Colloidal properties of ATV/PG-Lip

For optimization of ATV/PG-Lip colloidal properties, the effect of Lipoid S100, cholesterol and ATV content was studied (Table 1). A significant (p ≤ 0.05) decrease in liposomal vesicle size was observed on reducing Lipoid S100 concentration from 4% w/v (F1; 3922 ± 176.2 nm) to 2% w/v (F2; 760.3 ± 18.4 nm) with no significant effect (p > 0.05) on PDI. Similarly, a decrease in average vesicle size from 913 ± 35 nm to 268 ± 15.2 nm was previously reported with lowering phospholipid content from 900 to 300 mg [36]. Addition of cholesterol to the lipid mixture resulted in further significant (p ≤ 0.05) decrease in both vesicle size and PDI, as evident from comparing 0% w/v cholesterol (F2; 760.3 ± 18.4 nm and PDI 0.46 ± 0.01) to 0.5% w/v cholesterol (F3; 430.7 ± 13.8 nm and PDI 0.24 ± 0.001), reflecting improvement in the homogeneity of liposomal populations. These results are in agreement with previous findings showing reduction in particle size and PDI upon inclusion of cholesterol [37]. Based on its favorable colloidal properties, F3 (2% w/v Lipoid S100 and 0.5% w/v cholesterol) was selected for drug loading and further characterization.

Table 1 Formulation and optimization parameters of ATV/PG-Lip

Compared to F3 (430.7 ± 13.8 nm), ATV loading (F4–F7) resulted in a significant (p ≤ 0.05) reduction in vesicle size irrespective of the percentage of ATV loading. This could be attributed to ATV entrapment mainly in the liposomal lipid bilayer. This is due to ATV lipophilic nature [38] and slight solubility in phosphate buffer (pH 7.4), which consequently reduces repulsive forces between phosphate groups of lipid molecules leading to closer packing [20]. Similar size reduction has been previously reported upon lipophilic drug loading into liposomes for miconazole [20] and diclofenac [39]. Increasing ATV concentration from 0.2 to 0.6%w/v did not significantly (p > 0.05) affect vesicle size. However, a slight but statistically significant (p ≤ 0.05) increase in size was observed upon loading of 0.8%w/v ATV (F7; 307.9 ± 2.5 nm).

The average zeta potential for ATV-loaded formulations (F4–F7) was -18 ± 0.2 mV, similar to previously developed proposomes [13]. The negative charge observed is likely due to phosphate group ionization [40].

ATV entrapment efficiency

As can be seen in Table 1, ATV initial employed concentrations of 0.2–0.6% w/v resulted in percentage EE exceeding 80%. This was accompanied with a significant (p ≤ 0.05) increase in percentage DL on increasing ATV from 0.2% w/v (F4; 7.06 ± 0.25% w/w) to 0.6% w/v (F6; 18.33 ± 2.16% w/w). It is worth noting that percentage DL of ATV-loaded liposomes prepared by thin film hydration sonication method was 4.01 ± 0.05% w/w, as previously reported [38]. The comparatively enhanced ATV loading capacity observed in the current study could be attributed to ATV high solubility in PG (175.99 ± 2.08 mg/mL [41]) and further confirms PG-Lip as nano vector of high loading potential for ATV delivery.

However, by further increasing ATV initial concentration above 0.6% w/v, there was a significant (p ≤ 0.05) decrease in percentage EE (F7; 54.21 ± 4.9% w/v) with an insignificant (p > 0.05) change in percentage DL (F7; 16.48 ± 1.53% w/w; Fig. 1S), reflecting a maximum average percentage DL of ATV in PG-Lip of 17.41 ± 0.93% w/w. It should be stated that upon further increase in ATV concentration to 1% w/v, drug precipitation was evident, and it was hence excluded from the study.

Fig. 1

Characterization of the developed ATV/PG-Lip (a–c). Transmission electron microscopy images showing the morphology of blank PG-Lip and ATV/PG-Lip (a), scale bar = 200 nm. Physical stability data of ATV/PG-Lip based on vesicle size and polydispersity index when stored at 4 °C for 12 weeks (b), n = 3 at p ≤ 0.05. Cell cytocompatibility study for different formulations on human gingival fibroblasts (c), n = 7. Data represents mean ± SD. ns: statistically nonsignificant difference at p ≤ 0.05

Accordingly, F6 was selected as the optimized ATV-loaded formulation with optimum liposomal vesicle size (223.3 ± 2.1 nm), PDI (0.12 ± 0.001), and maximum percentage DL (18.33 ± 2.16% w/w) and was henceforth referred to as ATV/PG-Lip.

Transmission electron microscopy (TEM)

The morphology for both PG-Lip and ATV/PG-Lip was microscopically investigated. Both liposomal formulations showed spherical structure, with uniform vesicle size (Fig. 1a).

It is worth noting that microscopically determined vesicle size for PG-Lip and ATV/PG-Lip were lower than values determined using DLS. This difference in vesicle size was previously reported [42] and was attributed to the variation in sample preparation between TEM and DLS techniques. Since sample processing for TEM involves drying of the deposited dispersion on the grid prior to examination, subsequent water evaporation and possible shrinkage of the vesicles could occur. Also, TEM facilitates the examination of single liposomal vesicle, avoiding possible agglomeration.

Stability testing

The effect of storage on colloidal properties and ATV percentage EE of ATV/PG-Lip was monitored over 3 months at 4 °C (Fig. 1b). At the two-week interval, no change in either particle size or PDI was observed. A gradual increase in both parameters was noted thereafter, reaching a vesicle size of 436.5 ± 4.66 nm and PDI of 0.405 ± 0.018 by the end of the storage period (p ≤ 0.05), as previously reported [43], and could be attributed to vesicular aggregation. Regarding zeta potential, no significant (p > 0.05) change was observed after 3 months (-18.2 ± 0.3 mV).

Furthermore, successful ATV entrapment was maintained over 3 months, where the percentage EE remained above 80%. The potential of ATV/PG-Lip to efficiently retain ATV over time presents PG-Lip as a promising delivery system of reasonable stability overcoming a major drawback of liposomal systems which is encapsulated drug leakage [44].

In vitro cell cytocompatibility

For the investigation of cytocompatibility, the effect of PG-Lip and ATV/PG-Lip on the cell viability of human gingival fibroblasts was evaluated using MTT assay. Formulations were tested and compared to ATV solution at concentrations (1–30 µg/mL).

As seen in Fig. 1c, the results revealed that there was no significant (p > 0.05) reduction in cell viability for all tested formulations across different concentrations, verifying the cytocompatibility of the developed formulations.

3D printingOptimization of ink printability Morphological assessment

Different concentrations of polymer blend (PVA, HPMC, and chitosan; Table 2) were employed to optimize plain ink composition. PG (20% w/v final concentration) was used for plain ink preparation, as similarly included within liposomal formulation. By visually examining filaments continuity during extrusion, the printability of various inks was qualitatively assessed.

Table 2 Formulation and morphological assessment of plain inks for 3D printing optimization

The ink extrudability was improved by decreasing HPMC concentration from 4.5% w/v (ink A) to 3% w/v (ink B), where ink A showed extrusion failure resulting in complete nozzle clogging.

By increasing PVA concentration from 2% w/v (ink B) to 3% w/v (ink C), the ink showed less frequent nozzle clogging and less resistance to extrusion. This is in line with a previous study [19], where incorporating PVA and hyaluronic improved the extrudability of plain gelatin ink.

Additionally, the employed chitosan concentration greatly affected ink printability. By reducing chitosan concentration from 10% w/v (inks A–C) to 7% w/v (ink D), extrusion was relatively improved, however the ink formed a gritty non-continuous filament. Whereas decreasing chitosan concentration to 5 and 4% w/v (inks E and F, respectively) demonstrated more efficient extrusion, resulting in smooth continuous filaments. On the other hand, 3% w/v chitosan (ink G) formed a looser filament.

Accordingly, inks containing 10% w/v chitosan (inks A–C) were exempted from further trials and more experimental characterization was carried out for plain ink optimization using plain inks D–G (3% w/v PVA, 3% w/v HPMC and 7–3% w/v chitosan).

Viscosity measurements

In extrusion-based 3D printing, adjusting ink viscosity is critical for optimum extrusion without nozzle clogging and maintaining shape fidelity after printing [19]. Also, shear thinning property is a requirement for a continuous flow during 3D printing procedure [19]. In our study, characterization of the prepared inks was conducted via viscosity measurements. Initial viscosity values (at shear rate 1.2 s−1) were measured for plain inks containing decreasing chitosan concentrations (7–3% w/v; inks D–G) as demonstrated in Fig. 2a.

Fig. 2

Viscosity measurement of the developed inks for 3D printing (a, b). Initial viscosity values (at 1.2 s−1 shear rate) of tested plain polymer inks containing 3% w/v PVA and 3% w/v HPMC using different chitosan concentrations (3–7% w/v) (a). Data indicates mean ± SD, n = 3. Bars bearing different letters indicate statistically significant difference: a > b, at p ≤ 0.05. Viscosity values (at different shear rates) of tested plain inks containing 3% w/v PVA and 3% w/v HPMC using different chitosan concentrations and corresponding hysteresis loops of plain polymer inks; ink-D (7%), ink-E (5%), ink-F (4%) and ink-G (3%) (b)

Results showed that 7% w/v chitosan (ink D) demonstrated statistically (p ≤ 0.05) highest initial viscosity (32.63 ± 1.56 Pa.s.). This finding was reflected in the poor printability of ink D, forming non-continuous filament upon extrusion. However, there was no significant (p > 0.05) difference between initial viscosity values for 5–3% w/v chitosan (inks E–G), with an average value of (11.55 ± 1.4 Pa.s.).

All the tested plain inks (D–G) demonstrated a noticeable shear thinning behavior, as evident from viscosity vs stress curves (Fig. 2b), where the viscosity decreased by increasing the shear rate from (1.2 to 3.6 s−1; down-curve). Also, on decreasing shear rate (3.6 to 1.2 s−1; up-curve), viscosity values recovered gradually, generating a hysteresis loop and reflecting a thixotropic behavior [45]. Similarly, it was previously reported that chitosan gels possessed a shear thinning property [46]. By comparing hysteresis loops of inks E and F, it can be seen that both down- and up-curves were closer for ink E, implying more efficient and rapid viscosity recovery on release of shear [45]. Rapid recovery is desirable for achievement of high precision and shape fidelity of the printed structure.

Accordingly, ink E (3% w/v PVA, 3% w/v HPMC and 5% w/v chitosan) was selected as optimum plain ink and was henceforth referred to as plain polymer ink. Drug-loaded inks were prepared by mixing ATV in PG or ATV/PG-Lip with PVA (3% w/v), HPMC (3% w/v) and chitosan (5% w/v) and were referred to as ATV@ink and ATV/PG-Lip@ink, respectively. As control, PG-Lip@ink was similarly prepared using PG-Lip. Shear-thinning viscosity and elastic behavior were also verified for all the optimized inks (Fig. 2S).

Spreading ratio

Spreading ratio is considered as an important parameter to optimize ink printability. For production of highly precise hydrogel structures, lower spreading ratios are preferred [23]. In this study, we recorded percentage spreading ratio for the optimized inks (5% w/v chitosan); plain polymer ink, ATV@ink, PG-Lip@ink, ATV/PG-Lip@ink. For comparative analysis, spreading ratio was investigated for inks prepared using lower chitosan concentration (4% w/v chitosan); plain polymer ink F, ATV-, PG-Lip- and ATV/PG-Lip-loaded inks.

Interestingly, inks containing 4% w/v chitosan presented significantly (p ≤ 0.05) higher percentage spreading ratio than their respective counterparts as shown in Fig. 3a. This finding is due to the rapid viscosity recovery demonstrated by ink E (5% w/v chitosan) compared to ink F (4% w/v chitosan) and further justifies the opportune selection of Ink E. Moreover, results showed no significant (p > 0.05) difference in percentage spreading ratios for different optimized inks with 5% w/v chitosan (Fig. 3a).

Fig. 3

Optimization of 3D printing parameters (a–c). Optimization of ink spreading ratio for different tested inks containing 3% w/v PVA and 3% w/v HPMC, using 4% and 5% w/v chitosan (a). Data indicates mean ± SD, n = 3, different letters indicate statistically significant difference: a > b > c, at p ≤ 0.05. Effect of glutaraldehyde concentration on percentage swelling of ATV@3DP-film (b). FTIR spectra of chitosan powder and plain 3DP-film (c), verifying efficient chitosan crosslinking by glutaraldehyde

Collectively, plain polymer ink, ATV@ink, PG-Lip@ink, ATV/PG-Lip@ink were optimally prepared using 3% w/v PVA, 3% w/v HPMC and 5% w/v chitosan.

Crosslinking Optimization of crosslinker concentration

Structural strength is important to develop constructs with sufficient durability, which depends mainly on crosslinking [19]. In our work, we used Glut as an effective chitosan crosslinker [24]. For optimization of Glut concentration, monophasic square-shaped toroid constructs were 3D printed applying plain polymer ink and were crosslinked using 0.25, 0.5, 1 or 2% v/v Glut solutions. 3D-printed constructs were then dried before evaluation of film swelling over 6 h.

For 0.25% v/v Glut, the tested films were completely eroded/disintegrated after 0.5 h (not shown in Fig. 3b), while the film prepared using 0.5% v/v Glut lost integrity after 1 h (Fig. 3b). However, for both 1% v/v and 2% v/v Glut, structural integrity was maintained, and films swelling was noted over 6 h. Therefore, for higher safety and biocompatibility, the lower Glut concentration (1% v/v) was selected as optimum and was applied in this study.

Fourier transform infrared spectroscopy (FTIR)

For verification of efficient crosslinking of chitosan functional groups, FITR was conducted for uncrosslinked chitosan powder in comparison to crosslinked chitosan (in a monophasic construct printed using plain polymer ink).

As shown in Fig. 3c, uncrosslinked chitosan exhibited a characteristic broad band at 3324 cm−1, collectively due to the stretching vibrations of the OH and the functional NH2 groups in chitosan. The polysaccharide structure of chitosan was indicated by bands at 1373, 1021, and 2876 cm−1, related to stretching vibrations of C-N, C-O, and C-H, respectively [47]. The primary amine N–H bond bending is represented by the band seen at 1576 cm−1. The C = O stretching vibration in the amide group, produced by the incomplete deacetylation of chitin, is assigned to the band at 1648 cm−1 [47].

For crosslinked chitosan in film (Fig. 3c), a stretching band around 1642 cm−1 can be seen, which is of higher intensity than the similarly located C = O vibration band at 1648 cm−1 for uncrosslinked chitosan. This band corresponds to the crosslinking-typical imine bond (C = N), probably resulting from the crosslinking reaction between Glut and chitosan amino groups [47]. Also, a relative increase in the 2875 cm−1 band intensity can be related to the C-H crosslinked bond, probably overlapping with the -CH2- groups in the Glut structure [47]. Collectively, these findings confirm the efficient crosslinking in the ink polymer matrix, further validating the adopted crosslinking technique.

3D printing of composite mucoadhesive buccal films

We developed a monophasic square-shaped toroid film structure with dimensions: 10 mm × 10 mm × 2 mm. We further developed a biphasic design (upper- and lower-layer heights of 0.5 mm and 1.5 mm, respectively), for investigative analysis. The toroid architecture was selected for designing the buccal film, to allow more favourable circulation of buccal fluids, film interaction and drug release. Monophasic composite films were obtained via 3D printing of plain polymer ink, ATV@ink, PG-Lip@ink and ATV/PG-Lip@ink then drying at controlled temperature of 25 °C to obtain plain 3DP-film, ATV@3DP-film, PG-Lip@3DP-film and ATV/PG-Lip@3DP-film, respectively. Whereas the biphasic 3DP-film was printed using ATV@ink for upper layer and ATV/PG-Lip@ink for lower layer.

As shown in Fig. 4a–c, extrusion-based 3D printing provided accurate production of the tested CAD geometry. We further investigated the microstructure of monophasic ATV@3DP-film and biphasic 3DP-film using SEM. Cross sectional view of the monophasic ATV@3DP-film (Fig. 4d) showed a dense, compact structure of the applied polymeric matrix, further indicating structural integrity and efficient component homogeneity. On the other hand, cross sectional view of the biphasic film revealed distinctive upper and lower layers, where the upper layer reflected smooth compactness. Whereas the lower layer presented more porosity and matrix roughness, which might be attributed to the presence of liposomes. The clear demarcation of the upper and lower layers further verifies the reliability of the adopted 3D printing technique.

Fig. 4

3D printing of mucoadhesive buccal films (a–d). 3D computer-aided design of monophasic ATV/PG-Lip@3DP-film with dimensions 10 mm × 10 mm × 2 mm (a). Representative image of freshly prepared (b) and dried (c) ATV/PG-Lip@3DP-film. Scale bar = 10 mm. Scanning electron micrographs of the developed 3DP-films (d), illustrating top view and cross-sectional view of ATV@3DP-film, showing a relatively homogenous phase. Whereas the top view and cross-sectional view of biphasic 3DP-film show different microstructure for upper and lower phases. Scale bar = 200 µm

Characterization of the 3D-printed buccal films

The developed dried, composite films were characterized regarding swelling and time-driven disintegration based on assessing water uptake followed by monitoring change in swollen wet film weight. Film erosion was assessed based on final dry film weight. Additionally, residual moisture content and drug content uniformity were determined.

Swelling, disintegration and erosion

The degree of swelling of polymer blend is a critical parameter affecting film mucoadhesion due to detachment and relaxation of polymer chains occurring upon swelling [48].

Percentage swelling of monophasic plain 3DP-film, ATV@3DP-film, PG-Lip@3DP-film and ATV/PG-Lip@3DP-film in SSF is shown in Fig. 5a. Percentage swelling values for both plain 3DP film and ATV@3DP-film were significantly (p ≤ 0.05) higher than liposomal films (PG-Lip@3DP-film and ATV/PG-Lip@3DP-film). The effect of nanoparticles on swelling of polymeric matrices has been previously reported [18], and could be attributed to the physical crosslinking that is created between polymeric chains in the presence of liposomal vesicles, creating more compact and tighter structures. Also, the reduced swelling of liposome loaded films could be attributed to the lower hydrophilicity of liposomes compared to the polymeric matrix [49]. It is worth noting that initial swelling for both plain 3DP-film and ATV@3DP-film (342.4 ± 0.6% and 354.6 ± 1.4% at 0.5 h, respectively) was followed by an apparent decline in water uptake (230.50 ± 1.45% and 267.35 ± 2.6% at 1 h, respectively). This finding could possibly be partly referred to the diffusion of PG from the swollen films into the surrounding medium and consequent reduction of films weight. Nevertheless, this behavior was not observed for liposomal films, probably because of the relative partial confinement of PG within the liposomes in films. In addition, the longer diffusion path provided by liposomes [49] might have hindered entrapped PG diffusion, resulting in constant swelling behavior. For biphasic 3DP-film, percentage swelling values followed an intermediate pattern between ATV@3DP-film and ATV/PG-Lip@3DP-film.

Fig. 5

Characterization of the developed 3D printed films (a–e). Swelling (a) and loss in maximum swollen wet weight over time (b) of the developed 3DP-films. In vitro cumulative release profile for ATV solution and ATV/PG-Lip using the dialysis bag technique (c) and ATV@3DP-film, ATV/PG-Lip@3DP-film and biphasic film using the total immersion method (d). Release experiments were conducted at 37 °C in 5% ethanol phosphate buffer, pH 6.8. Mucoadhesive force and adhesiveness of different films (e). Data indicates mean ± SD, n = 3, p ≤ 0.05

As shown in Fig. 5a, films could undergo swelling, maintaining structural integrity for 4 h. Following swelling, films started disintegrating due to gel liquefaction which was reflected by wet weight loss. Percentage wet weight loss was recorded for different films (Fig. 5b), where films exhibited an increase in wet weight loss when compared to the maximum swollen weight due to disintegration. Both plain 3DP-film and ATV@3DP-film showed higher percentage wet weight loss (70.5 ± 3.2% and 60.8 ± 1.2% after 192 h, respectively) than liposomal films (25.8 ± 1.2% and 46.7 ± 1.5% for PG-Lip@3DP-film and ATV/PG-Lip@3DP-film, respectively). The higher weight loss for ATV/PG-Lip@3DP-film compared to PG-Lip@3DP-film can be possibly referred to ATV release. It is possible that ATV release has created over time more channels for the diffusion of the surrounding medium and consequently higher disintegration and gel liquefaction. The pattern for biphasic films wet weight loss similarly proceeded as detected for percentage swelling. This result was in line with the lower percentage swelling results for liposomal films, further pointing out the possible physical crosslinking.

Finally, percentage film erosion was determined after immersion for 8 days in SSF. Percentages loss in dry film weight for plain 3DP-film and ATV@3DP-film were 89 ± 0.6% and 87 ± 0.9% respectively. However, for PG-Lip@3DP-film and ATV/PG-Lip@3DP-film, corresponding percentages were 78 ± 1.2% and 76 ± 0.3% respectively.

Residual moisture content

Residual moisture content is an important parameter in buccal films due to its obvious impact on their stability against microbial infections [50].

The average moisture content for all tested films (plain 3DP-film, ATV@3DP-film, PG-Lip@3DP-film and ATV/PG-Lip@3DP-film) was 6.57 ± 0.75% w/w.

Drug content uniformity

Percentage drug content was utilized as an indicator of content homogeneity to measure printing accuracy and determine the effectiveness of process parameters for developing tailorable drug-loaded films with acceptable reproducibility [18].

Drug content values for ATV@3DP-film and ATV/PG-Lip@3DP-film were 96.9 ± 5.7% and 95.2 ± 0.75%, respectively, which indicated that the drug is uniformly contained in films, further validating the adopted printing procedure.

In vitro drug release

Drug release was investigated for ATV/PG-Lip compared to ATV solution using dialysis bag technique. Whereas drug release from ATV@3DP-film and ATV/PG-Lip@3DP was tested using total immersion method. For investigative analysis, drug release from biphasic 3DP-film was included.

As seen in Fig. 5c, ATV/PG-Lip presented a significantly (p ≤ 0.05) lower percentage release (48.6 ± 1.2%) compared to ATV solution (100 ± 0.3%) within 6 h. The complete diffusion of ATV from solution verifies the drug dialyzability and establishes the controlled drug release from ATV/PG-Lip. The kinetic analysis of release profiles (as detailed in Table 1S) showed that drug release from ATV/PG-Lip fitted Higuchi model, possibly implying ATV diffusion from liposomal bilayer.

Drug release study from the developed films demonstrated that ATV release rate for all time points increased in the following pattern: ATV@3DP-film > biphasic 3DP-film > ATV/PG-Lip@3DP-film (Fig. 5d). More specifically, approximately 50% ATV release was achieved after around 2, 4 and 6 h for ATV@3DP-film, biphasic 3DP-film and ATV/PG-Lip@3DP-film, respectively. These results could be interpreted considering percentage swelling for films (Swelling, disintegration and erosion), where the fluid uptake by ATV/PG-Lip@3DP-film was lower than that for ATV@3DP-film, probably contributing to the more controlled drug release from the former.

The intermediate drug release rate from the biphasic 3DP-film, compared to ATV@3DP-film and ATV/PG-Lip@3DP-film, is expectedly related to the cumulative ATV release at different rates from both ATV@ink and ATV/PG-Lip@ink constituting the biphasic 3DP-film structure. These results further establish 3D printing as reliable customization tool for tailoring drug release profiles.

As detailed in Table 1S, drug release from ATV@3DP-film followed first order kinetics. Whereas kinetic analysis of the release profiles indicated that the best-fit model for both ATV/PG-Lip@3DP and biphasic film was the Higuchi diffusion model, which was consistent with various studies of bucco-adhesive films [51]. Korsmeyer-Peppas n values (Table 1S) supported diffusion-driven release.

Mucoadhesive properties

Mucoadhesive films represent an efficient buccal pharmaceutical form; they provide retention for a longer period of time. In our work, we developed composite mucoadhesive 3DP-films, utilizing both HPMC and chitosan as mucoadhesive polymers [52]. Chitosan possesses mucoadhesive potential through the electrostatic interaction between its cationic amine groups and negatively charged mucin molecule [53]. However, it has some restrictions, which are mainly because of its low solubility and the weak mechanical strength of the formed gels. The mechanical strength of chitosan can be enhanced by blending with other polymers such as HPMC [54]. HPMC mucoadhesiveness is attributed to its non-ionic hydrophilicity leading to the diffusion and formation of interpenetration layer with mucus [53].

As shown in Fig. 5e, all the developed 3DP-films achieved an average mucoadhesive force of (2388.4 ± 18.4 dyne /cm2) and adhesiveness of (0.82 ± 0.025 mJ). Furthermore, mucoadhesion residence time for all the prepared films was beyond 24 h, indicating that inclusion of liposomal formulations within the polymer matrix did not affect the mucoadhesive attributes for liposomal films (PG-Lip@3DP-film and ATV/PG-Lip@3DP-film).

Taken together, the developed composite 3DP mucoadhesive films could afford reasonable swelling, acceptably controlled release and efficient mucoadhesive features. Thus, they would present good candidates for controlled buccal drug delivery.

In vitro antifungal activity

In this work, we investigated the effect of the developed formulations on the reported ATV in vitro antifungal activity.

Determination of minimum inhibitory concentration (MIC)

The MIC of ATV solution and ATV/PG-Lip against four different fluconazole-resistant candida strains (two standard and two clinical isolates) was determined using agar dilution technique.

As can be seen in Table 3, both ATV solution and ATV/PG-Lip demonstrated antifungal activity against all tested strains. The determined MIC for ATV solution against C. albicans ATCC 10231 (32 µg/mL) agreed with previous data [10], while the MIC for ATV/PG-Lip was 128 µg/mL. The determined MIC against both C. albicans and C. parapsilosis clinical isolates for ATV solution and ATV/PG-Lip were 64 and 256 µg/mL, respectively.

Table 3 MIC of ATV solution and ATV/PG-Lip against different Candida strains

The higher MIC for ATV/PG-Lip compared to ATV solution can be attributed to the confinement of ATV in the liposomal vesicle (ATV/PG-Lip) which is further immobilized within the agar gel matrix, as imposed by the experimental setting. In addition, the lower diffusion of liposomal vesicles compared to ATV solution through the agar gel might have resulted in the lower overall interaction between drug in ATV/PG-Lip and candida cells. This result highlights the challenges facing microbiological testing of nanoparticle dispersions. It also confirms the need for other elaborate testing adopting different microbiological techniques, as performed in our study.

Time-dependent antifungal activity

The growth curve for C. albicans ATCC 10231 was plotted to monitor the antifungal activity of ATV and ATV/PG-Lip (at MIC of 32 and 128 µg/mL, respectively) over time.

Results clearly showed that percentages fungal growth was in order of control untreated candida > ATV solution > ATV/PG-Lip. Moreover, the fungal growth of ATV/PG-Lip (262.3 ± 4%) was significantly (p ≤ 0.05) lower than ATV solution (405 ± 5%) after 24 h (Fig. 6a). This finding may be due to the high solubilization of ATV as ATV/PG-Lip and its subsequent sustained release. Also, this may be attributed to the ability of vesicles to bind to fungal cell wall, facilitating the drug penetration to fungal cells [55].

Fig. 6