture spheroids. Furthermore, the same morphological changes of LS monoculture spheroids after treatment with Cu(DDC) 2 liposomes were observed for LS co-culture spheroids. It was expected, that hSkMC and/or hDFa eventually influence the spheroid resistance towards Cu(DDC) 2 liposomes, which was not observed in this study. In particular fibroblasts, as cancer associated fibroblasts (CAFs), have tumor-promoting effects, supporting the proliferation and metastasis of cancer cells and enhancing resistances against anticancer drugs [45]. Attention should be paid to the fact, that fibroblasts are a highly heterogeneous cell population with phenotypic characteristics, which strongly differ from where they originate [46]. It could be possible, that the hDFa used in this study were not suitable to mimic CAFs in 3D spheroids. Therefore, a combination of fibroblasts and myoblasts is suggested to be ideal for the generation of CAFs [47].
Further work is considered to be of interest for the future. This includes the transfer of the liposome preparation process from lab scale to large scale. Different preparation techniques, such as high pressure homogenization [48], microfluidization [49], cross flow filtration [50], or dual asymmetric centrifugation [51,52] can be implemented for the production of large volumes of Cu 2+ liposomes. Formation of Cu(DDC) 2 complexes inside the liposomes and separation of non-encapsulated material may be achieved by using industrial high throughput columns and high pressure filtration under aseptic conditions. Additional neuroblastoma cell lines might be of interest for further development and formation of 3D spheroids. Labeling of liposomal components and fluorescence activated cell sorting after incubation and later destruction of the spheroid might be helpful to gain more knowledge about the distribution and intraspheroidal processing of the liposomes and its contents in the different regions of the spheroid.

Conclusions
This study shows that the treatment of 3D spheroids, compared to 2D monolayers, provides more information concerning the cytotoxic effectiveness of Cu(DDC) 2 liposomes. Accordingly, it was proven that Cu(DDC) 2 liposomes effectively reduced the viability of in vitro neuroblastoma 3D spheroids. As it turned out that primary surrounding cells will also be affected, the implementation of a targeting structure on the nanocarriers seems to be of importance for reducing unwanted side effects and will, therefore, be in the focus of further studies.