Development and Characterization of a High-CBD Cannabis Extract Nanoemulsion for Oral Mucosal Delivery

1. Introduction:

Cannabis-derived extracts rich in cannabidiol (CBD) possess substantial therapeutic potential for immune-mediated and oncologic oral conditions due to their anti-inflammatory, immunomodulatory, and pro-apoptotic properties [1,2,3,4,5,6]. CBD-rich cannabis extracts can modulate immune responses by various mechanisms. Several studies showed suppression of the activation of Cluster of Differentiation 4 helper T cells (CD4⁺) and Cluster of Differentiation 8 cytotoxic T cells (CD8⁺) T-cells, downregulation of pro-inflammatory cytokines such as Tumor necrosis factor alpha (TNF-α) and Interferon gamma (IFN-γ), and inhibition of the expression of cytotoxic effector molecules, including granzyme B, perforin, and Fas ligand [7,8,9]. Furthermore, CBD-rich cannabis extracts have been shown to promote selective apoptosis in malignant oral epithelial cells in head and neck squamous cell carcinoma (HNSCC) models [10,11,12,13,14,15]. Enrichment with the minor cannabinoid cannabichromene (CBC) further enhanced the cytotoxicity in HNSCC models through synergistic mechanisms [4].

Building on these findings, the biological effects of cannabis extracts are primarily mediated through the endocannabinoid system (ECS), a complex regulatory network that orchestrates critical physiological functions, including immune regulation, inflammation, nociception, and epithelial homeostasis [16,17,18,19,20]. The classical ECS comprises the cannabinoid receptor type 1 (CB1) and 2 (CB2), the endogenous ligands anandamide (AEA) and 2-arachidonoylglycerol (2-AG), and the enzymes responsible for their synthesis and degradation. The extended ECS further encompasses additional molecular targets, such as orphan G–protein–coupled receptors, transient receptor potential (TRP) channels, and peroxisome proliferator–activated receptors (PPARs), including PPARγ [21,22,23,24,25,26,27]. Through these pathways, CBD-rich extracts exert immunomodulatory, anti-inflammatory, and pro-apoptotic effects by modulating ECS receptors, reducing cytokine secretion, and suppressing cytotoxic mediators [7,8,12,13].

In our previous studies, we demonstrated that the CBD-rich cannabis extract (CAN296) and CBD:CBC formulations exert potent immunomodulatory and pro-apoptotic effects, showing therapeutic efficacy in T cell–mediated oral disorders such as oral lichen planus (OLP) and oral graft-versus-host disease (oGVHD), as in HNSCC models [4,28].

Building upon these therapeutic observations, it is essential to understand the clinical relevance of the target diseases. OLP and oGVHD are chronic, immune-mediated disorders that impair quality of life and carry a significant risk of malignant transformation [29,30]. OLP is a T-cell-driven inflammatory disease of the oral mucosa that presents as reticular, erosive, or plaque-like lesions [31], with a global prevalence of 0.5–4% [32]. It is classified as an oral potentially malignant disorder, with transformation to OSCC influenced by lesion type, duration, and patient-specific factors [33,34].

Chronic graft-versus-host disease (GVHD), a significant complication following allogeneic hematopoietic stem-cell transplantation (HSCT), occurs when donor T cells attack host tissues, resulting in persistent inflammation and tissue damage [35,36]. Oral manifestations often resemble OLP, including lichenoid lesions, mucosal atrophy, and ulcerations [37,38], and oGVHD similarly carries a malignant potential [39,40]. Both disorders share a common immunopathological basis characterized by T-cell dysregulation and cytotoxic epithelial injury. Activated CD4⁺ and CD8⁺ T cells secrete IFN-γ and TNF-α, while cytotoxic CD8⁺ T cells express Fas ligand (FasL) and release perforin and granzyme B at the epithelial basement membrane, inducing keratinocyte apoptosis [29,41,42,43,44]. Chronic immune activation, oxidative stress, and epithelial instability contribute to genomic damage and increased susceptibility to malignant transformation [45,46,47]. These mechanisms link OLP and oGVHD to HNSCC, the most common malignancy of the head and neck region [48]. HNSCC arises from the mucosal epithelium of the oral cavity, pharynx, or larynx and represents the sixth most common cancer globally, accounting for approximately 900,000 cases and over 400,000 deaths annually [49,50,51,52].

Current treatment modalities for OLP and oGVHD primarily rely on immunosuppressive agents. Topical corticosteroids (clobetasol, triamcinolone) are the first-line treatment for OLP and effectively reduce lesion severity and pain [53,54]. However, long-term use is associated with adverse effects, including mucosal atrophy, candidiasis, and reduced therapeutic efficacy due to tachyphylaxis [55,56]. Systemic corticosteroids, often used in severe cases of OLP and oGVHD, carry additional risks such as osteoporosis, diabetes, and increased susceptibility to infections [57,58]. Despite their efficacy, these agents do not address the chronicity or malignant potential of these disorders, underscoring the need for safer, targeted therapies capable of modulating immune dysregulation while minimizing toxicity. Calcineurin inhibitors, such as tacrolimus (TAC) and cyclosporine, offer an alternative mechanism of action by inhibiting T-cell activation through the suppression of interleukin-2 (IL-2) transcription [59,60]. While these agents have shown efficacy in managing OLP and oGVHD, their use is limited by systemic toxicity, including nephrotoxicity, hypertension, and neurotoxicity [61,62]. Moreover, the chronic and relapsing nature of both conditions often necessitates prolonged treatment, further exacerbating the risk of adverse effects [63,64]. Notably, conventional therapies fail to address the precancerous and malignant transformation potential of OLP and oGVHD.

Cannabinoids, particularly those found in CBD-type cannabis extracts, have shown potent antitumor effects in preclinical models of various cancers, including HNSCC (59, 60). Among over 500 bioactive compounds in Cannabis sativa, CBD and Δ⁹-tetrahydrocannabinol (THC) are the most studied components, while other cannabinoids and terpenes contribute to therapeutic activity via the “entourage effect” [65,66,67].

Cannabinoids exert anticancer effects through multiple mechanisms, including induction of apoptosis [68,69,70,71], inhibition of angiogenesis [72,73], suppression of tumor proliferation (67, 69), and modulation of the tumor microenvironment [74,75,76]. Notably, CBD-type extracts selectively target malignant cells while sparing healthy tissue [4,5], a property particularly relevant for oral diseases such as OLP and oGVHD, where chronic inflammation heightens carcinogenic risk.

This selectivity, together with the shared pathophysiology of OLP and oGVHD, including chronic inflammation, T-cell dysregulation, and malignant potential, supports the rationale for developing CBD-rich formulations. Accordingly, the present study focused on creating a stable, surfactant-based nanoemulsion system to enable localized mucosal delivery and sustained bioavailability of a CBD-rich extract.

For effective local treatment of oral lesions, formulation strategies must ensure mucosal compatibility, physical stability, and bioavailability. Due to the lipophilicity of cannabinoids in CBD-type extracts, aqueous and ethanol-based systems pose challenges for the delivery of these compounds. Nanoemulsion platforms can overcome these limitations by enhancing solubility and tissue penetration [77,78,79,80,81].

We hypothesized that incorporating the CBD-rich cannabis extract CAN296 into a Tween-dominant nanoemulsion system (≥80% Tween 80) would result in superior physicochemical performance compared with the crude extract oil, specifically by improving nanoscale droplet uniformity, wettability, and mucosal membrane interaction. These properties are expected to enhance local bioavailability and stability, thereby overcoming the limitations associated with the native, lipophilic extract. Accordingly, this study aimed to develop and characterize a stable, ethanol-compatible CBD-rich nanoemulsion optimized for oral mucosal application. The optimized formulation will serve as the basis for subsequent biological assays to evaluate its immunomodulatory and cytotoxic potential in oral inflammatory and precancerous conditions, including OLP, oGVHD, and early-stage HNSCC.

2. Results

2.1. Cannabis Emulsions with Up to 800 µg/mL Load Are Stabilized by 1% Surfactant Containing ≥80% Tween

Initial attempts using the crude, non-formulated cannabis extract (without surfactant) resulted in immediate phase separation, confirming that the native extract cannot form a stable aqueous dispersion. Therefore, subsequent experiments focused on determining whether a surfactant-based system could achieve and maintain nanoscale stability. To define the optimal formulation for long-term physical stability of cannabis emulsions, we systematically evaluated the influence of varying Tween:Span ratios on emulsion stability across a range of cannabis concentrations. Emulsions were prepared using a constant 1% surfactant system, with Tween 80 and Span 80 in different ratios (0:100 to 100:0), and cannabis extract was added at concentrations ranging from 0 to 4000 µg/mL. Each homogenized formulation was stored at room temperature and visually monitored for phase separation and clarity over an 8-week observation period (Figure 1).

Formulations without surfactants exhibited immediate phase separation at all cannabis concentrations, confirming that surface-active agents are essential for dispersion stability. When using mixed-surfactant systems with 35% or 65% Tween, the emulsions exhibited limited short-term stability, characterized by a turbid (milky) appearance. Specifically, formulations containing 400 µg/mL cannabis remained stable for up to 3 weeks, while those with 800 µg/mL remained stable for only 2 weeks before phase separation occurred. In contrast, emulsions containing ≥80% Tween (i.e., 80:20 and 100:0 Tween:Span) demonstrated prolonged physical stability throughout the 8-week observation period. At 400 µg/mL cannabis concentration, these formulations remained optically clear (i.e., transparent or translucent), while emulsions containing 800 µg/mL appeared turbid but homogeneous (milky), with no visible phase separation. None of the emulsions, regardless of the Tween:Span ratio, could stably incorporate cannabis concentrations of ≥ 1200 µg/mL; all such formulations exhibited visible oil phase separation during storage.

These results confirm that a 1% surfactant system requires at least 80% Tween 80 to achieve long-term stability for cannabis concentrations of up to 800 µg/mL. Beyond this threshold, the system's solubilization capacity is exceeded, indicating a formulation limit under the tested conditions. These findings guided our selection of the 80% Tween formulation for subsequent biological assays.

2.2. Dynamic Light Scattering Analysis Confirms Stability Limit of 800 µg/mL Cannabis in 1% Surfactant Emulsions with ≥80% Tween

To characterize emulsion stability in terms of droplet size distribution, we performed DLS analysis on representative formulations containing 1% surfactant (80% Tween, 20% Span) and varying cannabis extract concentrations (400, 800, 2000, and 4000 µg/mL). For comparison, two additional formulations lacking surfactant (0% Tween, 0% Span) were tested at cannabis concentrations of 2000 and 4000 µg/mL (Figure 2).

The formulation containing 400 µg/mL cannabis with 80% Tween exhibited a small and uniform particle size distribution, with a Z-average of 312.6 nm, a primary peak at 195.2 nm, and a moderate polydispersity index (PDI) (0.3647), indicating physical stability. The sample containing 800 µg/mL cannabis remained physically stable but showed larger, more heterogeneous particles (Z-average = 536.3 nm; primary peak = 279.9 nm; PDI = 0.4865). These two concentrations were classified as “Stable.”

At higher cannabis concentrations (2000 and 4000 µg/mL) with 80% Tween, the emulsions were unstable, with Z-averages exceeding 1000 nm and higher PDIs, including 0.8595 at 2000 µg/mL. These were marked as "Unstable". The control formulations without surfactant exhibited smaller average particle sizes (193.1 and 249.0 nm). These values reflect unstable multi-population distributions due to rapid phase separation and a lack of homogeneity; therefore, both samples were categorized as "Highly Unstable."

These findings confirm that surfactants are required to form uniform and stable cannabis emulsions, and that the maximum cannabis load that can be effectively stabilized with 1% Surfactant containing ≥80% Tween is approximately 800 µg/mL. This DLS-based analysis supports and extends the previously reported visual stability observations, providing particle size and polydispersity parameters that define the formulation limits under the tested conditions.

2.3. Transmission Electctron Microscopy (TEM) Reveals Progressive Particle Homogeneity with Increasing Tween Content

To visualize the morphological characteristics of the cannabis emulsions, representative samples containing 400 µg/mL cannabis extract and formulated with varying Tween:Span ratios were imaged using TEM. All emulsions were prepared with 1% total surfactant, except for the control sample, which contained no surfactant (0% Tween, 0% Span). Images were acquired at a fixed scale of 200 nm (Figure 4).

TEM micrographs revealed marked differences in droplet morphology and distribution depending on the surfactant composition. In the absence of surfactant (Panel A), significant, irregularly shaped oil droplets were observed, displaying high heterogeneity and frequent coalescence. Formulations containing low Tween content (35% Tween, Panel B; 65% Tween, Panel C) exhibited smaller droplets but retained notable morphological heterogeneity, with numerous irregular aggregates and variable droplet sizes.

In contrast, emulsions formulated with ≥80% Tween showed progressively improved particle uniformity. The 80% Tween sample (Panel D) produced more spherical, monodisperse droplets with reduced polydispersity. The 100% Tween formulation (Panel E) showed a dense field of uniformly distributed, nanoscale droplets with no visible aggregation. These findings are consistent with the DLS data, confirming that a higher Tween content promotes the formation of homogeneous nanoemulsions at a fixed cannabis concentration of 400 µg/mL.

2.4. Static Contact Angle (SCA) Measurements Reveal Nonlinear Wettability with Maximal Cohesion at 800 µg/mL

To be effective, nanoemulsions should exhibit increased wettability, which reflects improved spreading and adhesion on biological surfaces, thereby enhancing drug delivery and absorption. To assess surface wettability across different concentrations of CBD-rich cannabis, SCA measurements were performed on glass substrates using a drop shape analyzer (DSA). Droplets (5 µL) of formulations containing 0, 400, 800, and 1200 µg/mL cannabis extract were analyzed using the sessile-drop method (Figure 5).

The control formulation (0 µg/mL) exhibited a low SCA (~19°), indicating high spreading. At 400 µg/mL, the angle slightly increased (~23°), while 800 µg/mL yielded the highest value (~28°), reflecting reduced wettability but still within the range of good surface spreading. At 1200 µg/mL, the SCA returned to ~19°, indicating the greatest wettability among the concentrations tested. Statistical analysis confirmed that the 800 µg/mL formulation exhibited significantly higher contact angles compared to both 0 µg/mL and 1200 µg/mL (p < 0.01) (Panel A), indicating reduced wettability and increased droplet cohesion at this concentration. Together, these results demonstrate a non-linear, concentration-dependent relationship between extract content and surface spreading, with maximal droplet cohesion observed at an extract concentration of 800 µg/mL. A representative droplet image of the 800 µg/mL formulation is shown in Panel B.

2.5. Enhanced Stability of Nanoemulsions at 4 °C Compared to Room Temperature After 30 Days

To evaluate the impact of storage temperature on nanoemulsion stability, previously examined stable formulations containing cannabis extract (400 µg/mL and 800 µg/mL) were stored for 30 days at two temperatures. Samples were maintained at room temperature (~25 °C) and under refrigeration (4 °C), and droplet size distributions were analyzed by DLS (Figure 6). The largest detected droplet population (nm) was used as an indicator of aggregation or phase separation.

After 30 days, emulsions stored at 4 °C maintained droplet sizes close to their initial range (up to ~600 nm). In contrast, samples kept at room temperature exhibited a marked increase in size, with droplets expanding to over 4000 nm. Statistical analysis confirmed a significant interaction between storage temperature and cannabis concentration (two-way ANOVA, temperature × concentration; ****P < 0.0001). Pairwise comparisons indicated highly significant differences between storage conditions at 400 µg/mL (****P < 0.0001) and significant differences at 800 µg/mL (*P = 0.0259). Together, these results demonstrate that storage at 4 °C effectively minimized droplet coalescence and aggregation, preserving the nanoemulsion structure across both concentrations.

2.6. Significant In Vitro Retention of Cannabis Extract Nanoemulsion on Dialysis Membrane Suggests Mucoadhesive Potential

Surface adhesion and prolonged residence time can enhance local bioavailability for topical or mucosal drug delivery systems. To assess this property, the interaction of cannabis nanoemulsion with a dialysis membrane in an in vitro diffusion experiment was performed over 168 hours using a formulation containing 800 µg/mL cannabis extract (Figure 7). The extract-loaded nanoemulsion was placed in the donor compartment of the dialysis device, while the blank nanoemulsion was added to the receiver compartment. Eleven sealed capsules were incubated, and samples were destructively collected from both compartments at ten predefined time points and stored frozen. At the end of the experiment, ultra-high-performance liquid chromatography (UHPLC) was used to analyze all samples and quantify the CBD content. Extract concentrations were calculated by adjusting the CBD values based on the extract’s known CBD content (55%).

Over the course of the experiment, the donor compartment exhibited a gradual decrease in cannabis extract concentration, from 800 μg/mL at time zero to 116 μg/mL at 168 hours. In contrast, the receiver compartment showed only a modest cumulative increase, reaching 111.6 µg/mL by day 7. An increasing discrepancy between donor loss and receiver accumulation was observed throughout the time course, indicating that a significant portion of the extract was retained at the dialysis membrane. This membrane-retained fraction was calculated as the difference between donor depletion and receiver gain, and is shown as the purple bars in Figure 7.

2.7. Scanning Electron Microscopy (SEM) Visualization of Nanoemulsion Aggregates on Dialysis Membrane Surface

To visualize the localization of cannabis extract nanoemulsion on the dialysis membrane surface, an additional dialysis capsule was prepared under the same conditions described in Section 4.8, using formulations containing either 400 or 800 μg/mL of cannabis extract (Figure 8). After 24 hours of incubation, the capsule was disassembled, and the membrane was gently removed. It was then rinsed with distilled water, air-dried, and sputter-coated with a thin layer of iridium (2–3 nm) for SEM imaging.

SEM imaging qualitatively revealed deposition of nanoemulsion material on the membrane surface. At lower magnification (Figure 8A), scattered spherical droplets and small clusters were visible across the surface. At higher magnification (Figure 8B), visually denser regions of submicron aggregates were observed, forming irregular, closely packed structures. The intrinsic pore structure of the membrane was not resolved under the imaging conditions, likely due to dehydration and sputter coating, which may have obscured finer surface features. These observations are descriptive and based on qualitative visual assessment.

3. Discussion

Formulating a CBD-rich cannabis extract into a physically stable nanoemulsion required careful optimization of surfactant composition, extract concentration, and energy input to achieve nanoscale uniformity and mucosal compatibility. The optimized formulation, containing 1% total surfactant with ≥80% Tween 80 and ethanol as a cosolvent, demonstrated long-term physical stability with uniform droplet distribution, favorable wettability, and enhanced membrane interaction, supporting its suitability for oral mucosal delivery.

While our prior research demonstrated the anticancer effects of CBD-type cannabis extracts in HNSCC [4] and their immunomodulatory activity in OLP and oGVHD[28], the present work focused on overcoming the pharmaceutical challenge of formulating a physically stable Tween/Span-based nanoemulsion system incorporating ethanol as a cosolvent, suitable for oral mucosal application of CBD-rich cannabis extracts.

Cannabis-derived formulations are most commonly delivered via oral ingestion, inhalation, or sublingual sprays, each with notable drawbacks. Oral and edible products exhibit poor and variable bioavailability (6–20%) due to extensive first-pass metabolism and the high lipophilicity of cannabinoids [82,83,84]. Inhalation provides rapid systemic absorption but is associated with inconsistent dosing, short duration of effect, and potential respiratory irritation or airway inflammation [85,86]. Sublingual tinctures and sprays partially bypass hepatic metabolism, yet they often rely on ethanol-based vehicles that can cause mucosal irritation and provide limited residence time for local retention [82,87,88]. In contrast, the nanoemulsion platform developed in this study improves solubility, enhances mucosal wetting and adhesion, and enables localized, noninvasive delivery suitable for chronic inflammatory and precancerous oral conditions.

The nanoemulsion in this study was prepared using a MICCRA high-shear disperser, which applies controlled mechanical shear to achieve fine droplet dispersion without the need for ultrasonic or high-pressure homogenization. This moderate-energy method preserved the integrity of the cannabis extract, ensured mucosal biocompatibility, and maintained uniform droplet size and physical stability. In contrast, high-energy techniques such as ultrasonic cavitation and high-pressure homogenization have produced sub-200 nm translucent cannabis nanoemulsions with extended stability, but at the cost of elevated surfactant concentrations, which—although promoting smaller droplets and optical clarity—may be associated with reduced biocompatibility, potential mucosal irritation, and thermal or mechanical degradation of sensitive phytochemicals within the extract [89,90,91,92].

Initial attempts with the crude, non-formulated cannabis extract revealed that it could not be maintained as a stable aqueous dispersion, exhibiting immediate phase separation even under high shear. Consequently, comparative testing between the extract alone and the formulated nanoemulsion was not feasible. The study, therefore, focused on evaluating the physicochemical feasibility and performance of surfactant-based nanoemulsions as a prerequisite for subsequent biological testing.

Our findings demonstrate that emulsifying a cannabis extract with a Tween-dominant surfactant system enables long-term physical stability and the formation of nanoscale particles. A 1% total surfactant concentration containing at least 80% Tween 80 was required to stabilize cannabis emulsions at concentrations up to 800 µg/mL. At 400 µg/mL, the emulsions remained optically clear; however, at 800 µg/mL, they appeared milky but homogeneous, with no phase separation. Concentrations ≥1200 µg/mL exceeded the solubilization capacity of the system, resulting in visible phase separation and establishing a clear formulation limit under the tested conditions.

DLS and TEM corroborated these visual observations. Emulsions containing 400–800 µg/mL cannabis extract with 80% Tween displayed droplet sizes below 600 nm with moderate polydispersity indices, confirming the formation of a nanoscale, relatively uniform dispersion. In contrast, higher cannabis loads produced larger, heterogeneous populations and early aggregation behavior, as reflected in both DLS intensity/volume distributions and TEM images.

In this study, the formulation achieved nanoscale dispersions at extract concentrations of up to 800 µg/mL using low-energy homogenization and only 1% total surfactant, emphasizing ethanol compatibility, mucosal tolerability, and suitability for localized oral application. In contrast, a buccal CBD nanoemulsion system prepared with Tween 80/Labrasol exhibited rapid dissolution and stability in simulated saliva, thereby favoring systemic absorption over local mucosal retention [93].

SCA measurements provided further insight into the formulation’s interfacial behavior on solid surfaces. Both the 400 and 800 µg/mL formulations exhibited SCA within the optimal range for wetting the oral mucosa [94], suggesting favorable surface spreading and adhesion. Wettability followed a concentration-dependent pattern, with slightly increased droplet cohesion observed at 800 µg/mL, indicating altered interfacial dynamics at higher extract loads. These findings support the hypothesis that increasing the concentration of cannabis extract modulates droplet spreading and cohesion, factors that may enhance mucosal adherence and localized retention during buccal application.

Storage studies highlighted the effect of temperature on emulsion stability. After 30 days, emulsions stored at 4 °C maintained droplet sizes close to their initial range (up to ~600 nm). In contrast, samples kept at 25 °C exhibited marked coalescence (>4000 nm). Refrigerated storage mitigates coalescence and ripening in CBD nanoemulsions [95], consistent with colloid theory, which identifies Ostwald ripening as the principal coarsening mechanism in Tween/Span systems and outlines strategies to suppress it [96,97].

Comparable cannabinoid nanoemulsion systems demonstrated enhanced physicochemical stability under refrigerated storage, maintaining sub-200 nm droplet sizes and chemical integrity over extended periods. In contrast, room temperature conditions promoted droplet coalescence and cannabinoid degradation [95]. Refrigeration effectively minimizes kinetic instability and Ostwald ripening phenomena in surfactant-based cannabinoid formulations. This is consistent with stabilization effects also documented in lipid-based oral emulsions and intravenous hemp oil nanoemulsions. Together, these findings establish 4°C storage as a broadly applicable preservation strategy for cannabinoid delivery systems.

In vitro release testing using a dialysis membrane revealed a gradual depletion of the donor over 168 hours, with only modest accumulation in the receiver compartment. The discrepancy between donor loss and receiver gain, quantified as the membrane-retained fraction, indicates that a substantial portion of the extract remained associated with the dialysis membrane under the test conditions. This behavior suggests meaningful interaction between the nanoemulsion droplets and the membrane surface, consistent with mucoadhesive-like retention that could support localized delivery and sustained release on oral mucosal tissues.

Although differing in composition, mucoadhesive oral delivery systems have demonstrated that prolonged mucosal residence and controlled drug release enhance therapeutic outcomes in inflammatory oral diseases [98,99,100]. SEM analysis revealed membrane-associated retention and surface aggregation of the nanoemulsion, consistent with the therapeutic goal of maintaining localized drug presence while minimizing systemic exposure. SEM images showed scattered spherical droplets and densely packed submicron aggregates on the membrane surface, providing clear morphological evidence of deposition and supporting the formulation’s potential for sustained mucosal residence.

4. Materials and Methods

4.1. Phytocannabinoid Extraction and Sample Preparation

Air-dried Type III high-CBD cannabis CAN296 was obtained from a licensed Israeli medical cannabis distributor. Inflorescences were extracted in ethanol and decarboxylated by heating at 130 °C. The resulting extract was dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO, USA) and stored at –20 °C.

Phytocannabinoid profiling was performed using an UHPLC system coupled with a Q Exactive Focus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific, Bremen, Germany), as previously described [4]. Identification and absolute quantification of cannabinoids were achieved using analytical standards and external calibration curves (Thermo Scientific, Bremen, Germany), following previously established protocols [101].

4.2. Formulation Preparation

Cannabis nanoemulsions were prepared by first dissolving the CAN296 extract in 25% ethanol, which served as a cosolvent to aid in the dispersion of the lipophilic extract. The extract solution was then emulsified with 1% surface-active agent composed of Tween 80 and Span 80 (Sigma-Aldrich, St. Louis, MO, USA) at varying ratios. Cannabis concentrations ranged from 0 to 4000 µg/mL in 400 µg/mL increments. After 5 minutes of stirring, the mixtures were homogenized using a MICCRA D-9 disperser (MICCRA GmbH, Heitersheim, Germany) at 11,000 rpm for 3 minutes.

4.3. Stability Evaluation of Emulsions by Visual Inspection

Emulsion stability was evaluated through systematic visual inspection under consistent ambient lighting conditions. Samples were stored in clear, sealed vials at room temperature (~25°C) and assessed weekly over an 8-week period (days 7, 14, 21, 28, 35, and 42). Each formulation was observed for signs of turbidity, creaming, sedimentation, and phase separation. Emulsions were classified as Stable if they remained uniform (transparent or turbid) with no visible separation; Unstable if they exhibited droplet aggregation or partial separation with increasing turbidity or layering; and Highly Unstable if they underwent complete separation or sedimentation within the first week. These assessments served as a primary tool to evaluate formulation robustness across different surfactant ratios and cannabis concentrations.

4.4. Dynamic Light Scattering Analysis

Droplet size, PDI, and colloidal stability of cannabis nanoemulsions were assessed using DLS. Formulations contained 1% surfactant (80% Tween 80 / 20% Span 80) with 800, 2000, and 4000 µg/mL cannabis, along with surfactant-free controls. Measurements were performed at 25 °C on a Zetasizer ZS XPLORER (Malvern Panalytical, UK) in triplicate. Z-average, PDI, and intensity-weighted distributions were recorded and analyzed with ZS XPLORER software. Results were interpreted in conjunction with visual stability data to classify emulsions as stable, unstable, or highly unstable. It should be noted that DLS provides an intensity-weighted average that can overrepresent larger droplet populations in polydisperse systems; therefore, the measured Z-average and PDI values reflect relative rather than absolute particle sizes.

4.5. TEM Imaging

Morphological analysis of nanoemulsion particles was performed using TEM. Samples were adsorbed onto Formvar-carbon-coated copper grids (EMS, 200 mesh) and negatively stained with 2% uranyl acetate to enhance image contrast. After staining, the grids were allowed to air-dry before imaging. TEM was performed using a JEOL JEM-1400 Plus transmission electron microscope (JEOL Ltd., Tokyo, Japan) operated at 100 kV. High-resolution images were acquired using a Gatan Orius 600 CCD camera, allowing for detailed visualization of droplet morphology, particle size, and aggregation patterns across various surfactant formulations.

4.6. SCA Measurement

Wettability of the CBD-rich cannabis nanoemulsion was evaluated using a KRÜSS Drop Shape Analyzer (DSA; KRÜSS GmbH, Hamburg, Germany) equipped with ADVANCE software (v1.11.2.25901). Glass microscope slides (Thermo Scientific, Germany) were cleaned with ethanol, rinsed with deionized water, and air-dried before use. A 5 µL droplet of each formulation (0, 400, 800, and 1200 µg/mL) was dispensed with a microsyringe, and SCA values were measured immediately using the sessile drop method. Left- and right-angle measurements were automatically averaged, with six replicates per sample under ambient laboratory conditions (~25 °C).

4.7. Phytocannabinoid Identification and Quantification

Phytocannabinoid analyses were performed using a Thermo Scientific UHPLC system coupled with a Q ExactiveTM Focus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific, Bremen, Germany). Identification and absolute quantification of phytocannabinoids were performed using analytical phytocannabinoid standards and Cannabis samples at preset concentrations, as previously described [101].

4.8. In Vitro Release Kinetics Using a Dialysis Membrane System

The release of CBD from the nanoemulsion (800 µg/mL) was evaluated using a dialysis kit (Pur-A-Lyzer Mini 6000, 1 kDa molecular weight cutoff; Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) with a 1 kDa pore membrane. The cannabis-loaded nanoemulsion was placed inside the membrane chamber, and a blank nanoemulsion was added to the outer compartment. Twelve sealed vials were incubated at room temperature under constant agitation. At ten predefined time points (0, 0.5, 1, 2, 4, 8, 24, 48, 72 hours, and 7 days), samples were collected separately from the internal (donor) and external (receiver) compartments. UHPLC was used to analyze all samples for CBD concentration (Section 4.7), and release values were corrected based on the extract’s CBD content (~50%).

4.9. SEM Imaging

Samples were examined using an Extra-High-Resolution Scanning Electron Microscope (Magellan 400 L, FEI Company, Hillsboro, OR, USA) after plasma sputter-coating with a thin iridium layer (2–3 nm) to enhance surface conductivity. Each membrane sample was imaged in three representative regions at varying magnifications to assess surface morphology and potential pore obstruction. SEM images were processed using ImageJ software (NIH, USA) for qualitative visualization. Uniform brightness and contrast adjustments were applied to all images to enhance the visibility of surface features. Thresholding was used to improve the contrast between droplets and the membrane background, without compromising image integrity.

4.10. Statistical Analysis

All statistical analyses were conducted using GraphPad Prism version 10.6.1 for macOS (GraphPad Software, San Diego, CA, USA). Inferential statistical analyses were performed where applicable to assess differences between formulations and conditions. SCA data, storage stability measurements (droplet size), and in vitro release profiles were analyzed using one-way or two-way analysis of variance (ANOVA) with appropriate post hoc tests (Tukey’s or Šidák’s multiple comparisons), as specified. Data are presented as mean ± SD of at least three independent measurements unless otherwise noted. Visual inspection, DLS measurements, and TEM imaging were interpreted descriptively to evaluate qualitative trends in emulsion clarity, temporal stability, droplet size distribution, and polydispersity.

5. Conclusion

This research established a Tween-dominant nanoemulsion capable of stabilizing a robust concentration of CBD-rich cannabis extract. This optimized system remains stable under refrigeration, exhibits favorable wettability and membrane retention, and provides a physically stable, ethanol-compatible platform for oral mucosal delivery of cannabis extract. Future research should evaluate mucosal biocompatibility and in vivo retention to advance translational development for OLP, oGVHD, and early-stage HNSCC.