Cannabis sativa plant material, Białobrzeskie, Tygra, Henola varieties, was donated from the Experimental Station for the Cultivar Testing in Chrząstowo, belonging to the Research Centre for Cultivar Testing in Słupia Wielka. The agricultural details are presented in Supplementary Materials. The plant material for the study was collected after hemp plants reached the maturation phase, i.e., from the moment of seed formation to the first seed. Immediately after collection, two samples of 500 g each were separated and dried to an absolutely dry mass. The entire drying period lasted twenty hours. The temperature in the oven was maintained at no higher than 50 °C for the first 6 hours and the oven temperature was maintained at 105 °C for the remaining 14 h of drying.

Food-grade CO₂ was provided by Air Liquide Polska (Cracow, Poland). Soluplus^® (polyvinyl caprolactam-polyvinyl acetate-polyethylene glycol graft copolymer), was supplied by BASF SE (Ludwigshafen, Germany). Neusilin US2 (magnesium aluminometasilicate) was kindly provided by Fuji Chemical Industry (Minato, Tokyo). Cannabinoid standards (CBD–CAS: 13956-29-1, CBDA–CAS: 1244-58-2, and CBC–CAS: 20675-51-8) were purchased from Sigma-Aldrich (Poznan, Poland). Trifluoroacetic acid and acetonitrile (high-performance liquid chromatography [HPLC] grade) were provided by Merck (Darmstadt, Germany). The chemicals 2,2-Diphenyl-1-picrylhydrazyl, iron (III) chloride hexahydrate, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid), neocuproine, 2,4,6-Tri(2-pyridyl)-s-triazine, trolox, Trizma^® Base, Trizma^® hydrochloride, butyrylcholine iodide, acetylcholine iodide, acetylcholinesterase, butyrylcholinesterase, 5,5-dithiobis-2-nitrobenzoic acid, tyrosinase, galantamine, azelaic acid were purchased from Sigma-Aldrich (Schnelldorf, Germany). Sodium chloride, sodium dihydrogen phosphate, sodium hydrogen phosphate, and dimethyl sulfoxide were obtained from Avantor Performance Materials (Gliwice, Poland). Ammonium acetate, an analytical weighed amount of HCl, 1 N, and methanol were supplied by Chempur (Piekary Śląskie, Poland). Cupric chloride dihydrate, acetic acid (99.5%), and ethanol (96%) were supplied by POCH (Gliwice, Poland). Prisma HT, GIT, BBB lipid solution, an acceptor sink buffer, and a brain sink buffer were supplied by Pion Inc. (Forest Row, East Sussex, UK). High-quality pure water was prepared using a Direct-Q 3 UV purification system (Millipore, Molsheim, France; model Exil SA 67120). FaSSIF and FeSSIF were purchased from Biorelevant (London, UK).

The extract of Cannabis sativa inflorescences was obtained using the dynamic supercritical CO₂ (scCO₂) extraction process (SFT-120, shim-pol, Izabelin, Polska). In total, 6.5 g of dried plant material was placed in the extraction vessel and extracted under 6000 psi at 50 °C with 250 mL of CO₂. The extraction yield was calculated as the mass of extract obtained and subjected to drying (to remove any water from the eventually frozen needle) (g) divided by the mass (g) of plant material placed in the extractor and expressed as a percentage (%). The choice of the Henola extract (HiE) was based on the screening studies on three varieties (Białobrzeskie, Tygra, and Henola) of leaves and inflorescences and their neuroprotective potential (data not presented). After extraction, the antioxidant studies and inhibition of enzymes (acetylcholinesterase, butyrylcholinesterase, and tyrosinase) connected with neurodegeneration were repeated.

Next, the extracts were dried in a vacuum at 50 °C, weighed, and suspended in methanol (if the process was repeated to obtain more extract, at this stage the extracts were combined together), winterized, and filtered (Figure 2). For fluid extracts (HiE), carriers (Neusilin US2, Soluplus, or lactose for apparent solubility study) were added in a 1:1 mass ratio to the earlier weight of the extract. Systems were dried on rota-vapor at 50 °C until dry and grounded in mortar.

The cannabinoid profile (CBD, CBDA, and CBC) of the extract, and during the apparent solubility and permeability study, was analyzed using the ultra-high-performance liquid chromatography with the diode array detector (HPLC-DAD) method, Shimadzu Corp. (Kyoto, Japan). The previously described method was used [37]. The analysis was conducted on a CORTECS Shield RP18 stationary phase, 2.7 µm; 150 mm × 4.6 mm, with a mobile phase consisting of 0.1% trifluoroacetic acid (41%) and acetonitrile (41:59, v/v). The flow rate was set to 2.0 mL/min, and the column temperature was maintained at 35 °C. The injection volume was 10.0 µL, and the detection wavelength was set at 228 nm, with an analysis time of 50 min. The retention times for each cannabinoid were as follows: CBD at approximately 5.83 min, CBDA at approximately 6.42 min, and CBC at 14.57 min. The LabSolutions LC software (version 1.86 SP2) from Shimadzu Corp. (Kyoto, Japan) was used to obtain chromatograms. The method was validated according to ICH guidelines for current research, the validation parameters are collected in Table S1 (Supplementary Materials).

The dissolution rate was determined in the paddle apparatus (Agilent Technologies, Santa Clara, CA, USA). HiE had a thick, oily consistency, so for the purpose of apparent solubility study it was combined with lactose; the preparation steps were the same as for Neusilin US2 and Soluplus (HiE–control). The systems and control (600 mg) were placed into two gelatin capsules. The capsules were placed into coiled sinkers for floating prevention. The test was carried out in triplicate for 180 min in a pH 1.2 of 0.1 N hydrochloric acid, a pH 6.8 of phosphate buffer, Fasted State Simulated Intestinal Fluid (FaSSIF), and Fed State Simulated Intestinal Fluid (FeSSIF).

FaSSIF and FeSSIF dissolution media are more complex solutions specifically designed to simulate the conditions of the human small intestine under fasted and fed conditions. FaSSIF and FeSSIF contain natural surfactants present in the gut to simulate gastrointestinal fluids much more accurately than conventional dissolution media, and they simulate the conditions of the human intestine in a fasted state and after a meal [50]. Sodium taurocholate is included to replicate the role of bile acids in facilitating lipid absorption and emulsification. Lecithin is incorporated to mimic the presence of phospholipids, which play a vital role in the formation of mixed micelles that enhance the solubilization of lipophilic compounds. The buffer ensures a stable pH in the intestinal fluid (6.5 for FaSSIF and 5.0 for FeSSIF), and sodium chloride is added to ensure physiological osmolarity (a FaSSIF of 270 Osm/L and FeSSIF of 670 Osm/L) [51].

The vessels were filled with 500 mL of media at the temperature set at 310.15 K and the rotation speed of 100 rpm. At specific time intervals, 2.0 mL of the sample was taken out and immediately replaced with an equal amount of fresh medium at the same temperature. The percentage cumulative cannabinoid release (% CBD, CBDA, and CBC) was measured at different time points (5, 10, 15, 30, 45, 60, 90, 120, and 180 min) for each formulation. The samples were then passed through a filter with a pore size of 0.22 μm and analyzed using high-performance liquid chromatography (HPLC). Sample chromatograms from the dissolution study are presented in Figure S1 (Supplementary Materials). The standard deviation (SD) was also calculated for each time point and delivery system.

The differences and similarities between the apparent solubility profiles were determined by the two-factor values, f₁ and f₂, introduced by Moore and Flanner [52] with the use of the following equations:

where n is the number of time points, R_j is the percentage of the reference dissolved substance in the medium, T_j is the percentage of the dissolved tested substance, and t is the time point. Dissolution profiles are described as similar when the f₁ value is close to 0, or f₂ is close to 100 (between 50 and 100) [53]. The similarities and dissimilarities between profiles were marked in the figures with letters. If profiles share the same letter, they are similar.

The data from the dissolution studies were graphically correlated to mathematical models: zero-order, first-order, Higuchi’s model, and Korsmeyer–Peppas model in MS Excel (version 1808, Microsoft Corporation, Redmond, WA, USA) [54,55]. The mathematical equations of kinetic models are described below:

where F is the fraction of the released drug, k is the constant associated with the release, and t is the time (h).

The permeability of cannabinoids through biological membranes was measured using the Parallel Artificial Membrane Permeability Assay (PAMPA) model. The study was conducted in the gastrointestinal (GIT) and blood-brain barrier (BBB) models. The model consists of two 96-well microfilter plates, the donor and the acceptor plate. The wells were separated by a 120 μm thick microfilter disc coated with a 20% (w/v) dodecane solution of a lecithin mixture (Pion Inc., Billerica, MA, USA). The extract was diluted and the systems were dissolved (or suspended, centrifuged, and filtered) in dimethyl sulfoxide (DMSO) and placed in the donor solutions, which were adjusted to 6.8 for GIT application and pH 7.4 for BBB. The BBB permeability was only studied for extract, as Neusilin US2 does not leave the GIT. The plates were incubated at 310.15 K for 3 h for the GIT and BBB assay in a humidity-saturated atmosphere. After incubation, the plates were separated and the concentration of CBD and CBDA, as their concentration in the extract was the highest, was determined using the HPLC-DAD method. Each measurement was repeated six times. CBC was present in a quantifiable concentration only in the BBB study, thus, it was not determined in GIT conditions. The P_app was calculated using the following formulas:

where V_D is the donor volume, V_A is the acceptor volume, C_equilibrium is the equilibrium concentration (Cequilibrium=CD×VD+CA×VAVD+VA), S is the membrane area, and t is the incubation time (in seconds).

Substances with a P_app in the GIT model below 0.1 × 10⁻⁶ cm/s are considered to have poor permeability, compounds with 0.1 × 10⁻⁶ cm/s ≤ P_app < 1 × 10⁻⁶ cm/s are classified as mediocre permeable, and compounds found as well permeable have a P_app ≥ 1 × 10⁻⁶ cm/s [56]. Compounds whose P_app in the BBB model is <2.0 × 10⁻⁶ cm/s are known as poorly permeable. Compounds with questionable permeability have P_app values in the range of 2.0 to 4.0 × 10⁻⁶ cm/s. Substances that have a P_app value greater than 4.0 × 10⁻⁶ cm/s are regarded as highly permeable [57].

The extract and systems antioxidant activity was studied by four assays: DPPH, ABTS, CUPRAC, and FRAP. Two of them determine the ability to scavenge free radicals (DPPH and ABTS), whilst the other assays check the possibility of performing redox reactions (CUPRAC and FRAP). A linear regression equation between the trolox concentration and its scavenging percentage (DPPH and ABTS) or absorbance (CUPRAC and FRAP) was built. Thus, the results, presented as mg trolox/g plant material, were calculated through the equation according to the antioxidant properties of the extracts in all four assays [58,59]. Pure excipients showed no antioxidant potential under test conditions.

To perform the DPPH assay, a 96-well plate was used and the samples were measured spectrophotometrically [60]. The main reagent was a methanol solution of DPPH at a concentration of 0.2 mM. To initiate the assay, 25.0 µL of the system/trolox solution was mixed with 175.0 µL of the DPPH solution. The plate was then incubated in the dark at room temperature while shaking for 30 min. After the incubation period, the absorbances were obtained using a plate reader (Multiskan GO, Thermo Fisher Scientific, Waltham, MA, USA) at 517 nm. The absorbance (A) was also measured for a blank sample, which consisted of a mixture of DPPH solution and solvent at 517 nm. Each sample was tested for its absorbance at 517 nm. The inhibition of DPPH radicals by the studies’ systems/trolox was calculated using the equation:

where A_o is the absorbance of the control sample and A_i is the absorbance of the test sample. Each measurement was repeated six times.

As another assay to determine the scavenging radical potential, the ABTS study [61], was also performed. This study is based on the production of green cation radicals through the loss of electrons by nitrogen atoms of ABTS caused by potassium persulfate. During the assay, the green ABTS radical can be converted into a colorless neutral form in the presence of an antioxidant. In this assay, 200.0 μL of ABTS^•+ solution and 10.0 μL of the system/trolox solution were pipetted into 96-well plates and incubated for 10 min in the dark at room temperature while shaking [62]. After incubation, the absorbance values were measured at λ = 734 nm using a plate reader (Multiskan GO, Thermo Fisher Scientific, Waltham, MA, USA). The mixture of solvent and ABTS (control) and the wells filled with system and water (systems’ absorbance) at 734 nm were also studied. The inhibition of ABTS^•+ was calculated using the following equation:

where:

A₀—The absorbance of the control;

A₁—The absorbance of the sample.

To determine the reducing potential of the systems, the CUPRAC assay [63] was used. In this assay, the antioxidants’ phenolic groups undergo oxidation to form quinones, while the bluish neocuproine and copper (II) ion complex is reduced to the yellow neocuproine and copper (I) ion complex. To perform this study, a mixture of 50.0 µL of the system/trolox solution and 150.0 µL of the CUPRAC reagent was added to the plate and then incubated for 30 min at room temperature while shaking in the dark [62]. The control and systems’ own absorbance were also measured simultaneously. The absorbance was measured at a wavelength of 450 nm using a plate reader (Multiskan GO, Thermo Fisher Scientific, Waltham, MA, USA) after the 30 min incubation period. The analysis was performed using six replicates.

The FRAP technique was also used to determine the reducing properties of the systems, which involves reducing colorless Fe³⁺ ion to Fe²⁺ to form a dark blue complex with 2,4,6-tris(2-pyridyl)-1,3,5-triazine (TPTZ) [62]. In this method, 25.0 µL of the system/trolox solution and 175.0 µL of the FRAP mixture (consisting of 25 mL acetate buffer, 2.5 mL TPTZ solution, and 2.5 mL of FeCl₃·6H₂O solution) were applied to the plate and incubated in dark conditions at 37 °C for 30 min. The control and systems’ absorbance were also measured. Subsequently, the absorbance was measured at λ = 593 nm using a plate reader (Multiskan GO, Thermo Fisher Scientific, Waltham, MA, USA). The analysis was performed using six replicates.

The neuroprotective effect of cannabinoids was assessed against the possibility of inhibiting enzymes whose expression is associated with neurodegenerative changes.

As a standard inhibitor of esterases, galantamine was chosen, while for tyrosinase, azelaic acid was selected [64,65]. A linear regression equation that relates the standard concentration of a substance to its ability to inhibit an enzyme, as measured by the percentage of potential inhibition was created. An equation to calculate the standard equivalent for each extract based on its inhibitory properties in all three assays was obtained. The results were presented as a galantamine equivalent (GALAE) (mg galantamine/g plant material) for AChe and BChE assays and as an azelaic acid equivalent (AzAE) (mg azelaic acid/g plant material) [66,67,68,69,70,71].

The inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) was carried out using a colorimetric Ellman et al. modified assay [72]. This method requires artificial substrates (thiocholine esters). Thiocholine is liberated during the enzymatic reactions with 5,5′-dithio-bis-(2-nitrobenzoic) acid (DTNB), and the 3-carboxy-4-nitrothiolate anion (TNB anion) is formed. The potential to inhibit AChE and BChe was measured according to the increase in the thiocholine color in a 96-well plate. In total, 60.0 μL of 0.05 M Tris-HCl buffer (pH of 8.0), 10.0 μL of test solution, and 30.0 μL of AChE/BChE solution at a concentration of 0.2 U/mL were added to the wells. Subsequently, the plate was incubated for 5 min at 37 °C while shaking. Next, 30.0 μL acetylthiocholine iodide (ATCI)/butyrylthiocholine iodide (BTCI) at a concentration of 1.5 mM and 125.0 μL of 0.3 mM DTNB solution (5,5′-dithiobis-(2-nitrobenzoic acid)) were added to the wells. The plate was then incubated for another 20 min under the same conditions. A blank sample (the reaction mixture without the enzyme, with an increase in the volume of Tris-HCl buffer), a control sample (the solvent instead of the test sample), and a blank sample for the control sample (the reaction mixture without the enzyme, with an increase in the volume of Tris-HCl buffer) were also prepared. The measurements were performed at a wavelength of 405 nm. The analysis was performed using six replicates. The percentage of inhibition of AChE and BChE by the test samples was calculated using the following formula:

where:

A₁—The absorbance of the test sample;

A_1b—The absorbance of the blank of the test sample;

A₀—The absorbance of control;

A_0b—The absorbance of the blank of control.

The tyrosinase inhibition assay measures the activity of an inhibitor to prevent L-DOPA from accessing the tyrosinase active site. This leads to a decrease in the color intensity of the solution, which indicates enzyme inhibition [73]. To conduct the assay, 75.0 μL of 0.1 M phosphate buffer (pH 6.8) was added to each well of a 96-well plate, followed by 25.0 μL of the extract and 50.0 μL of enzyme solution (192 U/mL). The plate was shaken at room temperature for 10 min, after which 50 μL of 2.0 mM L-DOPA was added to each well and incubated for an additional 20 min under the same conditions. In addition to the test sample, a blank for the test sample (without enzyme, the volume of phosphate buffer was elevated), a control sample (with solvent instead of the test sample), and a blank sample for the control (without enzyme) were also prepared. The absorbance of the samples was measured at 475 nm. Each measurement was repeated six times. The percentage inhibition of the tyrosinase by the samples was calculated using an equation:

where:

A₁—The absorbance of the test sample;

A_1b—The absorbance of the blank of the test sample;

A₀—The absorbance of control;

A_0b—The absorbance of the blank of control.

Statistical analysis of results obtained in permeability assay, and in antioxidant activity study, was performed with the use of Statistica 13.3 software (StatSoft Poland, Krakow, Poland). Data are presented as mean values ± standard deviations. Experimental data were analyzed using the skewness and kurtosis tests to determine the normality of each distribution, and Levene’s test assessed the equality of variances. Statistical significance was determined using a one-way analysis of variance (ANOVA), followed by the Bonferroni post hoc test (to compare the experimental results acquired for cannabinoids in extract and in the systems). Differences were considered significant at p < 0.05.

Using extracts obtained from inflorescences with the scCO₂ extraction technique (the extraction yield was ~16.74%), cannabinoid delivery systems with increased solubility and permeability were obtained. As model carriers, biopolymer Soluplus and Neusilin US2 were applied. The systems of cannabinoids with carriers (Figure 2) were prepared using a solvent-evaporation method which enables the incorporation of a wide range of active ingredients into the resulting systems [74,75].

The extracts and systems have undergone the HPLC-DAD analysis to determine the cannabinoid content. In HiE, CBD was at the level of 6042.76 ± 82.19 μg/g plant material, CBDA at 2033.01 ± 67.98 μg/g plant material, whilst for CBC, 238.71 ± 11.20 μg/g plant material. The results of the systems analysis are presented in Table 1.

Two systems of the HiE with Neusilin US2, and Sol prepared in a 1:1 mass ratio (extract weight: carrier) using a solvent-evaporation technique, were enrolled in the dissolution study. The percentage cumulative cannabinoid release (% CBD, CBDA, and CBC) was measured at different time points (5, 10, 15, 30, 45, 60, 90, 120, and 180 min) for each system.

In 0.1 M hydrochloric acid, at pH 1.2 (Figure 3a), CBD was dissolved to the smallest extent compared to other media. After 60 min of the study, the percentage of dissolved CBD is only in HiE-Soluplus 4.08% ± 0.21%, in HiE-Neusilin US2 0.44% ± 0.09%, and even less in HiE. In the phosphate buffer at pH 6.8 (Figure 3b), the dissolution rate of CBD was overall greater for CBD in the co-dispersion delivery systems than in pH 1.2; however, CBD from HiE did not dissolve. After 60 min, the % CBD values were as follows: HiE-Soluplus at 39.83% ± 0.23% and HiE-Neusilin US2 at 33.21% ± 1.09%. CBD had the highest dissolution rate in HiE-Soluplus at pH 6.8 throughout the whole study.

The apparent solubility of CBD was also studied in FaSSIF and FeSSIF (Figure 4a,b). The dissolution profile of CBD was greater in FaSSIF and FeSSIF than in pharmacopeial media at pH 1.2 and 6.8. It is observed that in both advanced media, CBD was rapidly released from co-dispersion delivery systems. After 60 min, in FaSSIF, CBD was dissolved in HiE-Soluplus at 77.40% ± 1.15%, in HiE-Neusilin US2 at 75.47% ± 2.91%, and in HiE at 17.92% ± 1.79%. In FeSSIF (Figure 4b), CBD was released to the greatest extent, reaching after 60 min in HiE-Soluplus 99.25% ± 3.23%, in HiE-Neusilin US2 98.37% ± 1.82%, and in HiE 24.76% ± 2.48%. Co-dispersion delivery system HiE-Soluplus provided the best dissolution rate of CBD at each time point, which was statistically significantly different than CBD dissolution profiles in HiE-Neusilin US2 and HiE.

The dissolution of CBD is a complex process influenced by various factors. The dissolution kinetics of CBD were investigated using various mathematical models under different media conditions and in extract and co-dispersion delivery systems with Neusilin US2 and Soluplus (Table 2) [76,77,78,79]. Four mathematical models, namely zero-order kinetics, first-order kinetics, Higuchi kinetics, and Korsmeyer–Peppas kinetics were employed to analyze the dissolution data [80,81,82]. CBD in HiE displayed high R² values for zero-order and first-order kinetics, indicating a reliable and predictable release mechanism. The Higuchi model also showed notable correlations, suggesting diffusion-driven release. Moreover, the Korsmeyer–Peppas model displayed moderate to high correlations, and the n values indicated a Fickian diffusion. For both HiE-Soluplus and HiE-Neusilin US2, the Higuchi model consistently revealed diffusion-driven release mechanisms across pH conditions and biorelevant media. The Korsmeyer–Peppas model, which was also dominating for CBD in co-dispersion delivery systems, indicated the involvement of Fickian transport based on the n values (n < 0.5) [83].

The CBDA dissolution rate was also monitored under the same conditions. In hydrochloric acid, at pH 1.2, the overall results were the poorest (Figure 5a), it practically did not dissolve from HiE. After 60 min of the assay, CBDA was dissolved in HiE-Soluplus and HiE-Neusilin US2 at the level of 1.58% ± 0.21% and 1.88% ± 0.21%, respectively. The co-dispersion delivery system HiE-Soluplus provided the greatest dissolution rate of CBDA at pH 1.2. However, the overall results are poor and the profiles are statistically similar. The apparent solubility of CBDA was also studied in a phosphate buffer at pH 6.8 (Figure 5b). Similarly, to pH 1.2, CBDA in HiE practically did not dissolve during the study. After 60 min, HiE-Soluplus had a CBDA dissolution rate of 59.56% ± 0.23%, while HiE-Neusilin US2 was 49.64% ± 1.09%.

As for CBD, CBDA was also studied in FaSSIF (Figure 6a). The most noticeable differences are noted at the beginning of the study. After one hour of the assay, the dissolution percentages for CBDA in HiE-Soluplus, HiE-Neusilin US2, and HiE were 76.05% ± 2.91%, 60.20% ± 0.96%, and 23.25% ± 1.87% respectively. The results indicate that CBDA was dissolved to the greatest extent in HiE-Soluplus, which was statistically better than in HiE-Neusilin US2 and HiE. In the FeSSIF medium, the CBDA dissolution profile in HiE-Soluplus reaches the highest dissolution rate values and it differs significantly from the CBDA profile in HiE-Neusilin US2 and HiE (Figure 6b). The first time point, 5 min, shows the biggest variability in CBDA dissolution rate, where the percentage of CBDA released was 66.56% ± 1.09% for HiE-Soluplus, 30.13% ± 1.68% for HiE-Neusilin US2, and 11.52% ± 0.65% for HiE. The maximum dissolution rates are higher in FeSSIF than in FaSSIF.

The dissolution kinetics of CBDA was also studied (Table 3). CBDA consistently displayed the highest R² values in the Korsmeyer–Peppas and Higuchi models. The n values, fluctuating mostly from below 0.45 to three values below 0.89, suggest a potential dominance of Fickian diffusion. In three cases, the n values between 0.45 and 0.89 indicated the non-Fickian diffusion release mechanism which shows the relative complexity of the prepared co-dispersion delivery systems and may indicate that the CBDA release is controlled by more than one mechanism.

Following the methodology used for the apparent solubility study of CBD and CBDA, the dissolution profiles for CBC were determined in the same media and time points. In the study conducted at pH 1.2 (Figure 7a), the dissolution rate was similar for CBD and CBDA (the lowest). CBC did not dissolve in HiE during the study. At the last time point, 180 min, HiE-Soluplus showed the highest percentage of CBC released at 5.89%, whilst in HiE-Neusilin US2, CBC was dissolved in 5.18% ± 0.28%. CBC profiles were similar due to f₁ and f₂ factors. In the study where vessels were filled with phosphate buffer at pH 6.8 (Figure 7b), CBC in HiE was not dissolved, and the most noticeable differences were noted in the first minutes of the study. The CBC reached in HiE-Soluplus (120 min) of the study was 10.30% ± 1.36%. Whilst in HiE-Neusilin US2, it was 22.75% ± 1.00%. Both dissolution profiles of CBC in co-dispersion delivery systems were similar.

In fasted state intestinal conditions, the most dynamic changes, take place at 5 min of the study, where CBC is dissolved in HiE-Soluplus at 56.63% ± 2.80%, in HiE-Neusilin US2 at 12.28% ± 2.11%, and in HiE at 10.26% ± 0.81% (Figure 8a). After 30 min, CBC reached a plateau. The CBC profile in HiE-Soluplus is significantly better than in HiE-Neusilin US2 and HiE. The last environment in which the CBC dissolution rate was studied was FeSSIF (Figure 8b), where the greatest dissolution rate of CBC was obtained. After 15 min of the study, CBC was dissolved in 79.32% ± 2.30%, 58.68% ± 0.57%, and 26.25% ± 2.79% in HiE-Soluplus, HiE-Neusilin US2, and HiE, respectively. The CBC dissolution profile in HiE-Soluplus was significantly better than in HiE-Neusilin US2 and HiE.

The Higuchi and Korsmeyer–Peppas models consistently yield higher R² values compared to the zero-order and first-order models across different CBC formulations and pH conditions (Table 4). The release exponent (n) values are consistently below 0.5 across formulations and pH conditions, suggesting the release approximated the Fickian diffusion release mechanism indicative of controlled release predominantly driven by diffusion.

The results showed that HiE-Soluplus consistently provided the highest dissolution rate of cannabinoids compared to HiE-Neusilin US2 and HiE. The dissolution rate of CBD, CBDA, and CBC was highest in FeSSIF, followed by FaSSIF and the phosphate buffer at pH 6.8, while the lowest dissolution rate was observed in 0.1 M hydrochloric acid at pH 1.2. At pH 1.2, all three cannabinoids showed poor solubility. In a phosphate buffer with a pH of 6.8, the greatest improvement in solubility was observed for CBDA. CBDA, being the acidic precursor of CBD, might have some advantages in solubility compared to CBD and CBC. In FaSSIF, the maximum dissolution rate was similar for CBD, CBDA, and CBC. However, the fastest increase in dissolution rate was noted for CBC. In FeSSIF, the dissolution profiles for CBD, CBDA, and CBC were again similar, but the fastest increase in dissolution rate was observed for CBD. FaSSIF and FeSSIF provided a more similar composition to intestinal fluid than the pharmacopoeial media, containing surfactants that helped significantly increase the solubility of cannabinoids.

Increasing gastrointestinal permeability is important to obtain higher bioavailability as it allows for more efficient absorption into the bloodstream from the GI tract. Thus, a PAMPA study was performed.

The permeability coefficients of CBD in pH 6.8 were analyzed in HiE and co-dispersion delivery systems: HiE-Neusilin US2 and HiE-Soluplus (Table 5). The highest permeability coefficient was observed in HiE-Soluplus (3.09 × 10⁻⁷ ± 1.07 × 10⁻⁸ cm/s), followed by HiE-Neusilin US2 (2.73 × 10⁻⁷ ± 9.75 × 10⁻⁹ cm/s), and the CBD permeability was statistically the worst in the pure extract (1.86 × 10⁻⁷ ± 2.24 × 10⁻⁸ cm/s).

CBDA was better permeable than CBD through membranes in the study under the same conditions (Table 5). Its permeability coefficient reached 7.57 × 10⁻⁶ ± 1.21 × 10⁻⁷ cm/s in HiE, while the most noticeable and statistically significant increase was found in HiE-Soluplus, where CBDA reached 9.51 × 10⁻⁶ ± 4.66 × 10⁻⁸ cm/s and 7.56 × 10⁻⁶ ± 2.69 × 10⁻⁷ cm/s in HiE-Neusilin US2.

The BBB permeability was assessed for CBD, CBDA, and CBC in HiE. All P_app values were determined as higher than 4.0 × 10⁻⁶ cm/s, meaning that both cannabinoids cross the blood-brain barrier well.

HiE extract and the systems showed antioxidant activity (Table 6). In the DPPH model, the best result was obtained for Hi-Soluplus (0.97 ± 0.02 mg trolox/g plant material), while for HiE (0.85 ± 0.01 mg trolox/g plant material), however, these results are statistically similar. In the other scavenging radicals assay, ABTS, the HiE (17.04 ± 0.08 mg trolox/g plant material) antioxidant potential was improved the most by HiE-Soluplus (18.69 ± 0.17 mg trolox/g plant material). In the CUPRAC redox study, the greatest result was obtained for HiE-Soluplus (7.81 ± 0.22 mg trolox/g plant material). In FRAP, the most noticeable improvement in HiE antioxidant activity (11.58 ± 0.03 mg trolox/g plant material) was shown also for HiE-Soluplus (1.65 ± 0.03 mg trolox/g plant material). In general, the results show statistically significant improvement in ABTS and FRAP assays in antioxidant potential in the systems when compared to HiE, but the changes are subtle.

The inhibition of the enzymes connected to the development of neurodegeneration was also studied for the extract (Table 7). HiE inhibited an AChE of 20.78 ± 0.56 mg galantamine/g, while BChE was 17.49 ± 0.47 mg galantamine/g. Tyrosinase was also inhibited by the HiE 165.21 ± 7.11 mg azelaic acid/g. Preparation of the systems increased the inhibitory activity of the cannabinoids. The greatest enhancement was noted for HiE-Soluplus (AChE 21.06 ± 0.19 mg galantamine/g, BChE 17.54 ± 0.09 mg galantamine/g, and tyrosinase 171.30 ± 2.13 mg azelaic acid/g). The changes in biological activity are subtle, but there is a visible trend that the neuroprotective potential is increasing.

Data are available in a publicly accessible repository.