We first synthesized the chC10 and unveiled its ionic liquid crystalline mesophase, which underpins the desirable gel-like behaviour of the material ideal for loading into the SUF cavities. The formulation was then loaded into foil cavities, and its compatibility with the PDMS-based SUF was subsequently evaluated.

The ionogel is an ionic liquid crystal

Synthesis of chC10 was conducted by a salt metathesis reaction between the choline bicarbonate and decanoic acid, with the subsequent evaporation of CO2 to serve as an indicator for the progression of the reaction, as previously described and characterized (Fig. 1A) [32]. ChC10 is composed of choline and decanoate as the positively and negatively charged counterions together with an additional protonated mol of decanoic acid. Predominantly, the negatively charged decanoate and positively charged choline interact via Coulombic forces (ionic interactions) and the additional protonated acid interacts in the mixture via hydrogen bonding and Van der Waals interactions. The successful synthesis is confirmed by the depression of its melting point and the absence of the melting endotherm for decanoic acid. The melting endotherm of the ionogel was previously reported at − 2 °C and the water content of the insulin-loaded ionogel after freeze-drying is ~ 3% (w/w) [32]. The high viscoelasticity was previously described by rheological analysis with chC10 to resemble a gel-like material and thus the term ionogel [32]. Here, we show that its high viscoelasticity is also attributed to the structural order of the material shown by the birefringence in images obtained by polarized optical microscope (POM) and the characteristic Bragg peaks in the wide-angle X-ray scattering (WAXS) diffractograms (Fig. 1B and C). The presence of diffraction peaks at q2 = 2q1, q3 = 3q1 and q4 = 4q1 (q1 = 0.23, q2 = 0.46, q3 = 0.69, q4 = 0.92) is characteristic for smectic liquid crystal mesophases, indicating a well-defined lamellar ordering. The halo in the high q range is characteristic of the ‘liquid’ fatty acid domains in each smectic layer (Fig. 1C) [35]. The interlayer thickness (d) was determined to be 2.75 nm, which is consistent with the theoretical length of two fully extended decanoate molecules forming a bilayer, together with two choline molecules shared between adjacent layers, yielding a repeated smectic layer of approximately 3 nm. The slightly reduced d-spacing (2.75 nm vs. 3 nm) is characteristic of smectic IL crystals and is typically attributed to conformational kinking or interpenetration of the fatty acid alkyl chains, as well as shortened ionic and hydrogen bond distances arising from strong intermolecular interactions [36,37,38]. Figure 1D illustrates the suggested packing of the ionogel with the fatty acid tails interpenetrating the bilayer. Furthermore, to confirm the WAXS data, we aimed to identify the characteristic textures of smectic IL crystals by POM. Focal conic domains and Maltese crosses can be identified in the images (Fig. 1E) indicative of smectic liquid crystalline materials [39, 40]. IL crystals represent an intermediate state of matter between the solid and liquid phases, exhibiting partial structural order. Smectic liquid crystalline mesophases are typically characterized by a lamellar arrangement, in which molecules are organized into layered packings. Within these layers, molecular mobility is retained, while diffusion across layers remains restricted [35]. Hence, the chC10 ionogel forms a stable smectic liquid crystalline phase above its melting point, and the viscoelasticity mitigates the rapid in vivo dilution of liquid ILs, while its higher density compared to powder-loaded SUFs enables superior loading capacity [33].

The bulk density of pharmaceutical powders, such as the previously loaded powdered mixture of insulin: SDS: STI, typically ranges from 0.1 to 0.7 g/mL. Notably, the density of ILs has been reported to be approximately 1 g/mL for CAGE or higher for other carboxylate-based ILs while smectic crystalline mesophases show even higher densities due to their structural order. Consequently, the ionogel formulation is expected to increase the loading capacity of the formulation in the foil at least two-fold compared to previous loaded powder formulations, thereby enabling optimal drug loading and enhanced efficacy of the combined dosage form [41,42,43,44].

Fig. 1

Synthesis and characterization of decanoate 1:2 molar ratio (chC10) ionogel. A Synthesis of chC10 and visual appearance of the characteristic gel-like texture. B Polarized images of chC10 illustrating the birefringence typical of structured materials, N = 3, (C) WAXS scattering pattern for chC10 with the periodic peaks indicating a smectic ionic liquid crystalline mesophase. N = 2 (D) Schematic representation illustrating the suggested packing of the chC10. E Polarized image of chC10 in darkfield demonstrating Maltese crosses and focal conic domains, characteristic textures of smectic liquid crystals, N = 3

Incomplete release of formulation from the combined dosage form

The feasibility of loading the ionogel into SUF devices and release of the formulation from the foil was tested using the previously developed SUF with honeycomb microcavities (Fig. 2A). In this cavity design, used in prior studies, 78% of the device surface area is available for drug loading [14]. The insulin-loaded ionogel was deposited to fill the surface-patterned microcavities of the SUF, with excess material removed by scraping. Figure 2A and B display SEM images before and after loading the ionogel in the honeycomb microcavities of the SUF. After confirming successful loading of the chC10 formulation using SEM, we assessed the release of insulin and chC10 from the cavities of the SUF. Figure 2C depicts almost complete release of the insulin (90%) and full release of chC10 (100%) from the formulation not loaded into the SUF, in accordance with previous studies [32]. However, the ionogel-loaded device exhibited incomplete release of the formulation in the dissolution media, with only 41% of insulin and 63% of chC10 released over 6 h, reaching a plateau within 20 min. To investigate the underlying cause of this limited release, Raman confocal microspectroscopy was employed. First, Raman spectra from the empty SUF, the insulin-loaded chC10 formulation, and the formulation-loaded SUFs were acquired (Figure S4A). Samples retrieved after the release study were mapped using the dominant Raman bands of the SUF (∼492 cm⁻¹, Fig. 2D) and the formulation (∼1450 cm⁻¹, Fig. 2E). As shown in Fig. 2F, Raman spectroscopy revealed residual material along the microcavity edges, consistent with the formulation, as evidenced by the residue spectrum matching that of the formulation-loaded SUF (Figure S4B). We hypothesized that the incomplete release arose from poor wettability of the PDMS foil due to its hydrophobic nature, particularly in the deeper regions of the cavities.

Fig. 2

Visualization of the honeycomb microcavities loading and release of the formulation from SUF. Representative SEM images showing honeycomb microcavity design (A) before and (B) after loading the microcavities with the formulation. Scale bars represent 500 μm. C Release profile of choline decanoate and insulin from formulation loaded honeycomb cavity device (n = 3) and in the absence of the SUF as control (n = 2), data represent mean ± SD for the honeycomb microstructure (SUF) and mean ± range for the ionic liquid (control). Raman confocal microscopy of the devices after the release studies (D-E). The intensity maps of (D) the foil’s PDMS (492 cm-1) and (E) the insulin-loaded ionogel (1450 cm-1) were used to generate (F) a combined component map, where the dominant contributions of PDMS and the formulation are shown in blue and green, respectively. Scale bars represent 200 μm

Cavity designs improve formulation wettability and insulin release

In addition to the honeycomb design, two new cavity structures were developed. The first design, referred to as pyramidal cavities, features pyramidal frustums with side lengths of 400 μm and a depth of 150 μm. These are shown in comparison to the honeycomb structure in Fig. 3A and B. The second design, denoted as large cavities, consists of 4.25 × 4.25 mm cavities with depths of 100 μm, as seen in Fig. 3C.

Fig. 3

Schematic representation of the cavity designs used in the current study. A Schematic of honeycomb cavity design, dH=127 μm, WH=200 μm [13, 14], (B) schematic of pyramidal frustum cavity design, dP =150 μm, Δ = 447 μm, θ = 54.7, ws=47 μm and (C) schematic of large cavity design, dL =100 μm, WL = 4.25 mm

Compared to the original honeycomb design, the two new cavity designs possess either a tapered sidewall (54.7° vs. 90°) to improve cavity wetting and decrease mechanical adhesion of the ionogel or a larger cavity (4.25 mm vs. 200 μm) (Fig. 3). With these new designs, we aimed to unveil whether the aspect ratio, sidewall taper angle, cavity surface roughness and the depth of the cavities impact the wetting and subsequently the release of the formulation. Notably, the new cavity dimensions were selected to produce similar theoretical loading volumes across the two designs (B and C). For 15 × 50 × 0.95 mm sized SUFs used for the dissolution testing, the theoretical loading volumes were 71, 50, and 54 µL for the honeycomb, pyramidal, and large cavity designs, respectively (Table S1). Thus, we selected the new devices with more similar loading volumes to investigate and optimize the release of insulin and chC10 from the foil cavities.

The chC10 release profiles show that the large cavity design released a larger proportion of the chC10 (83%) than the other designs reaching a maximum release after 30 min (Fig. 4A). These results are more comparable to the formulation control group, which demonstrated complete release (100%) of chC 10 within 30 min. Insulin release followed the same trend as the dissolution of the chC10 (Fig. 4B). The large cavity design improved insulin release, with 74% release after 30 min compared to only 41% and 57% release for the honeycomb and pyramidal designs, respectively. Additionally, a different release profile is noticeable from the pyramidal compared to the large cavities design. The release profile of chC10 and insulin from pyramidal cavities follows a fast release until 30 min, as observed with the other tested designs; however, the release slows down over time, likely due to reduced dissolution deeper within the cavity, where the available surface area decreases. This observation suggests promising potential for designing microcavities that enable customized release profiles in future applications.

To support these findings, Raman microspectroscopy was performed to validate the release data. Raman imaging combined with optical microscopy revealed substantial residual material within the pyramidal cavities, particularly along the edges, with spectra matching that of the formulation (Fig. 4C and D), explaining the lower release. In contrast, no significant formulation residue was detected in the large cavity devices, confirming higher release efficiency (Fig. 4E and F). However, release from these cavities remained ~ 15% lower than the formulation control for both insulin and chC10, indicating incomplete release despite the absence of visible residue.

To investigate the reduced release from larger cavities contradicting the Raman microscopy findings, we hypothesized that the effect is not solely due to poor wetting from the small cavity geometry and the surface hydrophobicity but is also influenced by surface adsorption of insulin and chC10 onto the hydrophobic foil. To test this, insulin-loaded ionogel was predissolved and incubated in the same setup as in the release studies, with and without an empty foil present. Interestingly, the recovery of chC10 was lower when placed in a vessel with a foil, demonstrating 91% recovery after 30 min and 82% for the rest of the 2 h study. On the other hand, chC10 added to the vessel without a foil showed a recovery above 96% for the whole 2 h of the study (Figure S5A). These values are in good agreement with the lower amount of chC10 released from the large cavities (83%) compared to the formulation control (100%).

Even more detrimental effects were observed in terms of insulin recovery. Insulin recovery was ≈ 90% after 30 min for both in absence or presence of the foil. However, a drastic decrease was observed after 60 min, much more pronounced in the foil-containing vessels (48% recovery) compared to the vessels without the foil (70% recovery) (Figure S5B). This significant and continuous decrease in insulin recovery underlines a higher loss of insulin in the foil group, probably due to insulin adsorption onto the foil surface. The adhesion of peptides, proteins, and hydrophobic molecules to PDMS is a well-recognized challenge [45], which likely accounts for the observed insulin loss in the vessel with the SUF. This effect is further exacerbated at low insulin concentrations, as employed in this study, where adsorption even to the dissolution vessel becomes increasingly significant [45].

Fig. 4

Release studies from different cavity designs. A chC10 release from honeycomb, pyramidal, and large cavities and the ionogel (control). B Insulin release from honeycomb, pyramidal and large cavities and the ionogel (control). All release data represent mean ± SD for the tested SUFs and mean ± range for the control group n = 3 for all tested groups and n = 2 for the control group. C Optical microscopy images (left) and (D) confocal Raman microspectroscopy maps (right) from pyramidal cavities, scale bars 200 μm. E Optical microscopy image scale bar represents 400 μm and (F) confocal Raman microspectroscopy maps (right) large cavity devices, scale bar represents 400 μm. Raman images depict with green colour the remaining ionic liquid formulation, while blue represents the PDMS foil material

DBE-712 addition makes the PDMS SUF hydrophilic

Noteworthily, in addition to the hypothesis that cavity morphology impacts the wetting of device cavities, we further confirmed that adsorption of insulin and chC10 hampers their complete release. Hence, we chemically modified the PDMS SUF to improve wetting of the cavities and minimize the adsorption of insulin and chC10. Previous work used ultraviolet-ozone and polyvinyl alcohol (PVA) surface treatments to increase the hydrophilicity of a PDMS SUF [46]. This was done to obtain good adhesion between the SUF device and a protective top coating as well as improved wetting of the microscale compartments into which powder formulations were loaded; In this study, we aimed to chemically modify the PDMS to enable a fast, simple and uniform alteration of the surface properties of the devices [45].

To assess the impact of modifying the surface properties of PDMS, DBE-712 was added during the fabrication of the devices. DBE-712 has been shown to increase hydrophilicity and thereby reduce the fouling of peptides on the surface of microfluidic devices [46]. The impact of DBE-712 on modified PDMS SUFs was characterized by WCA measurement. A significant decrease in WCA was seen when comparing modified PDMS with plain PDMS (Fig. 5A). Most importantly, a rapid decrease in WCA from 110° to 70° (p < 0.0006) occurred in under 2 min for the modified PDMS. Additionally, the WCA decreased over time, to 21° and 78° (p < 0.0006) at 32 min for samples with and without DBE-712, respectively. The sharp decrease in WCA for the modified PDMS foil proves that adding the block co-polymer significantly impacts the surface free energy of the foil, making it more hydrophilic which enhances wetting and reduces adhesion of insulin and chC10.

After successfully improving the hydrophilicity of the foil, it was essential to verify the mechanical properties of the modified PDMS. The elastic modulus, necessary for the unfolding of the foil and the distention of the mucosa, was changed because of the addition of DBE-712 with a significant reduction (p < 0.0002) from 1.87 ± 0.08 to 1.24 ± 0.03 MPa (Fig. 5B). This reduction was expected as the addition of surfactants previously has shown to induce plasticizing effects [47]. Nevertheless, the mechanical properties were considered suitable due to their similarity to those from Ghavami et al., in which PDMS foils with elastic moduli of 1.12 MPa were used. The higher elastic modulus of the plain PDMS sample measured in our study is attributed to the higher temperature used for curing (90 °C for 16 h vs. 37 °C for 24 h) [48].

Modified PDMS foil prevents adsorption and improves release of the formulation

Release testing was performed identically to previous trials for comparison between the devices with and without the DBE-712 modification. As expected from our previous data and the significantly improved release from large cavities, the effect of modified PDMS was more noticeable for the pyramidal microcavities. A considerably faster release was evident from the modified pyramidal SUFs compared to the plain PDMS. The difference was observed already after 10 min as 80% of the chC10 was released from the modified PDMS SUF compared to 28% (p < 0.0013) from the non-modified PDMS (Fig. 5C). Most importantly, after 10 min of the study, complete insulin release was already observed from the DBE-712-modified PDMS compared to only 26% (p < 0.006) from the PDMS, showcasing that the hydrophilic surface not only prevented adhesion of insulin but also improved significantly the release rate of the formulation (Fig. 5D).

On the other hand, for the large cavities, there was no improvement of the chC10 release confirming the hypothesis that the wetting was mainly a problem for the microcavity designs with only slightly faster release observed for the large cavity modified PDMS (75%) compared to non-modified (53%) (p < 0.031) SUF for the first 10 min of the study (Fig. 5E). Interestingly, the modified SUF demonstrated improved release of insulin for both the pyramidal and the large cavity foils of approximately 100% and a faster release (p < 0.002) at the first timepoint of the study (Fig. 5F). This confirms the hypothesis of surface adsorption of insulin onto the PDMS foil and the improved wettability from the addition of DBE-712. Similarly, Raman microspectroscopy was used on pyramidal cavities to image the formulation-loaded SUF before and after the release testing (Fig. 5G and I). We detected a lower presence of formulation within the modified pyramidal microcavities, confirming our successful optimization of the foil and validating our hypothesis.

Fig. 5

Modified PDMS foil characterization and release performance. A Water contact angle (WCA) over time of PDMS and modified PDMS. The first part of graph displays the first 2 min of testing. Data represent mean ± SD (n = 3). B Tensile testing of modified (blue colour) and non-modified PDMS (green colour) with comparison of Young’s modulus and strain at failure. Data represent mean ± SD (n = 5). C Insulin release from modified and PDMS SUFs with pyramidal cavities. D chC10 release from modified and non PDMS SUFs with pyramidal cavities. E ChC10 release from SUFs prepared with modified and plain PDMS devices large cavity designs. F Insulin release from modified and PDMS SUFs pyramidal microcavity designs. All release data represent mean ± SD, n = 3. G Raman analysis of the modified PDMS device with pyramidal cavities after conducting the release experiment. The intensity map of the spectral band associated with (G) PDMS (492 cm-1), (H) insulin-loaded ionogel (1450 cm-1), and (I) the combination of both, reveals a reduction in the amount of residual formulation after the hydrophylic modification of the foil, scale bars represent 100 μm

Pyramidal cavity SUFs loaded with ionogel formulation significantly enhance insulin bioavailability

Following the previous in vivo studies conducted by Ghavami et al., it was necessary to fabricate smaller devices suitable for dosing in rodents [14]. Two cavity designs were considered when producing SUF devices with dimensions 10 × 15 × 0.45 mm (Fig. 6A). Honeycomb microcavities were not tested in the current study as the aim was to assess the impact of cavity dimensions on the in vivo performance of the combined dosage form and thus comparable loadings of the formulation were needed. For the optimal fit of the large cavities in the new dimensions of the foil, the large cavities were modified to 4.25 × 7 mm (Fig. 6A). Capsules containing the combined dosage forms with either pyramidal or large cavity SUFs along with the ionogel formulation and an empty foil control group were administered rectally, and blood glucose was monitored (Fig. 6B). Except for the empty-foil control, all groups showed a decline in glucose levels. The s.c. injection group demonstrated the most substantial reduction in blood glucose (~ 64% at 60 min). Among the foil designs, the combined dosage form with pyramidal microcavities produced the greatest blood glucose lowering (35%), compared with 18% for large cavities and 11% for the ionogel formulation without the foil. Most importantly, the BP achieved by the pyramidal microcavities was significantly higher (20.5%) compared to that of dosing using the large cavities design (12.2%, p = 0.022) and very significantly superior to the effect observed with the ionogel group (4.4%, p = 0.0006) (Fig. 6C). The blood glucose data indicate a highly improved permeation of insulin from the ionogel formulation when loaded in the foil, and interestingly, a significant impact of the cavity design.

Plasma insulin measurements corroborated the blood glucose findings. Consistent with the glucose data, plasma insulin increased sharply within 15 min after administering the pyramidal combined dosage form loaded with the ionogel, attaining levels equivalent to the s.c. injection and significantly greater than those observed in the other groups (Fig. 6D). The SUF with pyramidal microcavities achieved a relative bioavailability of 12.2%, compared with 5.2% for the large cavity SUF and 3.8% for the ionogel group, all administered with equal insulin-loaded ionogel doses confirming the superior performance of the ionogel-loaded SUF (Table 1). The AUC from the PK profiles showed statistical superiority (p = 0.045) only when compared to the ionogel group and not to the group receiving large-cavities SUFs (p < 0.11), likely due to a relatively higher variation in the data. However, the PD data showed statistical superiority of the pyramidal SUFs over both the large cavity SUF (p = 0.022) and the ionogel group (p = 0.0006) (Table 1).

Additionally, Cmax in rats receiving foils with the pyramidal cavities was 5-fold higher than the Cmax in the rats dosed with the large cavities SUFs (p = 0.0008) and 2.5-fold higher than the ionogel group (p = 0.0021) (Table 1), further validating the higher performance of the ionogel-loaded SUFs with the pyramidal microcavity design. A shorter Tmax was observed in the groups dosed with the SUFs with a median of 20 and 40 min for the pyramidal and large cavity designs, respectively, compared to 60 min for the ionogel group (Table 1). This shorter Tmax is likely attributed to the faster disintegration of the capsule due to the force exerted by the foil. Moreover, pyramidal SUFs and the s.c. administration led to comparable Cmax values of 158 mU/L and 140.5 mU/L (Table 1), respectively, further confirming the high efficacy of the combined dosage form with pyramidal microcavities in promoting insulin absorption.

Altogether, the short Tmax and high Cmax are consistent with our release data and support the premise that optimizing cavity geometry and surface properties is critical for enhancing absorption. Although additional studies are required to fully elucidate why the pyramidal cavity design outperforms the larger cavities, several mechanistic factors are likely involved. Notably, the pyramidal microcavities are four times deeper than the large cavities (400 μm vs. 100 μm), which increases the likelihood that a greater proportion of the formulation remains sequestered within the cavity during foil wrapping. In contrast, the shallower large cavities allow more of the formulation to contact the capsule shell dissolving together with the capsule and preventing the optimal interface of the formulation with the mucosa. Consequently, a higher payload is retained within the pyramidal microcavities at the time of capsule disintegration, plausibly contributing to their superior PK/PD performance.

In the current study, we demonstrated a pronounced synergistic effect of combining an ionogel formulation with a SUF. Notably, Ghavami et al. tested a honeycomb microcavity foil loaded with 6.7 mg of a powder formulation containing insulin, SDS and STI, and this resulted in a relative BA of 1.8% [14]. Here, the pyramidal microcavities loaded with chC10 showed a 6.7-fold higher relative BA, underscoring the superior performance and anticipated synergy of the ionogel formulation in combination with the SUF. However, direct comparison between the two studies should be done with caution due to the different amounts of insulin dosed. Jorgensen et al. dosed a miniaturized SUF with dimensions 7 × 7 mm to the jejunum of the rat with insulin, SDS, and STI in lower amounts than the 6.7 mg used by Ghavami et al. and achieved a relative BA of 0.12% [13]. The different sites of administration prevent a direct comparison but obviously the 100-fold improvement in BA signifies the efficacy of the ionogel formulation with the SUF. Rectal administration of a solution containing eicosapentaenoic acid (EPA) and insulin yielded a BP of 6%. When EPA and insulin were incorporated into a Pluronic® F-127 in situ gelling hydrogel, BP increased markedly to 27%. However, the formulation in that study was administered as a solution directly in the rectum of the rat [49] and not in a capsule as in the current study. In addition, rectal infusion of insulin with tauro-24,25-dihydrofusidate demonstrated 6% relative BA [50]. In another study, various permeation enhancers were delivered via a hydrogel directly to the rectum of the rats with dodecanoic acid (C12) achieving the highest BA of 7% while C10, octanoic acid (C8), and myristic acid (C14) demonstrated a 3.9%, 5.3%, and 2% BA, respectively [49]. In comparison, the relative BA and BP achieved in the current study,12.2% and 20.5%, respectively, are among the highest reported for non-injectable insulin delivery. Due to complications in miniaturizing the specific SUFs to fit in size 9 capsules suitable for oral administration to rats, oral dosing was not possible. Given our objective to evaluate the enhancement in absorption achieved by combining an advanced formulation strategy with the SUF platform, we opted for rectal administration as a proof-of-concept approach. This route enabled the use of larger capsules compatible with the rat model, allowing for administration of relevant amounts of the loaded formulation. Consequently, it facilitated a preliminary evaluation of the delivery system’s performance prior to advancing to more physiologically relevant animal models for oral administration.

Nevertheless, the limitations of the current model should be acknowledged, and the bioavailability values should be interpreted with caution, as the present study serves primarily as a mechanistic proof-of-concept. Permeability in small intestinal regions relevant to oral delivery may differ substantially from that observed in the rectal model [51]. In addition, protease levels in duodenal and jejunal fluids are substantially higher than in the rectum, where pancreatic proteases are absent, and may therefore promote more rapid peptide degradation. Finally, anaesthesia can affect intestinal motility, thereby prolonging the contact time between the combined dosage form and the mucosa, which may further contribute to an increased apparent bioavailability.

The physiological differences between rectal and oral administration must be considered when interpreting the results. The middle and inferior rectal veins drain into the systemic circulation via the inferior vena cava, bypassing hepatic first-pass metabolism. In contrast, the superior rectal vein drains into the portal vein, subjecting absorbed compounds to hepatic metabolism. Insulin was likely absorbed through both pathways, partially avoiding first-pass metabolism and contributing to the elevated plasma concentrations observed [52]. The consistently lower BA relative to BP across all groups is notable and aligns with prior findings in oral insulin delivery, likely reflecting differences in biodistribution. Oral absorption directs insulin to the liver, mimicking its physiological action but reducing systemic levels due to hepatic degradation, whereas s.c. administration bypasses the liver, resulting in higher peripheral plasma concentrations [53].

Fig. 6

In vivo data after rectal administration of the combined dosage form in a gelatin capsule. A Schemantic representation of the in vivo experiment. Computer-aided design (CAD) images showing the foils with pyramidal or large cavities produced for the study. Large cavity dimensions were modified to 4.5 × 7 mm for optimal fit to 15 × 10 mm size foils. B Pharmacodynamic data (PD) of blood glucose lowering (BGL) percentage is normalized to the blood glucose before dosing. C Area above the curve (AOC) of the BGL curves. D Pharmacokinetic data (PK) of insulin plasma concentration (Cins). E Area under the curve (AUC) for the insulin PK profiles. Data are presented as mean + SEM, N = 4, except for the subcutaneous (s.c.) administration and the empty foil controls where data represent mean + range, N = 2

Table 1 Pharmacokinetic and pharmacodynamic data for insulin dosed in the chC10 formulation alone (N = 4) or loaded in foils with pyramidal cavities (N = 4) or with large cavities (N = 4)

The next step is to evaluate the combined dosage form with pyramidal cavities in pigs, whose larger intestinal diameter allows testing of foils with greater dimensions and loading, closer to human application. Liquid formulations, including ILs, are known to dilute rapidly in vivo, affecting performance and translatability in larger animals with higher volumes of intestinal fluids [33]. Although chC10 is highly viscoelastic and hydrophobic, the SUF’s unidirectional release and protection against dilution are expected to be more effective in larger animal models, potentially leading to even greater absorption improvements.

While the current concept shows strong potential by significantly improving BA compared to previous foil-based devices, several limitations must be addressed before advancing the concept forward [13, 14]. Device retention under physiological conditions remains a concern, as some SUFs were excreted shortly after administration. Future work should focus on optimizing the device-mucosa interface and enhancing mucoad