Synthesis and optimization of gemcitabine-loaded nanoparticles by using Box–Behnken design for treating prostate cancer: In vitro characterization and in vivo pharmacokinetic study

2.1 Materials

GC hydrochloride was obtained from Eli Lilly Pakistan (Private) Limited as a donation. TPP and CS (50,000–90,000 Da molecular weight and 75–85% deacetylation degree) were bought from Sigma-Aldrich, USA. Other chemicals, including glacial acetic acid (GAA), sodium chloride, sodium hydroxide, phosphoric acid, potassium chloride, potassium dihydrogen phosphate, sodium dihydrogen phosphate, and disodium hydrogen phosphate, were supplied by Merck (Germany). Acetonitrile and methanol (MeOH) were provided by Fisher Scientific (UK). The reagents utilized in the study all were of analytic grade.

2.2 Preformulation studies

The preliminary studies were performed to determine the nanoparticle’s production zone [24]. For this purpose, various concentrations of CS (0.05–1% w/v) were prepared in 1% (v/v) aqueous solution of GAA, and the pH was adjusted to 5–6. Different concentrations of TPP (0.01–0.5% w/v) were prepared in double distilled water. Then, 2 ml of TPP solution was added dropwise to 5 ml of CS solution with magnetic stirring at room temperature. All the formulations were inspected visually and were categorized into three different zones: aggregates, opalescent suspension, and a clear solution. The opalescent system, which may be a suspension of small particles, was further studied, reaching a target concentration of CS in the range of 0.1–0.3% (w/v) and TPP 0.02–0.05% (w/v). These samples were microscopically observed and were categorized into aggregates and nanoparticles [25,26].

2.3 Design of experiments (DOE)

The traditional method for formulation development is time-consuming and costly, as during the process, we must change one factor at a time, and it does not provide information about the effects of interactions among the factors on the final formulation properties. The DOE is an efficient means for the design and optimization of experiments so that the data acquired can be evaluated to get valid and objective conclusions with a minimum number of trials. Therefore, by applying a statistical approach, a 3-factorial, 3-level BBD (Design-Expert 7 trial version software, Stat-Ease Inc., Minneapolis, USA) was used for the design of nanoparticle formulation and evaluation of change in concentration of the variables on responses [27,28]. BBD is suitable for investigating the quadratic response surfaces and constructing the second-order polynomial equations, thus permitting the optimization of the process with the least number of runs [29]. The independent variables selected were CS concentration (X ₁), TPP concentration (X ₂), and GC concentration (X ₃), while the responses included particle size (PS) (Y ₁), polydispersity index (PDI) (Y ₂), zeta potential (ZP) (Y ₃), and percent entrapment efficiency (% EE) (Y ₄) [30], as shown in Table 1.

2.4 Preparation of GC-loaded nanoparticles

The nanoparticles were formulated by a slight modification of the ionic gelation method, as reported by Calvo et al. [31]. TTP was used as a cross-linking agent. Briefly, three different concentrations of CS (0.14, 0.28, and 0.42% w/v) were prepared by dissolving the required quantity of CS in 1% v/v GAA solution under magnetic stirring at 60˚C for 2 h. The pH of each resulting solution was adjusted to 5 ± 0.03 with sodium hydroxide aqueous solution of 20% w/v and then filtered by using a syringe filter of 0.45 µm (Millipore, USA) [32]. Similarly, three TPP aqueous solutions (0.07, 0.14, and 0.178% w/v) were prepared, dissolving the desired quantity in double distilled water at room temperature and then filtered through a 0.22 µm syringe filter (Millipore, USA). Blank nanoparticles (PNP) were prepared by the dropwise addition of TPP solution at a rate of 0.3 ml/min (with the help of BD Luer-Lok 25G × 1 inch 3 ml Syringe) to CS solution under continuous magnetic stirring [33]. The prepared nanosuspension was kept at 800 rpm and at room temperature for 1 h. For the preparation of drug-loaded nanoparticles, the required quantity of GC was first dissolved in TPP solution, which was then added to the CS solution by following the same procedure and conditions as described for the preparation of PNP. The volume ratio of CS to TPP was 5:2 in all cases [34]. The nanoparticle suspension obtained was ultracentrifuged (Avanti 30 Centrifuge, Beckman Coulter Inc., Fullerton, CA, USA) at 10,000 rpm for 30 min [35]. The supernatant was carefully separated and collected for further analysis for % EE. The obtained nanoparticles were re-dispersed in double distilled water and freeze-dried. All the formulations mentioned in Table 2 were prepared using a similar method.

2.5 Entrapment efficiency and loading capacity

The % EE of GC in the GCNP was determined by separating the GC-containing supernatant by centrifugation (Avanti 30 Centrifuge, Beckman Coulter Inc., Fullerton, CA, USA) at 10,000 rpm for 30 min. The clear supernatant collected was then analyzed to determine the content of GC by using high-pressure liquid chromatography (HPLC) [32]. The instrument consisted of Alliance e2695 Waters equipped with 2998 PDA and Empower 3 software, an HPLC system, and a C8 column 4.6-mm × 25 cm; 5 µm packing was used. The mobile phase contained phosphate buffer (pH 2.4–2.6) and methanol in a ratio of 90:10 v/v. To prepare the buffer solution, 13.8 g of sodium dihydrogen phosphate and 2.5 ml of phosphoric acid were taken in 1 l of water. The mobile phase was filtered through Whatman filter paper and sonicated for 20 min by using Sonicator (Elmasonic-E30H-Hans, Germany). The flow rate was adjusted to 1 ml/min, 20 µl was the injection volume, and GC was detected at λ _max 275 nm. The sample solutions were filtered by using a syringe filter (0.22 µm) before injecting into the HPLC system. All the samples were determined in triplicate. The % EE and drug loading capacity (% LC) were calculated by using Eqs. 1 and 2, respectively [36]:

where Drug_t is the total quantity of GC used in the formulation of GCNP and Drug_f is the free GC in the supernatant.

2.6 Lyophilization

The nanoparticles obtained after ultracentrifugation were re-dispersed in double distilled water. Trehalose dihydrate 5% (w/v), as a cryoprotectant, was applied to each nanoformulation at a concentration of 10 % v/v. All the formulations with added cryoprotectant were freezed at −30°C for 12 h and then lyophilized at −50°C for 24 h by maintaining a vacuum of 0.133 mBar by using Labconco 7386020 4.5 Liter Benchtop Cascade Freeze Dryer [28,37]. All formulations mentioned in Table 2 and the OF were lyophilized in the same manner. Only lyophilized formulations were used for characterization and further studies.

2.7 PS, PDI, and ZP

The PS, PDI, and ZP of nanoparticles were established using DLS with the Zetasizer Nano ZS 90 (Malvern Instruments, Worcestershire, UK). DLC is a commonly used technique for determining PS in colloidal suspensions like CS nanoparticles, which swell in aqueous media, allowing DLS to measure the hydrodynamic diameter of the nanoparticles. As per the procedure, the lyophilized nanoparticles were redispersed in deionized water, vortexed for 1 min, and then analyzed for PS, PDI, and ZP. All measurements were taken in triplicate [30].

2.8 In vitro dissolution studies

The release study of the GC solution and the OF was performed using the membrane diffusion method [38]. Before use, the dialysis membrane (mol. cut-off of 12 kDa) was soaked in the dissolution medium, phosphate buffer saline (PBS; pH 7.4), overnight. In brief, 5 mg of GC and 10 ml of OF dispersion equivalent to 5 mg of GC in PBS were sealed in the dialysis membrane and kept in 50 ml of dissolution medium, previously maintained at 37 ± 0.5°C, and continuously stirred magnetically at 100 rpm. At specified intervals (0.5, 1, 2, 5, 8, 16, 24, 32, 48, 64, 80, and 96 h), 1 ml of the sample was taken from the medium. After each sampling, PBS was refilled in equal volumes to maintain the original testing conditions. The analysis for GC concentration was performed by using HPLC [27,36]. The sample solutions were filtered by using a syringe filter (0.22 µm) before injecting into the HPLC system. The release study for the GC solution was conducted in the same manner.

2.9 FTIR spectroscopy

FTIR spectroscopy was performed using a spectrophotometer (Bruker Alpha, Bruker Optics, Leipzig, Germany). This analysis aimed to detect potential interactions between the polymer and the drug. For this purpose, the IR spectra of CS, GC, PNP, and GCNP were recorded within the wavelength of 4,000 to 400 cm⁻¹.

2.10 DSC

GC, PNP, and GC-loaded nanoparticles were characterized using DSC (Perkin Almer, 60A, Rodgau, Germany DSC calorimeter) in the temperature range of 5°C to 600°C at a 10°C per minute scan rate under nitrogen flow.

2.11 XRD

The XRD spectra of CS, GC, and GCNP were obtained at room temperature in the diffraction angle range of 0–80° (θ). The analysis was performed using an X-ray diffractometer (D/max-2500pc, Rigaku Corporation, Tokyo, Japan).

2.12 SEM

Surface morphology and the sizes of PNP and GCNP were determined using SEM. This analysis provided detailed insights into the surface characteristics and dimensions of the nanoparticles, which are crucial for understanding their physical properties and potential behavior in biological systems. The lyophilized nanoparticles were redispersed in deionized water. A sample drop was placed on a metallic stub and vacuum-dried overnight in a desiccator. The samples were subjected to gold plating. A 15 nm thin film was applied for 45 s at 40 mA and 10⁻²mbar. Then, the sample was observed under an SEM microscope (Hitachi High Tech S4800 FE-SEM).

2.13 Stability studies

The stability study was conducted by dividing the OF into three parts and each was kept in a 15 ml tube. Two tubes, after covering with aluminum foils, were stored in a refrigerator at −4°C to −6°C, while the third tube was stored uncovered at room temperature. Then, the samples were analyzed for PS, surface charge, and GC release immediately after 1, 3, and 6 months [39,40].

2.14 Cell line studies

Cytotoxic assays were conducted by the MTT (3-[4,5-dimethylthiazole-2-yl]-2,5-diphenyl-tetrazolium bromide) colorimetric assay to compare the effects of GC, PNP, and GCNP at equivalent drug concentrations on human prostatic carcinoma cell line (PC-3, an androgen-independent cell line derived from bone metastasis). The untreated cells served as a control. Dulbecco’s modified Eagle’s medium containing 5% fetal bovine serum, 100 IU·ml⁻¹ penicillin, and streptomycin each was used as a growth medium, and cells were cultured in flasks maintained at 37°C in a 5% carbon dioxide incubator. After growth, PC-3 cells were diluted with the medium to produce cell cultures at a concentration of 5 × 10⁴ cells·ml⁻¹. This culture (100 μl/well) was seeded on a 96-well plate for 24 h. The solubility of GC in PBS was typically 10 mg·ml⁻¹. The stock solutions of GC and GCNP with a drug concentration of 5 mg·ml⁻¹ were prepared in fresh PBS. From these stock solutions, various concentrations (1.25, 2.5, 5, 10, and 20 µg·ml⁻¹) of GC and GCNP were prepared in PBS. After the incubation period, the added medium from each well was drained, and 200 μl from each concentration was added. PNP were also diluted in PBS and were added to the PC-3 cell culture. In contrast, PC-3 control was subjected to a drug-free medium. After 48 h, 200 μl of MTT with a concentration of 0.5 mg·ml⁻¹ was added to each well, followed by a further 4 h of incubation period. Then, to each well, 100 μl of dimethyl sulfoxide was added to solubilize the formazan crystals. The degree of intracellular reduction of MTT to formazan was determined by determining the optical density values at a wavelength of 570 nm using a Spectra Max Plus microplate reader (Molecular Devices, CA, USA) [41,42]. Cytotoxicity was represented as IC_50, which is defined as the concentration causing 50% growth inhibition in PC-3 cells [43]. % cell viability was expressed as % relative absorbance of the sample as compared to control.

2.15 PK analysis

In in vivo studies, rabbits were used because they are easier to handle and less expensive than small rodents such as rats and mice [44]. Their large size allows for more efficient administration of drugs via feeding tubes and easier blood sampling from the jugular vein. Twelve albino rabbits of both sexes, aged 5–6 months and weighing 2–3 kg, participated in the study. The rabbits were equally divided into reference and test groups randomly. The experimental procedures conformed to the International Council for Harmonization guidelines, which were approved by the Ethics Committee of Bahauddin Zakariya University, Multan. Rabbits were given unlimited water access but were deprived of food 24 h before dosing. A dose of 25 mg·kg⁻¹ body weight [45] of GC solution and suspension was administered orally via a nasogastric tube to the reference and the test groups. Subsequently, at time intervals of 0.5, 1, 3, 6,12,18, 24, 48, and 72 h, 0.5 ml of blood samples were drawn from the jugular vein and collected into the serum vials, and the standard method was used for protein precipitation. To 200 μl of rabbit serum, 100 μl of 6% perchloric acid (w/v) was added dropwise to each sample. The samples were vortexed for 1 min and incubated on an ice pack for 5 min. The samples were then ultracentrifuged (Avanti 30 Centrifuge, Beckman Coulter Inc., Fullerton, CA, USA) at 10,000 rpm for 15 min at 4°C. The supernatant from each sample was collected in an Eppendorf tube. The sample solutions were filtered by using a syringe filter (0.22 µm) before injection into the HPLC system [39,46].

2.16 Data analysis

For the design and optimization of the process, Design-Expert^®7 trial version software (Stat-Ease Inc., Minneapolis, USA) was used. Microsoft Excel for Microsoft 365 MS Office and Analyze Data add-in for Microsoft Excel were used for data analysis. The release data were analyzed by using DD Solver, an add-in program for Microsoft Excel. PK parameters were determined using PK Solver 2.0, an add-in program for Microsoft Excel.

1. Ethical approval: The research related to animals’ use has been complied with all the relevant national regulations and institutional policies for the care and use of animals. The animal study protocol was approved by the Institutional Review Board (or Ethics Committee) of Bahauddin Zakariya University Multan, Pakistan (protocol code 0006/UREC/2023) for studies involving animals.

3.1 Nanoparticle preparation and statistical analysis of experimental data by using design expert

For optimization, BBD with 3 levels, 3 factors, and 17 runs was used. All 17 formulations, as shown in Table 2, were prepared. In the experimental design, four responses were observed, as summarized in Table 3. The PS of the prepared formulations varied between 200.5 and 385.3 nm. Among these, the F13 formulation exhibited the smallest PS, whereas the F2 formulation was the largest. PDI values were scaled between 0.1 and 1, and a value below 0.5 is desirable [47]. The PDI results ranged from 0.251 to 0.752, reflecting the diversity in PSs within the formulations. Specifically, the F13 formulation demonstrated the lowest PDI, indicating a more uniform PS distribution, while F2 had the highest PDI. For higher physical stability, the ZP of the dispersion medium should be >+30 mV and <−30 mV [48]. The ZP values of the formulations were obtained from +24.3 to +34.2 mV. The F3 formulation had the lowest ZP at +24.3 mV, while F2 had the highest value of +34.2 mV, indicating greater stability. Higher positive ZP values correlate with improved stability of the formulations. EE varied between 60.3% and 78.3%. The F5 formulation exhibited the lowest, while the F13 formulation achieved the highest % EE for GC. The % LC of all the formulations varied between 6.56% and 11.13%. It was observed that % LC increases as the initial content of the drug increases in the formulation, and this corresponds to previous literature reports [32,49].

3.1.1 Analysis of PS

The fit summary suggested the quadratic model in PS analysis by the experimental design. The ANOVA results indicated the model F-value of 369.59 and a p-value <0.0001, which is well below 0.05. The lack of fit was 1.91, suggesting that it is not significant. The predicted R² and adjusted R² values were 0.9789 and 0.9952, respectively, and in agreement. All these values confirm that the selected model is fit. The regression equation for PS after elimination of non-significant terms is given as follows:

(3) PS = + 206.42 + 42.91 X 1 − 6.18 X 1 X 2 + 110.14 X 1 2 + 12.61 X 2 2 + 15.84 X 3 2

The above equation characterizes the second-order polynomial relation for PS, where +206.42 is the model intercept coefficient, +42.91 is the coefficient of linear term X _1, and −6.18 is the coefficient of the interaction term X ₁ X ₂. At the same time, +110.14, +12.61, and +15.84 represent the coefficient of quadratic terms X 1 2 , X 2 2 , and X 3 2 , respectively. Here, variables X ₁, X 1 2 , X 2 2 , and X 3 2 have positive effects while X ₁ X ₂ harms PS. Among the above variables, X ₁ and X 1 2 have a larger impact on the PS than on the interaction terms X 2 2 and X 3 2 . This means that the PS increases with increasing CS concentration (X ₁) [30], which may be associated with the availability of more cationic groups for binding with TPP, resulting in larger particles. It is also established that higher CS concentration causes an increase in the gelation medium viscosity and resistance against dispersion, which results in large-size nanoparticles (Figure 1) [50].

Figure 1

3D surface graphs representing effects of concentration of various independent variables on responses: (a) effect of CS and TPP concentrations on the PS, (b) effect of TPP and GC concentrations on the PS, (c) effect of CS and TPP concentrations on PDI, and (d) effect of TPP and GC concentrations on PDI.

3.1.3 Analysis of ZP

The quadratic model was suggested for the analysis of ZP by the experimental design. The ANOVA results indicated the model F-value of 45.64 and a p-value of <0.0001, which is significantly below the 0.05 threshold, indicating that the applied model is significant. Additionally, the lack of fit value was 3.38, and the predicted R² of 0.8003 was in reasonable agreement with the adjusted R² of 0.9617. These parameters collectively suggest that the applied model is significant. The regression equation for ZP after the elimination of non-significant terms is given as follows:

(5) ZP = + 30.14 + 3.64 X 1 − 0.7375 X 2 − 1.21 X 2 2 − 1.08 X 3 2

The above equation shows a second-order polynomial equation for ZP, where +30.14 is the intercept coefficient, +3.64 and −0.7375 are the coefficients of linear terms X ₁ and X _2, while −1.21 and −1.08 represent the coefficient of quadratic terms X 2 2 and X 3 2 , respectively. Here, variable X ₁ has a positive effect while X _2, X 2 2 , and X 3 2 have an adverse effect on ZP. Therefore, as CS concentration (X ₁) increases, ZP increases [51], while it decreases as the concentrations of TPP (X ₂) and GC (X ₃) increase (Figure 2).

Figure 2

3D surface graphs representing effects of concentration of various independent variables on responses: (a) effect of CS and TPP concentrations on ZP, (b) effect of TPP and GC concentrations on ZP, (c) effect of CS and TPP concentrations on % EE, and (d) effect of TPP and GC concentrations on % EE.

3.2 Optimization

After investigating the effects of selected factors on different responses, optimal levels for these factors were identified. Consequently, the variables were adjusted within a specified range, and optimization was performed. The OF is defined as one that achieves a lower PS, a lower PDI, a ZP within the desired range, and a high drug % EE (Table 4). The variable values of OF are provided in Table 5. The desirability for OF was found to be 0.970 (Figure 3).

Figure 3

Desirability plot of OF.

3.3 Predicted and measured responses for OF

The predicted and measured values for independent variables for OF are given in Table 6. The measured values of OF for PS, PDI, ZP, and % EE were 206.7 nm, 0.285, 27 mV, and 77.61%, respectively (Figure 4), showed favorable characteristics and are predicted to have high long-term stability. However, the % LC for OF determined was 11.90%.

Figure 4

PS and ZP analysis of OF: (a) particle and (b) ZP.

3.4 In vitro dissolution studies

Figure 5 describes the in vitro profile of GC from OF (GCNP) and GC solutions in PBS. It was observed that the release of GC from GC aqueous solution was fast and almost complete (p < 0.05) within 8 h, while OF at the initial level showed a burst release of 61 ± 1.1% up to 8 h, and then a slow release of 95.81 ± 1.1% was observed up to 96 h. The phenomenon of burst release was because of the molecular diffusion of a drug present on the nanoparticle’s surface [53]. Then, the slow release of GC can be due to the entrapment of the drug in the polymer matrix [54]. The GC release data were fitted into different kinetic models by using the software DD solver, an add-in program for Microsoft Excel. It was observed that the release of GC from the GC solution exhibited first-order kinetics, which was indicated by the highest R ² value (0.9941) [38]. Whereas OF obeyed the Korsmeyer–Peppas model with the highest R ² value (0.9507) and followed the non-Fickian drug release mechanism (n > 0.45), which means that the phenomena involved in drug release are drug diffusion processes from nanoparticles coupled with relaxation of polymer chains [55].

Figure 5

Cumulative % release of GC from GC solution and GCNP in PBS; pH 7.4 at 37°C (n = 3).

3.5 FTIR spectroscopy

FTIR spectroscopy was conducted to evaluate various interactions among GC and excipients. In an acidic aqueous medium, there is a positive charge on CS because of the protonation of the primary NH₂ group, which facilitates electrostatic interactions with TPP and nanoparticle formation. The FTIR spectra of GC, CS, PNP, and GCNP are shown in Figure 6. The CS spectrum shows the stretching bands at 1,025, 1,554, and 3,200–3,500 cm⁻¹ corresponding to C–O, C–H, and –OH groups, respectively, while the vibrational band at 1,650 cm⁻¹ represents the primary NH₂ group. Additionally, the peak at 2,865 cm⁻¹ indicates the methylene group stretching band in the CS [16]. For PNP, the amino band shifts to 1,536 cm⁻¹, suggestive of an ionic interaction between the NH₂ group of CS and TPP. The broad band at 3,500–3,200 cm⁻¹ is due to hydrogen bonding between TPP and CS. Because of these interactions, CS solubility reduces and thus facilitates the formation of nanoparticles. In the GC spectrum, a specific peak at 1,673 cm⁻¹ represents the ureido group, and a peak at 3,400 cm⁻¹ is attributed to the overlapping of NH₂ and –OH bands in this area [53]. While GC is incorporated into CS nanoparticles, a small new band appears at 1,734 cm⁻¹. Simultaneously, the free NH₂ band at 1,650 cm⁻¹ disappears, and the peak at 1,552 cm⁻¹ broadens. These changes suggest a high interaction between the NH₂ groups of CS and GC. The broad band in the range of 3,000–3,600 cm⁻¹ indicates the existence of strong hydrogen bonding between the nanoparticle matrix and the drug [16].

Figure 6

FTIR spectra: (A) GCNP, (B) PNP, (C) GC, and (D) CS.

3.6 DSC

To verify the physical status of GC in formulation, GC alone, PNP, and GCNP were investigated using DSC. As shown in Figure 7, the melting point of GC is detected as a sharp peak at a temperature of about 291.61°C [56]. The DSC thermogram of CS nanoparticles displayed a broad endothermic transition centered around 130°C, which is indicative of the CS’s glass transition temperature. This transition is associated with the thermal softening of the amorphous component of CS. The DSC curve for GCNP showed notable deviations from the individual components. Specifically, the sharp endothermic peak relating to the melting point of GC was completely absent. This indicates that GC has transformed from its crystalline form to a more amorphous state upon encapsulation within the CS matrix [16]. This alteration can influence the drug’s solubility, release profile, and stability within the nanoparticles.

Figure 7

DSC thermograms: (A) GC, (B) PNP, and (C) GCNP.

3.7 XRD

XRD provides insights into crystallinity and structural changes occurring upon drug encapsulation, which are critical for understanding the drug release and stability profiles of the nanoparticles. As displayed in Figure 8, the XRD of GC exhibited distinct sharp peaks at specific 2θ angles, which indicate its crystalline nature. The characteristic peaks observed for GC were at 2θ values of approximately 18.8°, 22.5°, and 27.2°, corresponding to its crystalline lattice structure [57]. The XRD pattern of CS, in contrast, showed two relatively broad peaks at 2θ of about 10° and 20° [58]. The XRD of GCNP displayed a significant alteration compared to that of the GC and CS. The intensity of GC’s characteristic peaks was markedly reduced, indicating a partial loss of its crystalline structure upon encapsulation. This suggests that GC is incorporated into the CS matrix in a more amorphous state. The diffraction pattern of the GCNP exhibited a combination of the broad peak from CS and a few residual peaks from GC, reflecting that while some degree of crystallinity of GC is preserved, it is largely modified by the encapsulation process. The XRD results confirm that GC is effectively incorporated into the nanoparticles with a significant reduction in its crystalline structure. This alteration in the crystalline nature of GC is expected to impact its solubility, stability, and release characteristics, which are crucial for its therapeutic efficacy in nanoparticle-based delivery systems.

Figure 8

XRD spectra: (A) CS, (B) GC, and (C) GCNP.

3.8 SEM

The size and morphology of GCNP and PNP were determined using SEM, as illustrated in Figure 9a and b, respectively. The images revealed that both types of nanoparticles were spherical with smooth surfaces. The average diameter of GCNP was less than 300 nm. However, it should be noted that the probable aggregation of the samples, resulting from the use of lyophilized forms, could alter the observed size of the nanoparticles. Additionally, the step-like shapes with isolated islands observed in the SEM images are attributed to the lyophilization process used in sample preparation. In Figure 9a, the whitish drug particles at the NP surface show the drug adsorption and entrapment in the NP. It can be estimated that GC adsorbed at the outer NP surface could be readily available in the body.

Figure 9

SEM images: (a) GCNP and (b) PNP.

3.9 Stability studies

For stability studies, PS, ZP, PDI, and release percentage were determined for OF [39], as detailed in Table 7. Over the 6-month storage period, a minor increase in PS was observed. The ZP values shifted from +27 to +30.3 mV. PDI increased from 0.285 to 0.617, and the percentage release decreased from 95.81 ± 1.1% to 86.53 ± 1.3%.

3.10 Cell line studies

The cell line study was conducted to compare the effects of GC, GCNP, and PNP in comparison to control at equivalent drug concentrations on the PC-3 cell line (Tables 8 and 9). As illustrated in Figure 10, PNP exhibited cell viability similar to that of the control after 48 h, suggesting the biocompatibility of CS. It was observed that PC-3 cells incubated with GCNP exhibited significantly higher toxicity (IC₅₀∼3.06 ± 0.32 μg·ml⁻¹) than GC (IC₅₀∼4.11 ± 0.81 μg·ml⁻¹) (Figure 11). The enhanced cytotoxicity of GCNP can be attributed to the superior muco-adhesion properties of CS, leading to higher local drug concentrations near the cell surface. This develops a concentration gradient that promotes drug influx into the cell. While GC relies on nucleoside transporters for cellular entry, GCNPs utilize endocytosis for drug internalization, thereby prolonging the cytotoxic effect [59,60] (Tables 8 and 9).

Figure 10

In vitro cell cytotoxicity assay of PC-3 cells incubated with control and PNP after 48 h (n = 3), ns: p > 0.05.

Figure 11

In vitro cell cytotoxicity assay of PC-3 cells incubated with GC and GCNP after 48 h (n = 3), ns: p > 0.05, * p < 0.05, and ** p < 0.01.

3.11 PK analysis

The PK parameters of GC after oral administration of GC solution and OF were determined by using PK solver 2.0, an add-in program for Microsoft Excel, and are illustrated in Table 10. The C _max for GC solution was 1,290 ± 78 ng·ml⁻¹, and this low value was due to the high first-pass effect and lower bioavailability of GC after oral administration [61,62]. While C _max for OF significantly increased to 3,070 ± 385 ng·ml⁻¹, it could be achieved because the drug in OF escaped the hepatic first-pass effect due to the size and surface morphology of nanocarrier systems [63]. t _max was 1 ± 0.01 h for GC solution and 6 ± 0.37 h for OF. The delayed t _max for OF can be elaborated by the fact that a significant amount of drug is released at the lymphatic site before reaching the blood. t _1/2 was 5.6 ± 0.08 h for GC solution, while it was 16.9 ± 0.31 h for OF (about 3 times higher), indicating a prolonged half-life for the OF. The AUC calculated for the GC solution was 9,512 ± 60 ng·ml⁻¹·min, and for OF, it was 52,832 ± 389 ng·ml⁻¹·min. The lower AUC of the GC solution was due to the higher metabolism of the drug to inactive metabolites. In contrast, nanoparticles achieved higher AUC due to slow and sustained release from nanoparticles, which help to protect the drug from metabolism. The mean residence times for GC and GCNP were 7.9 ± 1.27 and 21.9 ± 1.73 h, respectively. OF showed about five times higher bioavailability compared to the GC solution. All these parameters suggest that GCNP provides improved PK parameters after oral administration (Figure 12) [40].

Figure 12

Evolution of serum concentrations vs time oral administration in rabbits of GC and GCNP (n = 6). ns: p > 0.05, * p < 0.05, and ** p < 0.01 comparing GCNP to the GC for each time point.