The doxorubicin content in samples obtained from characterization studies was determined using UV-Vis Spectrophotometer (DU 700 series, Beckman Coulter Inc, Brea, CA) at λ = 480 nm (Prados et al., 2015). Serial dilutions of stock doxorubicin solution (1,000 μg/ml) were made to obtain concentrations ranging from 0.1 to 50 μg/ml. Linear regression analysis was performed between the absorbance and the concentration of doxorubicin to establish the calibration curve using SigmaPlot (Systat Software Inc, San Jose, CA).

The intracellular accumulation of doxorubicin was studied using fluorescence microscopy reported in the literature (Y.-K. Zhang et al., 2016). Images were collected using EVOS^® FL Auto Imaging System (Model AMFAD1000) (Thermo Fisher Scientific, Fair Lawn, NJ). In order to take the images, the cells were visualized at a total magnification of 1,200 × (40 × objective with an internal magnification of 30 × ) using two different modes (i.e., phase contrast and fluorescence). Phase contrast was used to locate a region of cells free from any cellular debris and/or any overlapping cells. Fluorescence was used to determine doxorubicin accumulation within the cells, which was achieved by selecting the RFP filter (built-in the instrument) at a preset wavelength (excitation: 552 nm; emission: 585 nm). For both phase contrast as well as fluorescence, images of cells were acquired using a monochrome camera also built-in the instrument.

All double-coated doxorubicin-loaded PBCA-NPs formulations (5 mg) were suspended in 1 ml DPBS by brief sonication. This homogenous suspension was then transferred to a folded capillary cell (DTS1070) (Malvern Panalytical Inc. Westborough, MA). After a brief equilibration period inside the sample chamber, the mean hydrodynamic particle size (nm), the polydispersity of size distribution (PdI) and zeta potential were measured. For particle size analysis, dynamic light scattering (DLS) along with Non-Invasive Back Scatter Technology built-in the Zetasizer Nano ZS (Malvern Panalytical Inc. Westborough, MA) was used. Zeta potential was measured using laser doppler micro-electrophoresis also built-in the Zetasizer Nano ZS. The fundamental size distribution generated by DLS is an intensity distribution, which can be converted to a volume distribution. This volume distribution can also be further converted to a number distribution. However, number distributions are of limited use as small errors in gathering data for the correlation function which may lead to huge errors in distribution by number. Therefore, the particle size graph representing an intensity distribution was chosen for data interpretation. In addition, it may be worthwhile to note that various parameters (e.g., viscosity, refractive index, dielectric constant of the tested samples) can impact both the particle size and zeta potential measurements of the fabricated nanoparticles. However, impact of these parameters is not performed for routine analysis. Most often a suitable buffer system representative of a suitable diluent is selected to understand the characteristics of these nanoparticles. Since DPBS solution in water does not absorb (i.e., absorption coefficient practically equal to 0) the usual wavelengths of the lasers used in DLS (500–780 nm). At a given wavelength, not concentrated solutions of Pi buffer have the same refractive index as water. Therefore, 0.8872 cP, 1.330 (dimensionless), 78.4 (dimensionless), respectively, for viscosity, refractive index and dielectric constant of water were used as default reference values of equipment during these measurements.

The amount of drug entrapped was determined by completely dissolving the lyophilized doxorubicin-loaded PBCA-NPs (5 mg) in 5 ml methanol:acetonitrile (50:50) solution. The resulting solution was centrifuged at 13,300 rpm for 10 min at 4 ± 0.5°C to pelletize any undissolved materials. The clear supernatant was analyzed for doxorubicin content by the UV-Vis spectroscopy method described previously. The entrapment efficiency of doxorubicin was then calculated as a ratio of the assayed doxorubicin in lyophilized doxorubicin-loaded PBCA-NPs to the total doxorubicin (i.e., 400 mg) used in the fabrication as reported in the literature (Papadimitriou and Bikiaris 2009).

SW620 and AD300 were seeded in their respective growth medium in 24-well plate and cells were allowed to grow overnight at 37°C, 5% CO₂. On the day of the experiment, cells were washed with DPBS (pH 7.4) three times. Subsequently, the cells were treated with 5 µM free doxorubicin in solution (medium: DPBS without FBS), all double-coated doxorubicin-loaded PBCA-NPs formulations (equivalent to 5 µM doxorubicin), and blank (growth mediums). The treated cells were then incubated for 2 h at 37°C, 5% CO₂. Since it is widely reported that FBS is likely to interfere with the formulations diluted to desired concentration with DPBS without FBS. Without FBS the cells have a very limited lifeline, which according to literatures is between 2 and 4 h. Therefore, to avoid any interference from such factors, 2 h time-point was chosen for intracellular accumulation studies. Following the incubation period, the treatment (free drug, double-coated doxorubicin-loaded PBCA-NPs formulations, or growth medium) was aspirated. The cells were gently washed with DPBS three times and then immediately examined the cells using the fluorescence microscopy method described previously. For this study, single-coated formulations, free doxorubicin, and growth medium served as controls.

Since it is difficult to estimate the amount of up taken doxorubicin from microscopic images, an empirical parameter [i.e., corrected total cell fluorescence (CTCF)], expressed as fluorescence intensity, was chosen for indirect quantification of doxorubicin within the cell (Jensen 2013). CTCF was obtained after subtracting the intensity of the blank cells (background) from cells exhibiting fluorescence. This was performed using ImageJ^® software (National Institutes of Health, Bethesda, MD). Higher fluorescence intensity was used as a surrogate for higher intracellular accumulation of doxorubicin.

The modified paraformaldehyde, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) colorimetric assay was used to detect the sensitivity of cells to doxorubicin as well as double-coated doxorubicin-loaded PBCA-NPs in vitro (Y.-K. Zhang et al., 2015). A general procedure followed for all cell lines used in this investigation has been described (Y.-K. Zhang et al., 2015). Briefly, cells were loaded in 180 µl of complete growth medium in 96-well plates in triplicate at cell density of 5,000–6,000 cells/well. After incubation at 37°C, 5% CO₂ for 24 h, cells were treated with different concentrations of doxorubicin solution or double-coated doxorubicin-loaded PBCA-NPs (20 µl/well). After 72 h incubation at 37°C, 5% CO₂, 20 µl of MTT solution (4 mg/ml) was added to each well. The plates were further incubated at 37°C, 5% CO₂ for 4 h, enabling viable cells to change the yellow-colored MTT into dark-blue formazan crystals. Subsequently, the MTT/medium was carefully aspirated from each well without disturbing the cell, and 100 µl of DMSO was added into each well. Plates were placed on shaking table to ensure thorough mixing of formazan into DMSO. Finally, the absorbance was determined at 570 nm using microplate reader (ThermoFisher Scientific, Waltham, MA) and data acquisition was performed by SkanIt™ software (ThermoFisher Scientific).

The particle size, polydispersity index (PdI), zeta potential and entrapment efficiency of various non-coated and coated doxorubicin-loaded PBCA-NPs formulations in comparison with blank PBCA-NPs nanoparticles (placebo) are summarized in Table 1. And, a typical particle size distribution profile of fabricated PBCA-NPs (i.e., formulation T0P0) is shown in Figure 2.

As expected, the mean particle size of doxorubicin-loaded PBCA-NPs (formulation T0P0) increased, due to doxorubicin loading, as compared to that of blank PBCA-NPs (154.5 versus 109.4 nm). This increase in size could be attributed to interference in the surface deposition of dextran 70 (surfactant) chains caused by the presence of doxorubicin during the polymerization step. As a result, their adsorption on the PBCA-NPs may have reduced, leading to an increase in the particle size as reported elsewhere (Vansnick et al., 1985). Further coating of doxorubicin-loaded PBCA-NPs with Tween 80 and PEG 20000 at 1%, respectively, the mean particle size of single-coated doxorubicin-loaded PBCA-NPs increased to 182.6 nm (formulation T0P1) or remained similar size at 154.4 nm (formulation T1P0). This could be attributed to the high MW of PEGs (MW > 5,000 Da) resulted in the formation of a layer-wise polymeric network on the surface of PBCA-NPs (Chognot et al., 2003). Furthermore, with double coating with Tween 80 at 1% and PEG 20000 at either 1% or 2%, mean particle size of the double-coated formulations increased from 154.4 nm (formulation T1P0) to 203.6 nm (formulation T1P1) and 256.9 nm (formulation T1P2). In addition, with double coating with Tween 80 at either 1% or 2% and PEG 20000 at 1%, mean particle size of the double-coated formulations increased from 182.6 nm (formulation T0P1) to 203.6 nm (formulation T1P1) and 221.0 nm (formulation T2P1). Finally, with double coating with Tween 80 and PEG 20000 up to 2%, mean particle size of the double-coated formulations increased from 154.5 nm (formulation T0P0) to 203.6 nm (formulation T1P1) and 276.2 nm (formulation T2P2). This could be attributed to the presence of Tween 80 facilitating the hydrogen bond formation with incoming PEG 20000 as a double coating, forming a layer-wise coherent coating over doxorubicin-loaded PBCA-NPs (Tejwani et al., 2000). On the other hand, the mean particle size slightly increased from 256.9 nm (formulation T1P2) to 276.2 nm (formulation T2P2). Therefore, it can be concluded that an overcoating with high molecular weight of PEG may impact the particle size of doxorubicin-loaded PBCA-NPs. PdI (a dimensionless quantity) is basically a representation of the particle size distribution of nanoparticles within a given sample population. All formulations (Table 1) exhibited low PdI values (< 0.2), this could be attributed to serial filtration step employed during the preparation and isolation of nanoparticles from reaction medium. A low PdI ( ≤ 0.2) is generally deemed acceptable particle size distribution of nanoparticles when selecting polymeric nanoparticles as a delivery system (Danaei et al., 2018).

The mean zeta potential of blank PBCA-NPs was observed to be -3.09 mV (Table 1). This inherent -ve zeta potential could be due to the resonance stabilized negative charge formed during the polymerization step (Duffy, Zetterlund, and Aldabbagh 2018). Interestingly, the zeta potential of uncoated doxorubicin loaded PBCA-NPs (formulation T0P0) showed a positive shift in zeta potential following the addition of doxorubicin as compared to the blank PBCA-NPs (3.13 mV versus -3.09 mV). This shift in the zeta potential of formulations could be attributed to the predominant positive charge of doxorubicin (pKa = 8.2) at a pH of 3.00 used during the polymerization reaction. For various doxorubicin-loaded PBCA-NPs formulations, zeta potential values varied from 2.72 mV (formulation T0P1) to 5.29 mV (formulation T1P1). No obvious trend in mean zeta potential values as a result of single- or double-coating was observed. These findings suggest that the coating did not impact the shear plane of the particle, which otherwise would have resulted in shifts in zeta potential values.

As displayed in Table 1, the entrapment efficiency of doxorubicin within the various doxorubicin-loaded PBCA-NPs formulations varied from 86.5% (formulation T2P1) to 89.9% (formulation T0P0). This could be attributed to adding doxorubicin during the polymerization step. Doxorubicin added during the polymerization step may act as a nucleophile and can form a part of the growing polymeric chain, yielding higher entrapment within the formed nanoparticles (Kreuter 1994). All formulations evaluated in this study have a polymeric matrix which was stored as freeze dried particles and hence are stable for long duration. Moreover, all formulations (except placebo) were slightly + ve (Table 1), when measured in DPBS (pH 7.4), since the cell culture medium also has a pH of 7.4, the zeta-potential is not likely to change. Therefore, nanoparticle-cell interaction is not likely to get impacted.

The cytotoxicity profiles of doxorubicin-loaded PBCA-NPs formulations from MTT assays performed on SW620 and AD300 cell lines are illustrated in Figure 7. For the comparison of the cytotoxicity profiles, IC₅₀ values (i.e., half-maximal inhibitory concentration) were calculated and displayed in Table 2. The IC₅₀ values of free doxorubicin in SW620 and AD300 cell lines were 0.094 and 4.796 µM, respectively. Interestingly, it was observed that double-coated doxorubicin-loaded PBCA-NPs (specifically formulation T2P2) showed significantly higher (p < 0.001) cytotoxicity than free doxorubicin, single-coated (formulation T0P1) (p < 0.001), and un-coated (formulation T0P0) (p = 0.004) in AD300 cell line. These results suggest that double-coated PBCA-NPs, specifically, formulation T2P2 have improved anticancer property in doxorubicin resistant AD300 cells. A trend of higher cytotoxicity with increased concentration of coating with Tween 80 was observed. This trend could be attributed to inhibition of overexpressed p-glycoprotein transmembrane receptors by Tween 80 as previously reported. On the other hand, synergistic effect on cytotoxicity in AD300 cell line was observed with double-coated doxorubicin-loaded PBCA-NPs with Tween 80 and PEG 20000. Trend analysis of the synergistic effect indicates that overcoating with PEG 20000 might aid in partial reversal of resistance, which could be attributed to folding of high molecular weight PEG (>5,000 Da) presenting a barrier comprising of conformationally random molecular chains to prevent drug efflux. This partial resistance reversal action of doxorubicin-loaded PBCA-NPs formulations was determined by calculating the resistance fold (i.e., ratio of IC₅₀ values obtained in AD300 cells to those obtained in SW620 cells). As shown in Table 2, a reduction in doxorubicin mediated resistance on AD300 cell line treated with formulation T2P2 as compared to doxorubicin solution was identified (23-fold versus 51-fold) indicating that double-coating with Tween 80 and PEG 20000 on PBCA-NPs may potentiate the sensitivity of the resistant cells towards doxorubicin. However, no significant change in IC₅₀ values was observed when SW620 cells were treated with free doxorubicin in solution or formulation T2P2. This is based on the assumption that, incorporating doxorubicin within the nanoparticle will not impact its efficacy. Since IC₅₀ is a drug related parameter, we anticipate it to be similar.

The cytotoxicity profiles of doxorubicin-loaded PBCA-NPs formulations from MTT assays performed on H460 and BCRP-overexpressed MX20 cells are shown in Figure 8. And, IC₅₀ values were calculated and compared in Table 3. The IC₅₀ values of free doxorubicin within H460 and MX20 cell lines were 0.058 and 0.813 µM, respectively. It was observed that double-coated doxorubicin-loaded PBCA-NPs (formulation T1P1) showed significantly higher (p < 0.001) cytotoxicity (about 1-fold) than free doxorubicin in comparison to single-coated PBCA-NPs (formulation T0P1) resulting in partial reversal by about 3.5-fold. These results suggest that these double-coated PBCA-NPs formulations might have improved anticancer property in mitoxantrone resistant MX20 cells. A similar trend (with respect to SW620 and AD300 cells) of higher cytotoxicity with increased concentration of overcoating with Tween 80 was observed. Consequently, formulation T2P2, exhibited significantly higher cytotoxicity in comparison to formulation T1P1. This finding suggests that an overcoat of Tween 80 at 2% might be required for enhanced cytotoxicity of formulations in resistant cells. Further, this trend could be attributed to inhibition of overexpressed BCRP transmembrane transports by Tween 80 as previously reported (Petkova et al., 2007). On the other hand, IC₅₀ value of formulation T2P2 was insignificantly higher than that of formulation T1P0 (i.e., 0.653 versus 0.626), indicating similar cytotoxicity potential of formulation T2P2 as compared to formulation T1P0. Although it may be argued, that PEG 20000 may not further enhance the cytotoxicity in mitoxantrone induced MDR cells. Overcoating with PEG 20000 may also aid in partial reversal of resistance, which could be attributed to folding of high molecular weight PEG (20,000 Da in this case) is presenting a barrier comprising of conformationally random molecular chains to prevent drug efflux (Gref et al., 2000). These results suggest that, cytotoxicity of these formulation is due to a combined effect of overcoating the nanoparticles with Tween 80 and PEG 20000. Furthermore, based on the calculated resistance fold value (Table 3), it can be observed that formulation T2P2 is also able to potentiate the sensitivity of doxorubicin in the mitoxantrone mediated resistant cell line (i.e., MX20), wherein the reduction of resistance from 13.9-fold (i.e., doxorubicin solution) to 4.9-fold was observed. However, contrary to their enhanced efficacy towards the resistant cell line, lower cytotoxicity (higher IC₅₀ values) as compared to doxorubicin solution for all doxorubicin-loaded PBCA-NPs formulations was observed in H460 cell line, indicating the specificity of these formulations towards the resistant MX20 cell line. Furthermore, this difference could be attributed to variability in the availability of doxorubicin at the cellular level. For example, free doxorubicin is instantly available within the cells, whereas, with nanoparticles time dependent drug release also needs to be taken into consideration.