All solvents used for the GC and HPLC analysis were of analytical grade. Acetonitrile (ACN) and formic acid (FA) together with terpene standards [(+)-3-carene, (+)-limonene, α-bergamotene, α-pinene, α-terpineol, ß-myrcene, camphene, β-caryophyllene, caryophyllene oxide, farnesene mix, fenchol, humulene, linalool, terpinolene] and n-alkane (C₈-C₃₀) standard solutions were purchased from Sigma-Aldrich (Prague, Czechia). Methanol (MeOH), n-hexane, dimethyl sulfoxide (DMSO), sodium hydroxide and ethanol (EtOH) were obtained from VWR Chemicals (Prague, Czechia). Standards of cannabinoids, namely, cannabidivarin (CBDV), cannabidivarinic acid (CBDVA), cannabigerol (CBG), cannabigerolic acid (CBGA), cannabinol (CBN), cannabinolic acid (CBNA), cannabidiol (CBD), cannabidiolic acid (CBDA), cannabichromene (CBC), tetrahydrocannabivarin (THCV), 19-tetrahydrocannabinol (THC), and tetrahydrocannabinolic acid A (THCA-A) were purchased from Cayman Chemicals (Ann Arbor, United States). Chitosan (practical grade), sodium tripolyphosphate (TPP) and acetic acid were purchased from Sigma-Aldrich. Microbiological growth media Mueller-Hinton Broth (MHB), Sabouraud dextrose agar (SDA) and Sabouraud dextrose broth (SDB) were bought from OXOID (Prague, Czechia), antibiotics clotrimazole (CLT), chloramphenicol (CLP), ampicillin (AMP), terbinafine (TB) were obtained from Sigma-Aldrich. Human spontaneously immortalized keratinocyte cell line (HaCaT) was obtained from Cell Lines Service GmbH (Eppelheim, Germany), Dublecco’s modified eagle medium (DMEM), Dublecco’s phosphate buffered saline (DPBS), penicillin:streptomycin and enzyme accutase were purchased from Biowest (Nuaillé, France). Fetal bovine serum (FBS) was bought from Biosera (Prague, Czechia), and 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl-2H-tetrazolium bromid (MTT) was obtained from Sigma-Aldrich.

Cannabis ethanolic extracts were prepared from three cannabis strains differing in cannabinoid profile. Namely Chocolope (Chl) (DNA Genetics, NL) with high THC(A) content, cultivated under controlled indoor conditions at the Department of Food Science, Faculty of Agrobiology Food and Natural Resources, Czech University of Life Sciences Prague in 2020. High CBD(A) strain with the working name “Jonas 1, J1” was obtained from Phyto Hemp s.r.o. (Czech Republic) as well as cultivar with high CBG(A) content (working name “Hemp G, HG”). These two genotypes were grown outdoor.

The extracts were prepared by maceration of 60 g of dried homogenized cannabis inflorescences for 48 h in 80% ethanol in the ratio 6:1 (solvent: flower; v/w). Subsequently, the extract was filtered, and the solvent evaporated using a Rotavapor® R-100 vacuum evaporator (Buchi, CHE) at 40 °C. Prepared extracts were stored at −20 °C. The cannabinoid profiles of prepared extracts were determined by HPLC/DAD (Thermo Fisher Scientific, USA). The terpene profile was determined by GC/MS (Agilent, USA). Both methods were described in detail in a previous study (37).

Chitosan nanoparticles were prepared by the ionotropic gelation (38) with slight modifications. Chitosan was dissolved at 0.5% w/v in 1% acetic acid (v/v). After 10 min of ultrasonication, the solution was stirred for 3 h until complete dissolution. Subsequently, pH was raised to 5.0 by 10 N sodium hydroxide and the solution was stirred for an additional 30 min. In the next step 0.25% TPP in distilled water was added to reach the required mass ratio of CS: TPP 3:1 (w/w), and the solution was stirred for another 30 min. Chitosan nanoparticles were formed immediately after the addition of TPP. The solution containing nanoparticles was centrifuged at 9000 g for 45 min at 4 °C. The obtained pellets were extensively rinsed with distilled water to remove all the residues of sodium hydroxide. Finally, the pellets were freeze-dried (Gregor Instruments, Czech Republic), weighed, ground and stored at 4 °C for further use.

To prepare chitosan nanoparticles with three different cannabis extracts, 2 mL of cannabis extracts (c = 50 mg/mL) in 80% ethanol were added to 200 or 400 mL of CS solution with 1% of acetic acid after 10 min of ultrasonication and adjusted pH to 5.0. The extract: CS ratios were selected based on physicochemical properties of crude extracts and it corresponds to 1 g and 2 g of CS, or 1:10 and 1:20 w/w extract: CS ratio, respectively. After that the same procedure as described above was followed. The supernatants from centrifugation were kept for subsequent indirect determination of the encapsulation efficacy. In total, seven types of nanoparticles were prepared (Table 1).

Particle size distribution, zeta potential and polydispersion index (PDI) of prepared empty and loaded nanoparticles were measured by DLS (Dynamic Light Scattering) with a Zetasizer Nano SZ instrument (Malvern Instruments, UK). The analysis was performed at a scattering angle of 173°, at a temperature of 25 °C, using 2 g/L solution of reconstituted nanoparticles (previously dried and ground to a fine powder) in deionized distilled water, at pH 5. The morphology of the nanoparticles was observed using a field emission scanning electron microscope (FE-SEM) Zeiss Ultra 55 (Zeiss, Germany) at 5 kV and 45 mA. Before observation, the samples were coated with platinum. Fourier transform infrared spectroscopy (FTIR) spectra were taken on Bruker Tensor 27 spectrometer (Bruker, UK). Finely, ground samples were placed on the ATR crystal, and 32 consecutive scans were performed on each sample. Spectra were taken from 4,000 to 400 cm⁻¹ and the resolution of the wavenumber was 2 cm⁻¹. Encapsulation efficiency (%EE) was determined indirectly from the supernatant obtained during the centrifugation. Supernatant was filtered through a 0.1 μm syringe filter (Milipore, USA), evaporated under the stream of nitrogen, diluted in MeOH, analyzed by HPLC/DAD and the EE was calculated according to Equation 1:

In vitro antimicrobial and antifungal activity was determined by broth microdilution methods according to the Clinical and Laboratory Standards Institute (CLSI) M07-A8 and M38-A2 for bacteria and fungi, respectively (39, 40), with slight modifications. The results were in both cases expressed as minimum inhibitory concentration (MIC₈₀), i.e., the lowest concentration that inhibited bacterial or fungal growth by 80% compared to the untreated control. All experiments were carried out in 3 technical and 3 independent replicates. Antimicrobial activity of the empty and loaded nanoparticles was tested against 7 pathogenic bacterial strains, namely Staphylococcus aureus ATCC 25923 and 29213, S. epidermidis CCM 50 and 4418, S. lugdunensis CCM 4069, S. saprophyticus CCM 2727 and Streptococcus pyogenes CCM 4425, While their antifungal activity was determined against 12 dermatophytes, namely Arthroderma insingulare (CCF 5417; 5943), Epidermophyton floccosum CCM 8339, Microsporum gypseum CCM 8342, 3 strains of Nannizzia fulva (CCF 6025; 5338; 5782), Nannizzia gypsea CCF 5215, two strains of Trichophyton rubrum (CCF 4934; 4879), Trichophyton interdigitale CCM 8377 and Trichophyton tonsurans CCF 4930. The strains were purchased from the American Type Culture Collection (ATCC), Czech Collection of Microorganisms (CCM), or kindly provided by the Culture Collection of Fungi, Department of Botany, Charles University, Prague (CCF).

Human keratinocytes cell line HaCaT (CLS, Germany) was maintained at 37 °C in a controlled atmosphere with 5% CO₂ and 95% humidity. Cells were cultivated in DMEM medium supplemented with penicillin (100 U/mL) and streptomycin (100 μg/mL) in T75 cm² flasks with surface treatment. HaCaT cells were sub-cultured by detachment using enzyme accutase at 80–90% confluency every 3rd or 4th day and fresh medium was replenished every 2–3 days.

The cytotoxic effect of crude and seven encapsulated extracts was determined by the MTT cell viability assay (41). Cells in the exponential growth phase were detached at 80–90% confluency and viable cells were counted using trypan blue solution in a Neubauer counting chamber under the microscope (Motic AE 2000, Spain). Cells were seeded in 96-well surface treated plates at 1.25 × 10⁵ cells/mL in concentration 200 μL/well and left in an incubator for 24 h. After that, the cells were treated with serially diluted particles/extracts/ATB in concentrations ranging from 4 to 1,024 μg/mL and incubated for 24, 48 and 72 h. After the required incubation period, the medium in all experimental groups, including controls, was aspirated and cells were washed twice with 200 μL of DPBS. Subsequently, 200 μL of serum free DMEM containing 500 μg/mL of MTT was added to the experimental and control wells. The plates were incubated for an additional 2 h in the dark. Subsequently, medium was aspirated and formed formazan crystals were dissolved in 200 μL of DMSO, and the absorbance was then measured using BioTek Synergy H1 microplate reader at 540 nm. The results were expressed as a 50% inhibition of viability (IC₅₀) compared to the untreated control. Untreated cells were used as a control group and pure DMSO as blank control. All experimental groups and concentrations were tested in pentaplicates, whereas control group was tested in eight replicates. The degree of cytotoxicity was evaluated according following criteria: IC₅₀ ≤ 20 μg/mL = high cytotoxic effect, IC₅₀ between 21 and 200 μg/mL = moderately cytotoxic, IC₅₀ ranged from 201 to 500 μg/mL = weakly cytotoxic and IC₅₀ ≥ 501 μg/mL = no cytotoxicity (42).

The results were expressed as mean values and standard deviation (SD) in Excel and STATISTICA 12 software (StatSoft, Tulsa, United States). One-way analysis of variance (ANOVA) followed by Tukey’s HSD Test with p = 0.05 was performed.

The total content of cannabinoids ranged from 174.43 mg/g for HG to 346.04 mg/g for Chl, respectively (Table 2). The cannabinoid profile varied according to the strain, THCA dominated in Chl, CBDA in J1 and CBGA in the Hemp G strain. In the case of J1, CBGA was also present in significant amounts. The remaining cannabinoids were present only in minor or trace amounts. This corresponds to the natural dominance of major cannabinoids described above (28). The terpene profile of the extracts was very poor, except for Chl strain. β- and γ- eudesmol were the most dominant ones in Hemp G and Jonas 1 strains. The major terpens in Chl strain were β-caryophyllene and β-eudesmol (Table 3).

The prepared particles had a form of a white to green fine powder with a mild odor of cannabis. The mean size of the chitosan nanoparticles with a slightly broad size distribution ranged from 89.1 nm for empty particles to 355.6 nm for CNG2 particles (Figure 1A). The size of particles increased with the addition of the extract. However, this effect was observed primarily in particles with a lower extract content, which may be attributed to the interaction between polymers and extract composition. These findings are consistent with a previous report (35). The range of PDI was below 0.4 (Figure 1A), indicating uniformity of the prepared particles (43). In some batches of nanoparticles, several formations of aggregates >400 nm have been observed, probably due to presence of impurities (dust) or due to the disruption of the equilibrium between the hydrogen bond attractions and the electrostatic repulsions between chitosan nanoparticles. These aggregates may exhibit distinct physicochemical properties that could affect the biological activity of the encapsulated extracts. However, the formation of these large microparticles can be restricted by increasing the CS: TPP ratio, or a higher volume of the reaction medium, i.e., water (44).

The infrared analysis of CS, and NPs was performed to characterize the chemical structure of nanoparticles (Figure 1B). FTIR spectra of CS exhibited two vibrations at 1655 cm⁻¹ and 1,573 cm⁻¹, which were attributed to the CONH₂ and NH₂ groups, respectively. The intensity of those functional groups decreases, and two new peaks appear at 1637 cm⁻¹ (b) and 1,543 cm⁻¹ (c) after the addition of TPP. That proves the crosslinking of ammonium groups with TPP, which corresponds with similar results in a study by Lifeng et al. (38). The broader region starting at 3292 cm⁻¹ (a) in CNP and loaded NPs compared to CS could be correlated with enhanced hydrogen bonding. Furthermore, the appearance of a peak at 1250 cm⁻¹ (d), due to P = O stretching, and a peak at 889 cm⁻¹ (e), due to P-O bending in chitosan nanoparticles, which is not present in native chitosan, clearly demonstrates the crosslinking of TPP resulting in nanoparticles (34).

The zeta potential of nanoparticles ranged from 25 to 41 mV (Figure 1C). The highest value was determined for unloaded NP (41 mV) and was similar to values measured for NP with lower extract content (~39 mV). The zeta potential value above ± 30 mV indicates greater stability and reduced aggregation, due to the higher repulsion between the particles (45). Therefore, those particles will be suitable for the preparation of stable suspensions. Moreover, the zeta potential was positive for all prepared samples due to positively charged amine groups in chitosan indicating good interaction with the bacterial cell membrane, which usually has a negative charge (38). The particle size and was also confirmed by FE-SEM images (Figure 1D).

The total encapsulation efficiency ranged between 44.65 ± 4.39% to 94.44 ± 0.93%. Only the major cannabinoids in both acidic and decarboxylated forms were detectable in the supernatant (Table 4). The absolute mass of cannabis extract encapsulated in nanoparticles, related to the amount of CS and TPP (3:1 w/w), as well as the extract:chitosan ratio, ranged from 29.55 ± 0.45 mg/g for CNG2 to 56.09 ± 1.88 for CNT1. Nanoparticles with an extract:chitosan ratio of 1:10 (w/w) exhibited a lower EE compared to nanoparticles with a ratio of 1:20 (w/w). On the other hand, the absolute amounts of encapsulated extracts were higher for all encapsulated samples at a 1:10 ratio than for nanoparticles at 1:20 ratio, as is shown in Table 4. Cannabigerolic acid was the only cannabinoid identified in all supernatants simultaneously showing relatively high EE. On the other hand, for the remaining trace cannabinoids, a strong encapsulation rate can also be expected. CBG together with CBDA showed a relatively lower EE compared to CBGA and THC(A). Interestingly the supernatant from CBDA dominant extract (Jonas 1) contained a higher concentration of CBD compared to crude extract. This could be explained by the decarboxylation of CBDA to CBD due to the higher temperature and the lowered pH during preparation of nanoparticles (46).

The crude extracts exhibited high activity against all tested bacteria (Table 5). The most susceptible bacterium was S. pyogenes with MIC₈₀ 4 μg/mL for all extracts. All extracts inhibited other bacteria by concentrations ranging from 8–16 μg/mL. The MICs are consistent with previous research (37). However, all ethanol extracts were slightly less effective against S. aureus compared to isolated THC, CBD and CBG (MIC 0.5–2 μg/mL) (21, 24). Although cannabinoids represent the extract’s major components, other biologically active substances, especially terpenoids, should not be overlooked. The antimicrobial activity of the identified terpenes in extracts has been proven by many studies (47). β-myrcene showed high activity against S. aureus and S. epidermidis (26), while β-caryophyllene has a proven effect against S. lugdunensis and S. saprophyticus (48). The unloaded nanoparticles and nanoparticles with encapsulated extracts also showed antimicrobial activity at 32–512 μg/mL concentrations. The bacteria were less susceptible to empty CNP and NP loaded with Hemp G extract (MIC₈₀ = 256–512 μg/mL). These results confirm that blank chitosan nanoparticles exhibited antimicrobial activity against the tested bacteria. Moreover, in the case of CNG1 and CNG2, the antimicrobial effect was equal to or even lower than that of the blank CNP. These data demonstrate that the encapsulated extract with a high CBGA content was ineffective against the tested bacteria after encapsulation (with the exception of S. pyogenes), and the primary antimicrobial effect can be attributed to the chitosan nanoparticles themselves. These findings are consistent with previous studies (34, 49). The strongest antimicrobial activity was demonstrated by CNT1, where MIC₈₀ ranged from 32 to 64 μg/mL. Regarding the EE, it is important to highlight that the extracts after encapsulation represented 2.96 ± 0.05–5.61 ± 0.19% of total mass of NPs. Therefore, the extract encapsulation in chitosan slightly increases its antimicrobial activity. Among the control antibiotics used, both controls showed strong antibacterial effects. AMP had stronger antimicrobial activity (0.0625–2 μg/mL) than CLP (2–8 μg/mL), however, the difference between CLP and cannabis extracts was only a few dilutions and in the case of S. pyogenes the extracts were comparable to CLP.

All strains of dermatophytes were less susceptible to crude or encapsulated extracts compared to bacteria. The MIC₈₀ values ranged from 64 to 512 μg/mL, with the modus ranging from 128 to 256 μg/mL. The Chl extract was the most potent, followed by J1 and HG, respectively. The most sensitive dermatophytes were Nannizzia fulva CCF 5338 and Trichophyton interdigitale CCM 8337, which belong to the most common cause of dermatomycoses together with T. rubrum (50). The most resistant strains were Arthroderma insingulare (CCF 5417 and 5,943) and T. rubrum CCF 4934. So far, very limited evidence of cannabis activity against dermatophytes exists. Turner and Elsohly (1981) reported potent effects of CBC and its analogues against Trichophyton mentagrophytes (MIC = 6.25–50 μg/mL) (51). In another study, cannabis extracts obtained from THC and CBD-rich cannabis plants using ethanol as a solvent demonstrated antifungal activity (MIC₅₀) against T. mentagrophytes ranging from 89.37 to 240 μg/mL (52). The pure CNP showed only low or no activity against dermatophytes in tested concentrations (MIC₈₀ was in most cases >1,024 μg/mL), even though CNP bactericidal effect is otherwise quite strong. This low antifungal activity of empty chitosan nanoparticles was previously reported (53, 54). The only exception was E. floccosum which was significantly inhibited by all NPs in the range of 256–512 μg/mL, while the widely used antibiotic TB was not effective in tested concentrations. However, these findings confirm the partial antifungal and antibacterial activity of pure chitosan nanoparticles against certain pathogenic bacteria and dermatophytes.

Similarly to antibacterial activity, CNT1 was the most potent against all dermatophytes (MIC₈₀ = 256–1,024 μg/mL). The antifungal activity of CNB1 was also confirmed and in the case of Nannizzia fulva CCF 6025 the activity of CNB1 (MIC₈₀ = 256 μg/mL), compared to the crude extract (MIC₈₀ = 256 μg/mL) was the same. Based on the encapsulation efficiency (Table 4) and the amount of encapsulated extract, that ranged between 2.96 ± 0.05 to 5.61 ± 0.19% of NPs, we can confirm that the incorporation of cannabis extracts to chitosan nanoparticles improved the antibacterial and especially antifungal activity by reducing the effective concentration of extracts, compared to the crude form, by more than 20 times in some cases.

As shown in Table 6, all crude extracts demonstrated moderate cytotoxicity to keratinocytes which increased with time. In general, the extracts were less cytotoxic than isolated compounds, i.e., previously reported IC₅₀ values for CBD were 1.83 μg/mL (55), CBG 3.7 μg/mL and CBGA 7 μg/mL (56). In a study involving several cell lines, the cytotoxicity of high-THC ethanolic extracts ranged from 8 to 49 μg/mL after 24 h exposure (57). In a paper investigating the effect of THC on skin protection against UV, concentrations of 1.25 μg/mL were found to be non-cytotoxic. In addition, the authors confirmed slightly increased cell viability (58). This is consistent with our observation as is shown in Figure 2A Pure Chl extract increased viability of keratinocytes by 15.64 ± 6.82% at a concentration of 8 μg/mL compared to control and even by 30.23 ± 3.39% at a concentration of 4 μg/mL, respectively.

Although nanoparticles were formulated with extracts varying in the spectrum of cannabinoids, the overall effect of NPs on keratinocytes was considered as non-cytotoxic as is shown in Table 6. In a paper reported by Ridolfi et al. (2012), a concentration of 500 μg/mL CNP did not affect the viability of keratinocytes after 24 h period, which is in the agreement with our results (59). After exposure of 24 h, the highest cytotoxic effect was observed for CNB1, where IC₅₀ was 653.942 ± 71.98 μg/mL, followed by CNG1 (734.81 ± 73.15 μg/mL) and CNT1 (774.81 ± 73.29 μg/mL). Regarding to EE, there was even a slight increase of cytotoxicity effect in some cases compared to pure extracts after 24 h and 48 h exposure period. However, the IC₅₀ of nanoparticles decreased with the exposure period as opposed to pure extracts, particularly after 72 h. In addition, regarding EE, the encapsulation reduced the cytotoxic effect of the crude extracts by up to 10 times after 72 h. Moreover, a similar effect on cell viability was observed for the chitosan nanoparticles with high-THC. Enhanced cell vitality by 19.88 ± 7.25% was observed for CNT1 after 72 h of exposure compared to control at a concentration of 256 μg/mL and 12.36 ± 11.4% at a concentration of 128 μg/mL, respectively, contributing to the evidence of a positive effect of THC on skin cells as was described above (Figure 2B). Unlike antibiotics, the effect of cannabis extracts is not limited to antimicrobial activity, since it may have several added benefits. For example, cannabinoids interact with CB1R and CB2R receptors in keratinocytes, which reduce the production of inflammatory factors (TNF-α, IL-1, or IL-6) that play an important role in wound healing (60). Furthermore, broad-spectrum extracts may contain other biologically active compounds that can contribute to the advanced effects of encapsulated extracts. These compounds may include terpenes that allow access of lipophilic cannabinoids to deeper layers of the skin or may enhance their activity. This “so called” entourage effect has been demonstrated for some terpenes identified in cannabis (61).

Although crude extracts exhibited moderate cellular cytotoxicity, in vitro models cannot completely reproduce human skin and its complex functions. Moreover, topical applications of cannabinoids are now commonly used in medical therapy and have negligible side effects (62). A review by Martins et al. (2022) shows that topical application of cannabinoids is very well tolerated across patients with a range of skin disorders, whereas the treatment period is usually several weeks (63). This again highlights the low cytotoxicity of the extracts in vivo.

Chitosan nanoparticles represent a very simple, inexpensive and effective way to apply drugs to the skin, improving their bioavailability by prolonging the dwelling time of topically applied drugs. Additionally, it could enhance the passage of the agent through epithelial cells by opening tight junctions between epithelial cells, while reducing their side effects (64). Moreover, encapsulation of extracts not only improves their application but also reduces their cytotoxic effect on keratinocytes.