All solvents used for the HPLC and GC analysis were of analytical grade. Acetonitrile (ACN) and formic acid (FA) together with terpene standards [(+)-3-carene, camphene, α-pinene, ß-myrcene, (+)-limonene, terpinolene, linalool, fenchol, α-terpineol, caryophyllene, α-bergamotene, humulene, farnesene mix, caryophyllene oxide] and n-alkane standard were purchased from Sigma-Aldrich (Prague, Czechia). Methanol, n-hexane, dimethyl sulfoxide (DMSO), and ethanol (EtOH) were purchased 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), Δ⁹-tetrahydrocannabinol (THC), and tetrahydrocannabinolic acid A (THCA-A) were obtained from Cayman Chemicals (Ann Arbor, United States). Extraction gas butane (BUT), and dimethylether (DME) were provided by RSonic (Berlin, DE) and Dexso GmbH (Pratteln, CH), respectively. Microbiological growth media Mueller-Hinton Broth (MHB), Sabourad dextrose agar (SDA) were bought from OXOID (Prague, Czechia), horse defibrinated blood was purchased from Thermo Fisher Scientific (Waltham, United States), antibiotics clotrimazole (CLT), chloramphenicol (CLP), ampicilin (AMP), terbinafine (TB), and RPMI 1640 medium from Sigma-Aldrich.

The dried inflorescences from two cannabis genotypes – Forbidden Fruit (FF) (Czech Seed Bank, Czechia) and Chocolope (CHP) (DNA Genetics, NL) were used for the preparation of extracts. The plants were cultivated under controlled indoor conditions at the Department of Food Science, Faculty of Agrobiology Food and Natural Resources, Czech University of Life Sciences in Prague in 2020. Each genotype was cultivated on the table (2 m²) equipped with drip irrigation system. Plants were placed in pots containing Euro Pebbles (expanded clay) as growth medium. Air ventilation unit maintained the room temperature and humidity between 22 and 30°C, and 40 and 70%, respectively. The microclimatic conditions were adjusted according to the plant growth phase. Double-ended high-pressure sodium lamps were used to provide a suitable spectrum of light at a power of 1,000 W. The detailed information about plant nutrition were described previously by Janatová et al. (2018). Each cannabis genotype was extracted by three types of solvents – ethanol (EtOH), butane (BUT), and dimethyl ether (DME). To prepare first type of extract, 60 g of dried homogenized cannabis inflorescences were macerated for 48 h in 96% 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. The Dexso extractor was used for the preparation of butane and dimethyl ether extracts. The extractor was filled with 30 g of plant material to which 500 mL bottle containing compressed BUT or DME was connected. The gas flow was allowed to enter, and the extract was collected at the bottom to the teflon paper. To remove the residues of the gases, the extracts were placed for 2 h under the vacuum at Glass Vac (BVV, United States).

Ten milligrams of each extract were dissolved in 1 mL of MeOH, filtered through 0.45 μm nylon filters (Agilent, United States), transferred to a HPLC vial, and submitted for analysis.

The apparatus consisted of UltiMate 3000 HPLC system (Thermo Fisher Scientific, United States) and was equipped with UV detector. Cannabinoids analysis was performed on an Excel SuperC18 column (250 × 4.6 mm, 3 μm, 90 A; ACE, Scotland). Compounds were separated using a gradient elution employing water + 0.075% FA (A) and ACN + 0.5% FA (B) as mobile phase under following gradient (A:B): 0′, 30:70; 1′, 30:70; 10′, 0:100; 12′, 0:100; 14′, 30:70. Temperature of the column compartment was set at 40°C. Injection was 10 μL and flow rate was 1 mL/min. Cannabinoids were detected at wavelengths between 190 and 400 nm. Quantification was done under 210 nm. The evaluation of the acquired data was performed using the Chromeleon 7.2 software (Thermo Fisher Scientific, United States). Standard calibration curves were prepared in a concentration range of 2–100 μg/mL with six concentration levels (100, 50, 20, 10, 5, and 2 μg/mL). UV peak areas of the standards (at each concentration) were plotted against the corresponding standard concentrations (in μg/mL) using weighted linear regression to generate a standard curve.

The extracts were dissolved in methanol at a concentration of 2 mg/mL, transferred to the vial, and submitted to the GC-FID or GC-MS analysis.

To identify the terpenic profile, the cannabis extracts were analyzed by GC 7200 coupled with 7890 B qTOF mass detector (Agilent). Separation of individual volatile components was performed on a HP-5MS column (30 ms, 0.25 mm i.d., 0.25 μm; Agilent). Helium was used as a carrier gas; the flow was set at 1 mL.min^–1. The oven temperature program began at 60°C with a 3.5 min hold; the temperature was increased at a rate of 3.5°C/min to 155°C and then again at a 30°C min^–1 rate to 300°C and held for 10 min (total run time: 45 min). The temperature of the quadrupole was maintained at 230°C and of the ion source at 230°C. The compounds were measured in a scan mode in a range of 55–700 Da. The retention indices of each analyte were calculated from the retention times of n-alkanes by linear interpolation as previously described by Kováts. Identification of the analyzed volatiles was carried out by comparing the spectra with the spectra of available standards and/or by comparing their spectra and retention indices with the NIST Mass Spectral Search Program, version 2.2. The relative quantification of terpenes was performed by 7890A GC coupled with FID detector (Agilent) under the same chromatographic conditions. The detector conditions were set up as follows: t = 300°C; flow of gases - air: 400 mL/min; H₂: 30 mL/min; and N₂ make up flow: 5 mL/min. All samples were analyzed in triplicate.

The data were processed in Excel and STATISTICA 12 software (StatSoft, Tulsa, United States). As the data did not show normal distribution, the significant differences were evaluated by non-parametric Kruskal-Wallis test.

The extraction yield by both gases was on average 4% higher compared to traditional maceration in ethanol (Table 1). The color of BUT and DME extracts was golden, while EtOH extracts were dark green. This indicates a higher content of chlorophyll content in EtOH extracts. High variability in cannabinoid content was observed if different solvents were used for the extraction of a single strain (Table 1). The most significant differences in the cannabinoid composition were determined in extracts prepared with ethanol. They contained less than half of the cannabinoids compared to the BUT and DME extracts (that is, CBGA ranged from 5.7 to 9.5 mg/g and 19 to 21 mg/g in EtOH and BUT/DME extracts, respectively (Table 1). The THC content in all extracts was similar. In general, the predominant cannabinoid in all extracts was THCA (190–617 mg/g), while the least present was THCV which was found only in Chocolope EtOH extract. The lower content of cannabinoids in EtOH extracts can probably be attributed to the higher polarity of the solvent and thus lower extraction ability of these compounds into the resulting extract (Politi et al., 2008; Romano and Hazekamp, 2013).

The higher variability in the terpene profile was observed between strains and between extracts of one strain prepared with different solvents (Table 2). The main terpenes (>7% each) in Forbidden Fruit extracts were ß-caryophyllene, ß-myrcene, selina-4(15),7(11)-diene, selina-3,7(11)-diene followed by humulene and ß-farnesene. The Chocolope strain extracts contained mostly ß-caryophyllene, ß-myrcene, (+)-limonene, linalool, and selina-3,7(11)-diene. The results indicate that both strains belong to caryophyllene-dominated cultivars (Lewis et al., 2018), which are also characterized by relatively high levels of humulene (Fischedick, 2017) and low levels of the often predominant α-pinene (Lewis et al., 2018).

The solvents had a statistically significant effect on the presence and relative ratio of individual terpenes in the extracts (Table 2). Both EtOH extracts compared to BUT and DME ones lacked most of the identified monoterpenes (i.e., thujene, α-pinene, ß-myrcene, and (+)-limonene). These terpenes were probably lost during the evaporation of the solvent under vacuum and increased temperature (Romano and Hazekamp, 2013). In conclusion, the gas extraction was more effective, BUT and DME extracts contained significantly higher amounts of active compounds. On top of that, the extraction was less time demanding (did not require the solvent evaporation step). Moreover, DME is a non-toxic gas commercially used for collagen extraction (EFSA, 2015). Therefore, it could be considered as a suitable alternative to ethanolic extraction.

All extracts showed very good to moderate activity against 18 of the 19 microorganisms tested. The MIC of the extracts ranged from 4 to 128 μg/mL and 32 to 256 μg/mL (Table 3) for bacteria and dermatophytes, respectively.

The most susceptible bacteria were S. aureus strains (MIC = 4–8 μg/mL). However, almost 10 times higher MIC values (MIC = 64–128 μg/mL) were determined for S. pyogenes. Such a significant decrease in the activity of cannabis extracts could be attributed to the presence of blood in the cultivation medium. As previously reported, in this type of media, cannabinoids become less effective due to their strong affinity to blood proteins (Garrett and Hunt, 1974; van Klingeren and Ten Ham, 1976). All extracts were also slightly less effective against S. aureus strains compared to isolated THC, CBD, CBG, and CBC (MIC = 0.5–2 μg/mL) (Appendino et al., 2008; Blaskovich et al., 2021). Generally, all the extracts except the EtOH ones were from more than 60% composed of cannabinoids (Table 1), therefore, they can be considered antimicrobial principles of cannabis. However, the role of the remaining biologically active compounds, especially terpenes, should not be omitted. The antimicrobial activity of the terpenes identified in extracts (Table 2) has been confirmed by many studies (Schofs et al., 2021). For example, ß-caryophyllene and α-pinene are very effective against S. aureus, S. epidermidis, S. saprophyticus and S. lugdunensis (Dahham et al., 2015; Ranarivelo et al., 2020). While myrcene has proven effects against S. epidermidis (Inoue et al., 2004) and limonene also has strong bactericidal effects against a diverse range of bacteria (Han et al., 2020, 2021).

The most sensitive dermatophytes were T. rubrum CCF 4879 (the most common species causing dermatophytosis in developed countries) and M. canis CCF 6025, while A. insingulare was resistant even to the highest concentrations of the extracts tested. The mode values of the MICs demonstrated that the EtOH extracts were slightly less effective (MIC = 128 μg/mL) compared to the DME and BUT extracts (MIC = 64 μg/mL). Even though some of the cannabis skin preparations are among other disorders also indicated for the treatment of bacterial and fungal skin diseases (Hashim et al., 2017; Anastassov et al., 2019), the scientific evidence about their activity against dermatophytes is very limited. So far, only Turner and Elsohly (1981) reported significant antifungal effects of CBC and its’ analogs against Trichophyton mentagrophytes (MIC = 6.25 – 50 μg/mL^–1). Also hemp water extracts proved to have moderate activity against T. rubrum and interdigitale (MIC = 500 μg/mL) but not against N. gypsea (Orlando et al., 2020). There is no doubt that cannabinoids have a beneficial effect on various skin problems such as skin inflammation, fibrosis, itch, pain, and improved wound healing (Hashim et al., 2017; Cintosun et al., 2020). However, their contribution to the antifungal activity remains questionable and a possible mechanism of action is yet to be elucidated. On the other hand, all extracts contained terpenes with established activity against dermatophytes. For example, essential oils from cannabis with high concentrations of ß-caryophyllene or ß-caryophyllene alone were very effective against various dermatophytes (Iordache et al., 2016; Orlando et al., 2021). Tavares et al. (2010) reported a strong MIC (1–5 μg/mL) of myrcene against a variety of microorganisms, including dermatophytes. Moreover, limonene, α-pinene, γ-eudesmol, germacrene (separated or as a apart of essential oil) also proved significant to moderate antidermatophytic activity (MIC = 0.8–250 μg/mL) which was sometimes even better than some used antifungals, e.g., griseofulvin (Sanguinetti et al., 2007; Singh et al., 2010; Pinto et al., 2013; Alves et al., 2018; Danielli et al., 2018). It is very probable that both groups of biologically active substances, i.e., cannabinoids and terpenes, have an important role in the antifungal activity of cannabis extract.

The MIC values of cannabis extracts presented in this study are higher in comparison to the currently used antifungals. Furthermore, as mentioned above, some of the isolated compounds can be more active than the extracts. Therefore, the presented results may raise the question about the real therapeutic potential of cannabis extracts in the treatment of various skin infections. Current antimicrobial and antifungal agents were design solely to kill the target microorganism (Gnat et al., 2020). On the other hand, the effect of cannabis extract can be more complex. Not only it can inhibit or kill the microorganism growth, but both cannabinoids and terpenes due to their anti-inflammatory, antipruritics and antinociceptive properties can have positive effect on other symptoms accompanying the skin infections. Cannabinoids can also interact with the skin endocannabinoid system that plays an important role in its homeostasis (Martins et al., 2022). Terpenes are understood as skin penetration enhancers (Aqil et al., 2007), thus can facilitate the penetration of highly lipophilic cannabinoids to deeper skin layers or by other ways enhance their activity. This so called “entourage effect” was quite recently confirmed for some of the terpenes present in cannabis (LaVigne et al., 2021). From the empirical evidence provided by many users of cannabis topicals, it is known that the preparations are usually well accepted and reported to be effective (Martins et al., 2022). Although topicals are often homemade or from unofficial sources and contain unknown concentrations of various cannabis chemotypes (Mahmood et al., 2022). Therefore, their low toxicity during long-term use can be expected. These facts confirm that the therapeutic potential of cannabis in the treatment of skin infections should not be overlooked. Especially for the treatment of dermatomycosis that is usually long and can be associated with adverse effects. The side effects are unacceptable for elderly, pregnant women and children (Gnat et al., 2020).

To confirm the safety and efficacy of cannabis extracts that will lead to their use in medical practice, a lot of work has to be done. Among others identification of most active cannabis chemotypes, standardization of the extracts, determination of effective dose, pharmacokinetics, toxicological risk assessment and clinical studies are needed.