The application of plant material or seeds of Cannabis sativa for the preparation of early antibiotic treatments of different infections have already been reported from folk medicine at the end of the 19 th century, as described in early review articles by Kabelik [98,99].

First systematical evidence for the antimicrobial activity of constituents of Cannabis sativa was given by the Dissertation Krejci in 1950 [100]. His comprehensive bacteriological in vitro investigations on the antibacterial effects of extracts of Cannabis sativa on different bacteria were later published in 1958 [101]. Krejci reported a good activity of the cannabis extracts against the growth of Gram-positive bacteria, while Gram-negative bacteria were not affected. Furthermore, he prepared a salve containing 2% extracts and demonstrated its therapeutic efficacy in the treatment of skin infections with Staphylococci [101]. Independently, Drobotko et al. described the occurrence of antibacterial activity in extracts from hemp in 1951 [102], while Ferenczy et al. reported the finding of the occurrence of antibacterial activity in extracts obtained from seeds of hemp in 1956 [103,104]. Ferenczy et al. demonstrated the potency of the antimicrobial activity using dilutions of their extracts and observed that their activity was pH-dependent [104].

Since these early findings, numerous of reports on the antimicrobial activity of different extracts prepared from Cannabis sativa on Gram-positive, Gram-negative bacteria, and different fungi [105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129], which have also been partially reviewed elsewhere [98,130,131].

The additional scientific knowledge from these studies was rather limited, as in most cases, complex mixtures with undetermined composition or partially determined mixtures of over 60 compounds [112] have been utilized for the evaluation of the antimicrobial activity, thus, clear compound specific pharmacological conclusions are impossible. Considering the variety of natural products from different classes, which have been isolated from Cannabis sativa [132,133,134] and the complex pharmacology of phytocannabinoids mentioned above, studies on isolated single cannabinoids are required to fully understand their antimicrobial potential.

The first early study on the antimicrobial activity of specific, isolated cannabinoids from Cannabis sativa were reported by Schultz and Haffner in 1958 [27]. They isolated cannabidiolic acid (CBDA) (11) and compared its antibacterial activity against S. aureus, Bacillus subtilis, and Escherichia coli to cannabidiol (CBD) (10), the corresponding diacetate (31) as well as its sodium salt (32) (Figure 5).

While CBD (10) and CBDA (11) show a potent antibacterial activity against the Gram-positive bacteria S. aureus and B. subtilis, an activity against the Gram-negative E. coli was not observed. As the poor water solubility of the compounds was limiting their investigations, the diacetate 32 and its sodium salt 32a were generated, which showed significantly increased water solubility. Nonetheless, these derivatives were inactive against all tested bacteria [27].

A second study comparing the antibiotic activity of CBD (10), CBDA (11), and the acetylated cannabidiolic acid 32 was published by Krejci et al. in 1959 [28]. In contrast to the investigations of Schultz and Haffner [27], 32 displayed still potent antibacterial activity, although decreased compared to CBD (10) and CBDA (11) [28].

In 1965, Melchoulam et al. mentioned the antibacterial activity of CBG (17) and CBGA (18) against Gram-positive bacteria in a short side note of their publication on the isolation and structure determination of natural cannabinoids. However, neither information on bacterial strains nor specific MIC (Minimal Inhibitory Concentration) data were given [135].

In 1976, Klingeren and Ham published a study on the antibacterial activity of Δ⁹-THC (1) and CBD (10), in which both cannabinoids were screened for their antibacterial activity against the Gram-positive bacteria S. aureus, Streptococcus pyogenes, Streptococcus milleri and Enterococcus faecalis and the Gram-negative pathogens E. coli, Salmonella typhi and Proteus vulgaris [26]. Both compounds displayed significant antibacterial activity against all Gram-positive bacteria tested (Table 1, Entry 1 and 2), but were in accordance with the results of Krejci et al. [28] inactive against the Gram-negative ones. Furthermore, the authors proved on cultures of S. aureus that both compounds are bacteriostatic as well as bactericidal.

In 1981, Turner and ElSohly published a report on the in vitro evaluation of antimicrobial activity of CBC-type cannabinoids against different Gram-positive bacteria and fungi including an in vivo study in rats investigating the anti-inflammatory activities in an carrageenan-induced rat paw edema assay [25]. In 1982, the same research group published an expanded study on the antibacterial and antimycotic activities of further CBC-type and CBG-type cannabinoid derivatives [29]. The results of both studies are summarized in Table 2. In this study, CBC (19) showed the highest activity of all tested CBC-type cannabinoids against S. aureus and B. subtilis with a MIC of 1.56 µg/mL and 0.39 µg/mL, respectively, even superior compared to Streptomycin (Table 2, Entry 1). Interestingly, isoCBC (33) (Table 2, Entry 2), bearing the lipophilic side chain and one hydroxyl group in exchanged positions displayed similarly potent activity against B. subtilis with a MIC of 0.78 µg/mL. The truncation of the lipophilic side chain in both isoCBC- and CBC-type cannabinoids led to a clear 2–8-fold drop in antibacterial activity (Table 2, Entries 3, 4 and 5), with the exception of isoCBC-C₁ (35) exhibiting a MIC of 0.78 µg/mL against B. subtilis (Table 2, Entry 6). Moreover, the double bonds in the GPP-derived side chain seem to be less important for high activity as the reduced derivative 38 (Table 2, Entry 7) showed comparable antibacterial activity as CBC (19). Interestingly, a moderate antimycobacterial activity Mycobacterium smegmatis was observed for some of the compounds with best activity for CBC-C1 (37) and 38 exhibiting MIC values of 3.12 µg/mL comparable to Streptomycin (Table 2, Entry 4). The structure–activity relationships for the antibacterial activity of CBC-type cannabinoids can be concluded as illustrated in Figure 6. The antimycotic activity of the compounds was less pronounced. Only CBC-C₁ (37) displayed a moderate antimycotic activity against Sachcheromyces cerevisiae with a MIC of 6.25 µg/mL, clearly less active compared to Amphotericin B. The activity against Trichophyton mentagrophytes, a pathogenic fungus, was comparable to Amphotericin B with 2 two-fold higher MIC of 6.25 µg/mL.

The authors also demonstrated that all tested CBC-type cannabinoids in this study reveal a higher anti-inflammatory activity compared to phenylbutazone or aspirin [25].

The CBG-type derivatives isoCBG-C₁ (39) and CBG-C₁ (40) displayed comparably strong antibacterial activity with MIC values in the range of 1.56–6.25 µg/mL [29]. According to the trends observed for the CBC-type derivatives, the reduction of the double bond in the GPP-derived side chain was well tolerated, as derivative 41 showed similar antibacterial activity against S. aureus and B. subtilis and good activity against M. smegmatis with a MIC of 3.12 µg/mL. Again, the antimycotic activities of the CBG-type compounds were less pronounced with only moderate activity. CBE (14) showed no noteworthy antibacterial or antimycotic activity [29].

Between 2008 and 2015, a consortium of researchers around Ross and ElSohly published a series of studies on the isolation and antimicrobial evaluation of novel minor cannabinoids from high potency variants of Cannabis sativa [136,137,138,139,140]. In their studies, CBG (17) (Figure 7) exhibited selective moderate antimicrobial activity against Mycobacterium intracellulare with an IC₅₀ value of 15.0 μg/mL [137]. Furthermore, CBGA (18) displayed moderate antileishmanial activity with an IC₅₀ value of 12.0 μg/mL against Leishmania donovani, while both compounds did not show any cytotoxic activity on African green monkey kidney fibroblast [137]. Compound 42 (Figure 7), a cannabinol derivative bearing an additional hydroxy function in the lipophilic side chain, showed moderate anti-MRSA (IC₅₀ = 10.0 µg/mL), moderate antileishmanial (IC₅₀ = 14.0 µg/mL), and reasonable antiprotozoal activity against Plasmodium falciparum (D6 clone) and P. falciparum (W2 clone) with IC₅₀ values of 3.4 and 2.3 µg/mL, respectively [138]. The isolated cannabichromanone derivatives 43 and 44, both differing in their oxidation pattern, displayed antiprotozoal activity against P. falciparum (D6 and W2 clone) with IC₅₀ values of 3.7 and 3.8 µg/mL, and 4.7 and 3.4 µg/mL, respectively [139]. Meanwhile, the cannabichromanone derivatives 45 and 46 isolated in the same study showed no antiprotozoal activity. However, all of these cannabichromanone derivatives 43–46 showed moderate antileishmanial activity with IC₅₀ values of 14.0, 14.0, 12.5, and 9.0 µg/mL, respectively [139]. Compound 47, a prenylated ester of cannabinol, showed moderate antimycotic activity against C. albicans ATCC 90028, with an IC₅₀ value of 8.5 µg/mL, and good antiprotozoal activity against P. falciparum (D6 clone) and P. falciparum (W2 clone), with IC₅₀ values of 2.7 and 2.4 µg/mL, respectively. A CB1R binding assay indicated that 47 is non-psychotropic [136].

Compound 52 (Table 3, Entry 5), a hydroxylated derivative of CBNA, exhibited good antimycotic activity against Candida albicans (IC₅₀ = 4.6 µM) as well as low antimycobacterial activity against M. intracellulare (IC₅₀ 30.6 µM) [140]. The CBG derivative 49 (Table 3, Entry 2) and CBG derivative 51 (Table 3, Entry 4) possessed moderate anti-MRSA activity with IC₅₀ values of 24.4 and 6.7 µM, respectively. Furthermore, both compounds displayed additional low antileishmanial activity against L. donovani. A good anti-MRSA activity with an IC₅₀ value of 3.5 µM was observed also for the hydroxylated CBN derivative 53 (Table 3, Entry 6) as well [140]. The CBC-type derivative 48 (Table 3, Entry 1) and CBG derivative 50 (Table 3, Entry 3) showed moderate antiprotozoal activity against P. falciparum (D6 and W2 clone), while 50 additionally displayed a moderate antileishmanial activity with an IC₅₀ value of 10.7 µM [140]. In these studies, rather, single values of inconsistently screened different antimicrobial activities were reported, thus, clear structure-activity relationships are hard to conclude.

In 2008, Appendino et al. [24] reported a more focused structure–activity relationship study on the anti-staphylococcal activity of different naturally occurring cannabinoids and synthetic derivatives against a broad range of various multi-drug resistant strains of S. aureus, including EMRSA-15, one of the main epidemic methicillin-resistant strains [141], SA-1199B, a multidrug-resistant strain that possesses a gyrase mutation, and in addition overexpresses the NorA efflux mechanism [142], RN-4220, a macrolide-resistant strain [143], XU212, and a tetracycline-resistant strain overexpressing the TetK efflux pump [144], as summarized in Table 4.

In their studies, all major cannabinoids Δ⁹-THC (1) (Table 4, Entry 10), CBD (10) (Table 4, Entry 1), CBG (17) (Table 4, Entry 5), CBC (19) (Table 4, Entry 4), and CBN (7) (Table 4, Entry 12) obtained by isolation from Cannabis sativa display potent anti-staphylococcal activity against all tested strains, with MIC values in the range of 0.5–2.0 µg/mL in most cases, even superior, compared to the reference antibiotics norfloxacin, erythromycin, tetracycline, and oxacillin. Given their non-psychotropic activity profiles, the authors selected CBG (17) and CBD (10), and synthetically generated several derivatives for further structure–activity studies, which are illustrated in Figure 8. The structural exploration revealed that in both cannabinoid types, the phenolic hydroxyl moieties are crucial for activity as mono- or di methylation, as mono-or deacetylation led to complete loss of anti-staphylococcal activity (Table 4, Entries 3, 7, 8, and 9). In light of the potent activity of the monophenols Δ⁹-THC (1) (Table 4, Entry 10), CBC (19) (Table 4, Entry 4), and CBN (7) (Table 4, Entry 12), the complete loss of activity for the mono methylation is not fully consistent and indicates a more complex SAR picture.

The carboxylic acid precursors CBGA (18) (Table 4, Entry 6) and CBDA (11) (Table 4, Entry 2) similar to Δ⁹-THC (2) (Table 4, Entry 11) retained their potent activity; however, esterification of the carboxylic acid moieties was detrimental for the activity. Thus, compounds 57, 58, 62, and 63 showed no noteworthy activity (Table 4, Entries 3, 8 and 9). Additionally, bigger substituents in the same position, such as in the prenylated CBD-derivative 67, were also not tolerated and led to complete loss of activity. While the biosynthetic cannabinoid precursor olivetol 68 showed very low activity (Table 4, Entry 17), no activity at all was observed for resorcinol 69 missing the lipophilic side chain (Table 4, Entry 18). Thus, indicating that both the lipophilic C5 side chain as well as GPP-derived lipophilic side chain are crucial for high activity. In addition, the substantially decreased activity of camagerol (66), a dehydroxylated CBG-derivative revealed that the overall polarity in the GPP-derived lipophilic side chain is important for high activity as well. Interestingly, the synthetic isomers of CBD (Compound 64, Table 4, Entry 13) and CBG (Compound 65, Table 4, Entry 14), bearing the lipophilic side chain and one hydroxyl group in exchanged positions displayed potent activity comparable to their corresponding natural counterparts indicating that both moieties, although crucial for activity, can vary in their position at the phenolic core. A trend that has been observed for CBG-C₁-type and CBC-type cannabinoid derivatives earlier [29], suggesting that they may rather serve as lipid affinity modulators influencing water solubility and cellular bioavailability [24].

Recently, a very comprehensive study on the antimicrobial, antibiofilm and anti-persister cell activities of phytocannabinoids from Cannabis sativa against MRSA was published by Brown and co-workers comprising in vitro and in vivo investigations of 18 commercially available cannabinoids, pre-cannabinoids and synthetic isomers [23]. The results of this study are summarized in Table 5.

Seven of the tested cannabinoids showed potent antibacterial activity against MRSA USA300, a highly virulent and prevalent pathogen, including Δ⁹-THC (1), Δ⁸-THC (5), CBN (7), CBD (10), CBG (17), CBCA (20), and exo-THC (22) (Table 5, Entries 1, 5–7, 11, 14 and 16), in accordance with previously published data [24,25,26,29]. In contrast to the study by Appendino et al. [24], in most cases, with the exception for CBCA (20) (Table 5, Entry 14), the authors observed a moderate loss of activity for the corresponding carboxylic acid bearing pre-cannabinoids. Thus, Δ⁹-THCA (2), CBDA (8) and CBGA (18) (Table 5, Entries 2, 8 and 12) were 2–8-fold less active than their decarboxylated congeners. Furthermore, the cannabinoids Δ⁹-THCV (3) and CBDV (12) (Table 5, Entries 3 and 9), bearing a truncated n-propyl lipophilic side chain were 2- and 4-fold less active compared to their congeners Δ⁹-THC (1) and CBD (10), respectively, bearing the n-pentyl side chain. The negative impact of side chain truncation and carboxylic acid substitution on the anti-staphylococcal activity was additive detrimental as both Δ⁹-THCVA (4) and CBDVA (13) (Table 5, Entries 4 and 10) exhibited an additional 4-fold decrease in anti-staphylococcal activity compared to Δ⁹-THCV (3) and CBDV (12), respectively. Noteworthy, both major human metabolites of Δ⁹-THC (1), 11-hydroxy-Δ⁹-THC (11-HO-Δ⁹-THC, 26) and nor-9-carboxy-Δ⁹-THC (11-nor-9CTHC, 27) (Table 5, Entries 17 and 18) as well as CBL (21) Table 5, Entry 15) were inactive at the highest screen concentration.

Two major features strongly contributing to the high virulence and strong persistence of MRSA in both human host organism and environment are its high notion to form biofilms on necrotic tissue and medical devices [13,14] and its ability to form non growing, dormant persister subpopulations [145,146,147,148,149]. Both making the pathogen less susceptible for host immune factors displaying high levels of tolerance to antibiotics, such as gentamicin, ciprofloxacin, and vancomycin [145,146,147,148,149], and can led to severe chronic infections [150,151,152].

Recently, antibiofilm activity of ethanolic extracts of Cannabis sativa, likely containing phytocannabinoids, has been reported by Frasinetti et al. [121]. However, the study of Brown and co-workers for the first time allowed a direct attribution of antibiofilm activity to the tested cannabinoid derivatives [23]. As summarized in Table 5, the antibiofilm activity of all tested cannabinoids against MRSA USA300 strain correlated well with their specific anti-staphylococcal activity. Thus, the major cannabinoids, such as Δ⁸-THC (5), CBGA (18), CBG (17), and exo-THC (22) completely suppressed biofilm formation at concentration of 2 µg/mL (Table 5, Entries 5, 11, 12 and 16), CBG (17) displaying the most potent antibiofilm activity with an IC₅₀ value of 0.5 µg/mL [23]. Remarkably, in addition, CBG (17) was demonstrated to eradicate preformed biofilms of MRSA USA300 at concentrations of 4 µg/mL [23]. Brown and co-workers also screened for the ability of the cannabinoids to kill persister cells (Table 5). The anti-persister activity correlated with the activity against actively dividing cells and CBG (17) showed the most active cannabinoid killing persisters in concentration-dependent manner starting at 5 µg/mL, again. Impressively, CBG (17) was able to completely eradicate a population of ~108 CFU/mL MRSA persisters below detection thresholds within 30 min of treatment [23].

In order to elucidate the molecular target of CBG (17), Brown and co-workers repeatedly challenged MRSA USA300 with various lethal concentrations ranging from 2–16-fold MIC or sequential subcultures in solution with sublethal concentration of CBG (17) to select for spontaneous resistance. However, no spontaneous resistant mutants were obtained, suggesting a low frequency of resistance of below 10⁻¹⁰ for MRSA [23]. A comprehensive chemical genomic analysis on the model bacterium B. subtilis at sublethal concentrations of CBG (17) than led to the identification of 41 transposon mutants, which revealed an enrichment of genes encoding for proteins that are localized at the cytoplasmic membrane, and genes encoding for proteins involved in processes taking place at the cytoplasmic membrane. Thus, indicating that CBG (17) acts via a disruption of the integrity of the cytoplasmic membrane, which was confirmed through treatment of MRSA cells with a membrane-potential sensitive fluorescent probe in the presence of CBG (17), leading to a dose-dependent increase of fluorescence occurring at MIC of CBG (17) [23]. Unfortunately, a hemolytic activity of CBG (17) on human erythrocytes was observed at 32 µg/mL, above its MIC of 2 µg/mL, but with a low therapeutic index of 16:1 [23].

However, a pharmacokinetic study of 120 mg/kg doses of CBG (17) in rats and mice have reported no signs of acute toxicity [153]. Brown and co-workers demonstrated the antibacterial efficacy of CBG (17) in vivo at 100 mg/kg doses in a murine systemic infection model, displaying a significant reduction of bacterial burden by a factor of 2.8 log₁₀ in CFU compared to the control, while being well tolerated in mice [23].

Finally, Brown and co-workers discovered that the antibacterial activity of several cannabinoids could be significantly increased against Gram-negative pathogens, which have been reported to be less susceptible towards cannabinoids [26,140] in the presence of sublethal concentrations of polymyxins perturbating the outer membrane of Gram-negative bacteria. The authors could show, that CBG (17), which was inactive against E. coli (>128 µg/mL), was active in the presence of a sublethal concentration of 0.062 µg/mL of polymyxin B with a MIC of 1 µg/mL.

This observation uncovered the hidden broad-spectrum antibiotic activity within this synergistically acting polymyxin B adjuvant combination therapy approach, and allowed Brown and co-workers to demonstrate the potent efficacy of CBG (17) against clinical isolates of Gram-negative priority pathogens such as Acinetobacter baumannii, E. coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa.