Antimicrobial Activity of Some Steroidal Hydrazones

Twelve steroid based hydrazones were in silico evaluated using computer program PASS as antimicrobial agents. The experimental evaluation revealed that all compounds have low to moderate antibacterial activity against all bacteria tested, except for B. cereus with MIC at a range of 0.37–3.00 mg/mL and MBC at 0.75–6.00 mg/mL. The most potent appeared to be compound 11 with MIC/MBC of 0.75/1.5 mg/mL, respectively. The evaluation of antibacterial activity against three resistant strains MRSA, E. coli and P. aeruginosa demonstrated superior activity of compounds against MRSA compared with ampicillin, which did not show bacteriostatic or bactericidal activities. All compounds exhibited good antifungal activity with MIC of 0.37–1.50 mg/mL and MFC of 1.50–3.00 mg/mL, but with different sensitivity against fungi tested. According to docking studies, 14-alpha demethylase inhibition may be responsible for antifungal activity. Two compounds were evaluated for their antibiofilm activity. Finally, drug-likeness and docking prediction were performed.


Introduction
The development of new antimicrobial agents is still attracting the interest of medicinal chemists since the resistance of bacterial pathogen strains is a major problem.
One of the reasons for the fast multiplication of bacteria is their ability to exchange genes with each other, leading to the development of resistance. On the other hand, the interest in the discovery of new antimicrobial agents is because during the past 30+ years, the FDA has approved only two new antimicrobial drugs: linezolid and daptomycin.
Despite the fact that many compounds have been synthesized and tested, their clinical use has been restricted due to the high risk of toxicity and pharmacokinetic deficiencies. Thus, the scientists have directed their efforts at developing novel approaches to antimicrobial therapy, aiming to overcome the resistance problem [1][2][3][4]. Another big problem is biofilm formation, which plays a crucial role in bacterial infection and antimicrobial resistance. There is increasing proof that cells in biofilms, on a biotic or abiotic surface, are 1000-fold more resistant to conventional drugs than planktonic cells [5,6]. The problem is that upon being established, biofilms become difficult to eliminate and as a result, chronic and persistent infections [7] appear. As reported in the literature [8,9], one of the main Gram-positive pathogens causing biofilm-associated infections is Staphylococcus aureus. Thus, another need is for the development of new agents that are able to inhibit S. aureus biofilm formation.
Hydrazones of different chemical classes possess diverse biological and pharmacological properties such as antimicrobial, anti-inflammatory, analgesic, antifungal, antitubercular, antiviral, anticancer, antiplatelet, antimalarial, anticonvulsant, cardio protective, anthelmintic, antiprotozoal, anti-trypanosomal, anti-schistosomiasis etc. [10][11][12]. Hydrazones contain two connected nitrogen atoms of different nature and a C-N double bond that is conjugated with a lone electron pair of the terminal nitrogen atom. These structural fragments are mainly responsible for the physical and chemical properties of hydrazones. The combination of thehydrazono group with other functional groups leads to compounds with a unique physical and chemical character [13]. It is noteworthy that there is an approved FDA drug with a hydrazone scaffold, namely levosimendan, a calcium sensitizer used in the management of acutely decompensated congestive heart failure ( Figure 1). On the other hand, steroidal compounds are a class of bioactive substances playing a major role in living organisms with a wide representation in the natural world. Steroidal derivatives attracted the interests of scientists, especially medicinal chemists, due to their wide range of biological activities [10,[13][14][15]. They are known to possess antimicrobial [16,17], antioxidant [17] and anticancer [17] activities. In the last few decades, the efforts have concentrated on rational modification of steroid molecules due to their lower toxicity, vulnerability to multi-drug resistance and high bioavailability to penetrate the cell wall and to be linked to nuclear and membrane receptors. Vollaro et al. [18] reported the investigation of the in vitro effect of pregnadiene-11-hydroxy-16α,17α-epoxy-3,20-dione-1 (PYED-1) on biofilm formation.
Encouraged by these observations, and based on our previous work [25][26][27], herein we report the synthesis of two novel 5α-steroidal hydrazones and the evaluation of antimicrobial activity of newly and earlier synthesized compounds.
Thus, the purpose of our study was in silico and biological evaluation of the antimicrobial potential of twelve steroidal hydrazino derivatives, including action on the resistant strains.

Chemistry
In the continuation of our research on new bioactive N-containing 5α-steroids, ten steroidal hydrazine derivatives, that we synthesized earlier on the basis of steroidal ketones [25][26][27][28][29], and two new compounds were prepared and evaluated for their antibacterial and antifungal actions.

PASS Predictions
PASS prediction of antimicrobial activities was performed for previously synthesized compounds (1, 3-11), as well as for two new designed ones. The antibacterial activity was predicted only for two compounds with Pa values in the range 0.164-0.313, and antifungal activity for almost all compounds with Pa values in the range 0.143-0.470. The calculated Pa values for all compounds were less than 0.5, indicating their relative novelty compared to the structures of the compounds from the PASS training set [30]. This may be proof that the studied compounds have some features dissimilar from those of well-known antimicrobial agents, which may indicate their innovative potential.

Antibacterial Activity
Synthesized compounds were tested for their antibacterial activity against a panel of nine bacteria species, using the microdilution method for the determination of minimal inhibitory and minimal bactericidal concentrations (MIC and MBC, respectively). As reference drugs, ampicillin and streptomycin were used. The antibacterial activity of tested compounds (Table 1) in general was low to moderate, except in some cases where it was good, with MIC ranging from 0.37 to 3.00 mg/mL and MBC at 0.75-9.00 mg/mL, presented in Table 1. The order of activity can be presented as follows: 11 = 12 > 4 = 5 > 3 > 1 > 6 > 7 > 10 > 8 > 9 > 2. Compound 11 appeared to be the most potent among those tested, with MIC and MBC of 0.75/1.5 mg/mL, respectively, but less than for both reference drugs. The most sensitive bacterium was found to be B. cereus, whereas S. aureus was the most resistant one. The structure-activity relationship studies revealed that the presence of 3-nitrobenzohy drazide at position 17 of 3α-hydroxy-5α-androstan-17-one 11 is beneficial for antibacterial activity. The replacement of the nitro group in the benzene ring by Br led to compound 12 havingthe same good influence as the previous one, on activity. The replacement of the substituted benzene ring by isonicotinoylhydrazide, and 3α-hydroxy-5α-androstan-17-one by 3α-hydroxy-5α-pregnan-20-one, resulted in compound 4 with slightly lower activity. It is interesting to notice that the presence of thethiosemicarbazide substituent at position 20 (5) of the steroid ring, in place of isonicotinoylhydrazide (4) as the substituent in position 20, exhibited the same activity as compound 4. Replacement of hydroxy group in position 3 by phenylacetoxy (3) decreased the antibacterial activity more, while replacement by theazide group (2) was detrimental.
The evaluation of antibacterial activity of these compounds against three resistant strains, MRSA, E. coli and P. aeruginosa revealed that compounds were more potent against MRSA than ampicillin, which did not show bacteriostatic or bactericidal activity, while against the two other resistant strains, it did not show bactericidal activity. The order of activity of the tested compounds against resistant strains can be presented as 6 = 7 > 1 > 3 > 4 > 11 = 12 > 5 > 8 > 10 > 9 > 2, with compounds 6 and 7 being the most potent (MIC/MBC at 0.75-1.50 mg/mL and 1.50-3.00 mg/mL, respectively). It should be mentioned that compounds 3-5, 8-12 did not show any activity against the resistant E. coli strain. It is interesting to notice that compounds 6 and 7 were more potent against resistant E. coli than E. coli strains, while the opposite was observed for compounds 1 and 2. On the other hand, compounds 1,3,6,7 and 8 exhibited better activity against P. aeruginosa and against resistant P. aeruginosa, while compounds 2,4,10-12 demonstrated the same activity against both of these strains. In general, our compounds showed better activity against the resistant P. aeruginosa strain than against two other resistant strains, being less potent than the reference drug streptomycin.
In the case of structure-activity relationship studies against resistant strains, it was found that the presence of athiosemicarbazide substituent (6) as well as anisonicotinoylhydrazide one (7) in position 17 of 3α-hydroxy-5α-androst-9(11)-en-17-one was favorable for the activity against resistant strains.
Finally, it should be mentioned that compounds tested have different behavior against ATCC and resistant strains. The only common behavior against both strains was observed for compound 2, which demonstrated a negative effect on antibacterial activity in both cases.

Inhibition of Biofilm Formation
After the observation of the antifungal activities of compounds, antibiofilm activities were assessed. We observed that compounds 1 and 8 possessed higher antifungal activity against C. albicans and all tested micro fungi than other used compounds. The strain used for the antibiofilm assay was C. albicans. Incubation with compounds 1 and 8 hasreduced the ability of C. albicans (Figures 2 and 3) to attach to the surface and begin the process of biofilm formation. A concentration equal to the previously determined MIC has reduced the biofilm biomass by 33% and 15% for compounds 1 and 8, respectively. When applied in 0.5 and 0.25 MIC concentrations of compound 1, inhibition percentages were almost the same, about 18% (Figure 3). The reference drug, Ketoconazole, possessed better biofilm activity than the compounds, reducing the biofilm biomass by 50%, 47% and 25% for MIC concentrations 0.5 MIC and 0.25 MIC, respectively (Figure 2).  Even twice as low concentrations (0.5 MIC) of compound 8 limited the biofilm forming ability and induced more than 16% inhibition in C. albicans. The impact on the fungal biofilm was less profound and the 0.25 MIC concentration of 8 was able to reduce the biofilm formation byless than 5% (Figure 3).

Docking to Antifungal Targets
In order to investigate the possible mechanism of antifungal activity of compounds, all of them along with the reference drug ketoconazole were docked to lanosterol 14αdemethylase of C. albicans and DNA topoisomerase IV. The results are presented in Table 3. Based on docking studies, all compounds bind to the CYP51 Ca enzyme similarly to the reference drug ketoconazole (Figure 4). The most active compound 8 binds to the Fe of the heme and interacts hydrophobically and aromatically with the heme. Additionally, compound 8 forms a hydrogen bond between the oxygen atom of the C=O group and the side-chain hydrogen of Tyr64. Hydrophobic interactions were also detected between residues I Tyr118, Thr122, Ile131, Tyr132, Leu376, Met508 and the compound ( Figure 5). Ketoconazole also forms aromatic and hydrophobic interactions with the heme group. It has been shown, however, that compound 8 forms a more stable complex with the enzyme, possibly due to its interaction with heme's iron. It is likely that this is the reason why this compound has a better antifungal effect than ketoconazole.  The superposition of compound 3 and ketoconazole ( Figure 6) explains its good antifungal activity. Similar to ketoconazole, compound 3 inserts the binding site of the enzyme, forming an additional hydrogen bond with residue Met508. In addition, it exhibits the same hydrophobic and aromatic interactions with the heme group as ketoconazole, which explains its good inhibition profile.

Drug-Likeness
All tested compounds were evaluated for their drug-likeness and bioavailability scoresand the results are presented in Table 4. According to the prediction, the bioavailability score for most of the compounds was about 0.55, except for compounds 3, 9, 10 and 12 with 0.17 values. Despite these compounds exhibiting two violations of Lipinski's rule of five, they have excellent drug-likeness scores ranging from 0.74 to 1.54. Thus, it can be concluded that they have good oral bioavailability and drug-likeness profile (Figure 7).
The antifungal activity of all investigated samples was tested on strains obtained from the Mycological Laboratory, Department of Plant Physiology, Institute for Biological Research "Sinisa Stankovic", National Institute of Republic of Serbia, Belgrade, Serbia. The following fungi Aspergillus fumigatus (ATCC 1022), Aspergillus niger (ATCC 6275), Trichoderma viride (IAM 5061), Penicillium funiculosum (ATCC 36839), P. verrucosum var. cyclopium (food isolates) and Candida albicans (ATCC 10231) were tested. The detailed explanation is given in our previous papers [32,33].
Commercial antibiotics, ampicillin and streptomycin, and fungicides, bifonazole and ketoconazole, were used as positive controls. EtOH 30% was used as a negative control. All experiments were performed in duplicate and repeated three times.

Inhibition of Biofilm Formation
The potential of compounds to inhibit biofilm formation was investigated as previously described, with some modifications [36] C. albicans ATCC 10231 was incubated in 96-well microtiter plates with an adhesive bottom (Sarstedt, Germany), with MIC and sub-MIC concentrations of tested compounds/referent drug in YPD medium at 37 • C for 24 h. Afterwards, wells were washed thrice with sterile PBS (Phosphate buffered saline, pH 7.4) and biofilms were fixed with methanol for 20 min. Then, methanol was removed and stained with 0.1% crystal violet (Bio-Merieux, France) for 30 min. The plate was slowly washed, air dried and 96% ethanol (Zorka, Serbia) was added to dissolve bounded crystal violet. The absorbance (620 nm) was read on a Multiskan™ FC Microplate Photometer, Thermo Scientific™. Lastly, the percentage of inhibition of biofilm formation was calculated by the formula: Percentage of inhibition = ((A 620(control) − A 620sample )/A 620control )) × 100

Docking
AutoDock 4.2 ® software was used for the in silico studies and a detailed procedure is reported in our previous paper [37].

Conclusions
This work presents the synthesis of two new steroid derivatives and the study of antibacterial and antifungal activities together with previously synthesized compounds against a panel of bacterial and fungal pathogens of twelve steroid derivatives, two of which are new. The antibacterial activity of tested compounds was low to moderate with minimal inhibitory concentration being 0.37-1.5 mg/mL and minimal bactericidal being 1.5-3.0 mg/mL, except against B. cereus which was good. The antibacterial activity against resistant strains MRSA, E. coli and P. aeruginosa was superior against MRSA than ampicillin, which did not show bacteriostatic or bactericidal activity, while against the two other strains, it did not show bactericidal activity. All compounds exhibited moderate to good antifungal potency with MIC and MFC in the range of 0.37-3.00 mg/mL and 0.50-6.00 mg/mL, respectively. Compound 7 demonstrated the best activity among all tested with MIC/MFC of 0.37/0/75 mg/mL, respectively. The most sensitive fungal to compounds tested was T. viride, while C. albicans was the most resistant one. Despite this, almost all compounds except for 11 and 12 were more potent than ketoconazole against C. albicans, the deathiest fungal. Antibiofilm activity assessed for the two most potent compounds 1 and 8 in concentrations of MIC, 0.5 MIC and 0.25 MIC revealed that it was lower (33 and 15%, respectively, in concentration of MIC) than that of ketoconazole. According to docking results, it seems that the inhibition of CYP51 reductase is responsible for the antifungal activity of the compounds. All compounds showed good drug-likeness scoresin the range of 0.71-1.54. Three compounds showed two violations to the Lipinski rule.