Antimicrobial Properties of 8-Hydroxyserrulat-14-en-19-oic Acid for Treatment of Implant-Associated Infections

Abstract

Treatment options are limited for implant-associated infections (IAI) that are mainly caused by biofilm-forming staphylococci. We report here on the activity of the serrulatane compound 8-hydroxyserrulat-14-en-19-oic acid (EN4), a diterpene isolated from the Australian plant Eremophila neglecta. EN4 elicited antimicrobial activity toward various Gram-positive bacteria but not to Gram-negative bacteria. It showed a similar bactericidal effect against logarithmic-phase, stationary-phase, and adherent Staphylococcus epidermidis, as well as against methicillin-susceptible and methicillin-resistant S. aureus with MICs of 25 to 50 μg/ml and MBCs of 50 to 100 μg/ml. The bactericidal activity of EN4 was similar against S. epidermidis and its Δica mutant, which is unable to produce polysaccharide intercellular adhesin-mediated biofilm. In time-kill studies, EN4 exhibited a rapid and concentration-dependent killing of staphylococci, reducing bacterial counts by >3 log₁₀ CFU/ml within 5 min at concentrations of >50 μg/ml. Investigation of the mode of action of EN4 revealed membranolytic properties and a general inhibition of macromolecular biosynthesis, suggesting a multitarget activity. In vitro-tested cytotoxicity on eukaryotic cells was time and concentration dependent in the range of the MBCs. EN4 was then tested in a mouse tissue cage model, where it showed neither bactericidal nor cytotoxic effects, indicating an inhibition of its activity. Inhibition assays revealed that this was caused by interactions with albumin. Overall, these findings suggest that, upon structural changes, EN4 might be a promising pharmacophore for the development of new antimicrobials to treat IAI.

INTRODUCTION

Implant-associated infections (IAI) are still a cause of high morbidity and social costs despite the substantial improvement of early diagnosis and treatment. Optimal management of IAI requires surgical intervention and use of antibiotics against adherent bacteria (1). Staphylococci, including Staphylococcus aureus (both methicillin-susceptible and methicillin-resistant strains) and S. epidermidis, are the bacteria most frequently associated with IAI (2). These bacteria are able to persist on the implant surfaces, forming biofilms. Biofilms are multilayered communities of bacteria embedded in self-produced extracellular matrix characterized by an oxygen and nutrient gradient throughout its structure, inducing sessility in centrally situated cells (3, 4). The biofilm matrix is mainly composed of polysaccharide intercellular adhesin (PIA), proteins and extracellular DNA (4). PIA produced by ica operon-encoded enzymes has been implicated in the virulence and immune evasion of S. epidermidis and the pathogenesis of IAI (5).

Therefore, to successfully treat IAI, antimicrobials need to penetrate the biofilm and act independently of the bacterial physiological state. Thus far, most of the known antibiotics are dependent on the metabolic status of bacteria hindering the eradication of quiescent pathogens (6, 7). The only antibiotic with a proven activity against staphylococcal biofilm is rifampin (8). However, due to a rapid resistance development, it has to be combined with other antibiotics. This and the recent increasing emergence of drug-resistant bacteria highlight the need for new antimicrobials to combat IAI.

An especially compelling approach is the investigation of antimicrobials from natural sources. The large Australian plant genus Eremophila (Myoporaceae), of which a few species have been traditionally used by Aborigines to treat various ailments (9), is native to arid areas of Australia and produces unique secondary metabolites, among others nine classes of diterpenoids, including the most commonly occurring serrulatanes (10). Screening of organic extracts of Eremophila species revealed a selective effect against Gram-positive bacteria (11).

The aim of the present study was to evaluate the activity in vitro and in vivo and the mode of action of one of the compounds extracted from leaves of Eremophila neglecta, 8-hydroxyserrulat-14-en-19-oic acid (EN4), as a new candidate for the treatment of IAI.

MATERIALS AND METHODS

Antimicrobial agents, media, and chemical reagents.

EN4 was extracted from freshly collected E. neglecta plant material with a purity of ≥95% (as determined by nuclear magnetic resonance analysis) as described previously (12) and stored at −20°C. EN4 was dissolved in 1% dimethyl sulfoxide (DMSO) (Merck, Darmstadt, Germany) in phosphate-buffered saline (PBS) (Reagens, Basel, Switzerland) (DMSO-PBS) up to a concentration of 400 μg/ml. Nisin (Sigma-Aldrich, Buchs, Switzerland) was solubilized in 0.02 M HCl. Daptomycin (DAP; Cubicin; Novartis, Bern, Switzerland) was dissolved in 0.9% saline (Bichsel, Interlaken, Switzerland) supplemented with 50 μg of calcium ions/ml (CaCl₂). Actinomycin D (Sigma-Aldrich), ciprofloxacin (Ciproxin; Bayer, Zürich, Switzerland), vancomycin (Vancocin; Teva Pharma, Aesch, Switzerland), chloramphenicol (Applichem, Darmstadt, Germany), and chlorhexidine dihydrochloride (Sigma-Aldrich) were prepared according to the manufacturer's instructions. Bacterial media were purchased from Becton Dickinson (BD; Allschwil, Switzerland). ³H-labeled precursors of macromolecules were obtained from Hartmann Analytic (Braunschweig, Germany). Unless otherwise stated, all chemical substances were purchased from Sigma-Aldrich.

Microorganisms and growth conditions.

The strains used in experiments are listed in Table 1. Stocks of microorganisms were prepared using a cryovial bead preservation system (Microbank; Pro-Lab Diagnostics, Richmond Hill, Ontario, Canada) and stored at −75°C. Unless otherwise stated, for the preparation of overnight culture, a bead was incubated in 1 ml of tryptic soy broth (TSB) for 7 h at 37°C, diluted 1:100 in fresh TSB, and incubated overnight at 37°C. For experiments with exponential-phase bacteria, the overnight culture was diluted 1:100 and further incubated for 5 to 6 h at 37°C, followed by two washes with 0.9% saline (Bichsel, Interlaken, Switzerland), and adjusted to an appropriate level of CFU/ml. All cultures were prepared without shaking. CFU were determined by plating of aliquots of 10-fold dilutions of bacterial cultures on Mueller-Hinton agar (MHA), followed by 24 h of incubation at 37°C.

Table 1.

Microorganism strains used in this study

Species	Straina	Relevant properties	Source or reference
Staphylococcus aureus	Western Samoan Phage Pattern A (WSPPA)*	MRSA ST30	13
	ME230*	MSSA, isogenic ΔSCCmec mutant of WSPPA	M. Enderb
	SA113 (ATCC 35556)†	MSSA, production of PIA-mediated biofilm under aerobic conditions	14
Staphylococcus epidermidis	1457‡	Efficient biofilm producer	17
	1457 Δica‡	Isogenic Δica mutant, devoid of PIA-mediated biofilm	17
Mycobacterium tuberculosis	ATCC 27294 (H37Rv)§	Rifampin-sensitive strain	15
	ATCC 35838 (H37Rv-RIF-R)‡	Rifampin-resistant strain
Streptococcus pyogenes	ATCC 19615§	Laboratory strain isolated from human	16
Streptococcus pneumoniae	TIGR4 (JNR.7/87)	Encapsulated, virulent strain
Enterococcus faecalis	ATCC 19433§	Laboratory strain
Pseudomonas aeruginosa	PA01¶	Laboratory strain isolated from human
Escherichia coli	ATCC 25922	Human clinical isolate
Candida albicans	SC5314∥	Laboratory strain isolated from human
Candida glabrata	t608919‡	Clinical isolate	This study
Candida krusei	ATCC 6258‡	Reference laboratory strain isolated from human

^aSources: *, B. Berger-Bächi, Zürich, Switzerland; †, F. Götz, Tübingen, Germany; ‡, D. Mack, Swansea, United Kingdom; §, R. Frei, Basel, Switzerland; ¶, U. Jenal, Basel, Switzerland; ∥, S. Leibundgut-Landmann, Zürich, Switzerland.

^bM. Ender, ETH Zürich, Switzerland (unpublished data).

Susceptibility in vitro. (i) Gram-positive and -negative bacteria.

The minimal inhibitory and bactericidal concentrations (MIC and MBC, respectively) of EN4 for logarithmically growing bacteria were determined by using a macrodilutions method according to Clinical and Laboratory Standards Institute guidelines (18).

For staphylococci, the MBC was also assessed in the stationary growth phase (MBC_stat), which reflects the metabolic status of biofilm-embedded microorganisms (24). MBC_stat was determined using staphylococcal cultures in nutrient-limited medium consisting of 0.1 to 0.25% TSB in PBS. The use of this medium sustains bacterial counts within a range of 5 ± 0.5 log₁₀ CFU/ml for at least 24 h. MBC_stat was defined as the lowest concentration of EN4 that reduced the inocula by ≥99.9% in 24 h.

(ii) Mycobacteria.

Susceptibility of Mycobacterium tuberculosis ATCC 27294 and ATCC 35838 to EN4 was determined by microwell alamarBlue assay as previously described (19). Briefly, 0.5 McFarland inocula of M. tuberculosis were diluted 1:25 in 2-fold concentrated medium (5.9 g of Middlebrook 7H9 broth, 1.25 g of Bacto Casitone, 3.1 ml of glycerol, 100 ml of oleic acid-albumin-dextrose-catalase [OADC], 400 ml of distilled water) and mixed with solutions of EN4 (final concentrations of 50, 100, and 200 μg/ml) or 1% DMSO-PBS as untreated controls in a flat-bottom 96-well plate (BD). After 5 days of incubation at 37°C 30 μl of 0.01% (wt/vol) resazurin (Sigma, St. Louis, MO) was added to untreated M. tuberculosis, and the plate was reincubated for 24 h at 37°C. If the color turned pink (indicating growth), resazurin was added to the rest of the wells. The MIC was determined as the lowest concentration of EN4 preventing color changes of resazurin (i.e., inhibiting growth of M. tuberculosis) after 24 h of incubation at 37°C.

(iii) Candida.

The susceptibility of Candida to EN4 was determined according to CLSI guidelines (48). Five colonies of Candida were dissolved in 0.9% saline, and the turbidity adjusted to a 0.5 McFarland standard (1 × 10⁶ to 5 × 10⁶ CFU/ml). The suspension was diluted to 5 × 10² to 2.5 × 10³ CFU/ml with RPMI 1640 (Gibco, Paisley, United Kingdom) supplemented with 17.2 g of morpholinepropanesulfonic acid (MOPS) and 10 g of glucose per 500 ml (pH 7.0). Then, 0.9 ml was transferred to glass tubes containing 100 μl of 2-fold EN4 dilutions in 10% DMSO-RPMI 1640. The tubes were incubated for 24 h at 35°C without shaking. The MIC was determined as the lowest concentration of EN4 that inhibited visible fungal growth.

Time-kill studies.

Glass tubes containing TSB with EN4 at concentrations representing 0.5× to 4× the MIC were incubated with 10⁶ CFU of the test strain/ml at 37°C without shaking. Bacterial survival in the antimicrobial-free culture containing 1% DMSO-PBS in TSB served as a control. At the indicated time points, aliquots were removed and washed with 0.9% saline to avoid potential drug carryover effect. The CFU were determined by plating aliquots of appropriate dilutions on MHA. A bactericidal effect was defined as a ≥3-log₁₀ CFU/ml reduction in the initial inoculum.

Susceptibility of adherent staphylococci in vitro.

An exponential-phase culture of the test strain was diluted in TSB supplemented with 0.5% glucose (Braun Medical AG, Sempach, Switzerland) to 10⁴ CFU/ml. A total of 100 μl was seeded in flat-bottom 96-well plates. After 18 h of incubation at 37°C, the wells were washed and treated with 2-fold dilutions of EN4 for 24 h at 37°C. The activity of EN4 against biofilm was assessed by crystal violet (CV) staining as previously described (20). Briefly, planktonic bacteria were discarded, and adherent bacteria were washed. The plates were then incubated for 60 min at 60°C. Staining was performed with 100 μl of 0.5% CV solution per well for 20 min at room temperature, followed by a thorough washing under running tap water. The bound CV was extracted with 100 μl of 33% acetic acid, and the extracts were transferred to a fresh 96-well plate to be read. The optical density of the samples was measured at 590 nm using a Molecular Devices reader (Applied Biosystems, Rotkreuz, Switzerland). To determine the effect of EN4 on bacterial viability, adherent bacteria were carefully detached by pipetting up and down and then plated on MHA (21).

Macromolecular biosynthesis assay.

Exponential-phase S. aureus Western Samoan Phage Pattern A (WSPPA) was prepared in completely defined medium (CDM) for peptidoglycan, RNA, and DNA or in CDM-Leu for protein biosynthesis. CDM contained the following substances per liter: 1.77 g of Na₂HPO₄, 1.36 g of KH₂PO₄, 0.2 g of MgSO₄·7H₂O, 0.5 g of NH₄Cl, 0.5 g of NaCl, 294.1 g of sodium citrate tribasic dehydrate, 1.5 g of glucose, 160 mg each of various amino acids (i.e., l-alanine, l-valine, l-isoleucine, l-aspartic acid, l-glutamic acid, l-serine, l-threonine, l-cysteine hydrochloride, l-arginine, l-leucine, l-lysine, l-proline, l-phenylalanine, l-tryptophan, and l-histidine monohydrochloride), 1.6 g of glycine, 0.05 mg of cyanocobalamine, 0.04 mg of p-aminobenzoate, 0.01 mg of biotin, 0.1 mg of nicotinic acid, 0.1 mg of d-pantotheic acid hemicalcium salt, 0.15 mg of pyridoxine hydrochloride, 0.1 mg of thiamine hydrochloride, 0.1 mg of riboflavin, 69.5 μg of ZnCl₂, 0.1 μg of MnCl₂·4H₂O, 6 μg of BH₃O₃, 0.347 mg of CoCl₂·6H₂O, 2.6 μg of CuCl₂·2H₂O, 24 μg of NiCl₂·6H₂O, 36 μg of NaMoO₄·2H₂O, 0.15 mg of FeCl₂·4H₂O, 120 mg of NaOH, 5 mg of uracil, 5 mg of cytosine, 5 mg of adenine, and 5 mg of guanine. The concentration of l-leucine in CDM-Leu was 22.5 mg/liter instead of 160 mg/liter.

The culture (2 × 10⁷ CFU/ml) was aliquoted into prewarmed glass tubes and subsequently treated with 4× the MIC of EN4 (100 μg/ml), vancomycin (8 μg/ml), actinomycin D (25 μg/ml), ciprofloxacin (8 μg/ml), chloramphenicol (32 μg/ml), or chlorhexidine (CHX; 3.12 μg/ml), followed by the immediate addition of 0.1 μCi [³H]N-acetylglucosamine/ml, 1 μCi of [³H]uridine/ml, 1 μCi of [³H]thymidine/ml, or 3 μCi of [³H]leucine/ml to investigate the biosynthesis of peptidoglycan, RNA, DNA, and proteins, respectively. Untreated control samples were exposed to 1% DMSO-PBS. After 1 h of incubation at 37°C, 0.5-ml portions of the bacterial suspensions were transferred to ultracentrifuge tubes and precipitated on ice with 1 ml of ice-cold 10% trichloroacetic acid (TCA) for at least 1.5 h. After this time, free fractions of precursors were removed from the precipitates by washing with 5% TCA–1.5 M NaCl and with 5% TCA. Upon solubilization with 0.1% sodium dodecyl sulfate (SDS)–0.1 M NaOH, the precipitates were transferred to scintillation tubes (Perkin-Elmer, Groningen, Netherlands) and thoroughly mixed with 2 ml of scintillation cocktail (Ultima Gold; Perkin-Elmer, Waltham, MA). Radioactivity reflecting the amount of incorporated precursors was measured using a Tri-CARB 1900TR liquid scintillation analyzer (Packard, Meriden, CT). Incorporation was measured in counts per minute (cpm) and expressed as a percentage of the untreated control.

Flow cytometric analysis.

Exponential-phase S. epidermidis 1457 (5 × 10⁷ CFU) was incubated for 10 min at 37°C with 100 μg of EN4 or CHX/ml. Upon centrifugation, the samples were stained with 0.75 μM propidium iodide and 0.0625 μM SYTO9 from the Live/Dead BacLight bacterial viability kit (Invitrogen, Oregon). After 15 min of incubation at room temperature protected from light, the green and red fluorescence intensities of 10⁵ stained bacteria per sample were recorded using a CyAn ADP flow cytometer (Dako Cytomation, Glostrup, Denmark) and analyzed with FlowJo software (TreeStar, Stanford, CA).

ATP leakage assay.

Exponential-phase WSPPA (7 × 10⁸ CFU) was exposed for 10 min at 37°C to EN4 (50, 100, and 200 μg/ml), CHX (50 and 100 μg/ml), nisin (32 μg/ml), ciprofloxacin (8 μg/ml), or 1% DMSO-PBS (untreated control). In order to assess ATP release, the bacteria were harvested by centrifugation, and 100 μl of each supernatant was transferred to flat-bottom black-walled 96-well plate (Perkin-Elmer, Groningen, Netherlands). ATP content was determined using the BacTiter-Glo assay (Promega, Madison, WI). Luminescence was measured for 30 min with SpectraMAX GeminiXS (Molecular Devices/Bucher Biotec, Basel, Switzerland). The results are shown as the area under the curve.

Transmission electron microscopy.

Exponential-phase WSPPA (5 × 10⁹ CFU) was incubated for 1 h at 37°C with EN4 (100 and 200 μg/ml), CHX (250 μg/ml), or 1% DMSO-PBS (untreated control) and harvested by centrifugation. The pellets were resuspended in 2% glutaraldehyde in PBS, followed by 2 h of fixation at room temperature. Postfixation was carried out with reduced osmium tetroxide (1.5% KFeCN and 1% OsO₄ in PBS) for 40 min, followed by a second postfixation with 1% OsO₄ for 40 min. After a washing step, the bacteria were embedded in 2% agarose and cut in blocks, which were dehydrated using 50, 70, 90, and 100% ethanol (10 min each). The samples were further incubated in acetone (10 min). Infiltration with Epon-acetone (50:50) for 1 h was followed by incubation in pure Epon for 3 to 4 h. The samples were embedded in fresh Epon and polymerized for 48 h at 60°C. Sections (60 nm) were cut at Ultracut E (Leica), stained with 6% uranyl acetate, and analyzed using a Morgagni transmission electron microscope (FEI).

EDTA cotreatment.

The effect of EDTA on the susceptibility of Gram-negative bacteria to EN4 and nisin was investigated as previously described (22). Briefly, an overnight culture of P. aeruginosa PA01 was prepared in brain heart infusion (BHI) broth, diluted 1:100 in fresh BHI medium, and incubated at 37°C to a population density of 5 × 10⁷ CFU/ml. Bacteria (5 × 10⁵ CFU) were transferred to microcentrifuge tubes containing buffer (50 mM Tris-HCl [pH 7.2]) supplemented with (i) 1 mM EDTA, (ii) 5 mM EDTA, (iii) 100 μg of nisin/ml, (iv) 100 μg of nisin/ml and 1 mM EDTA, (v) 200 μg of EN4/ml, or (vi) 200 μg of EN4/ml and 5 mM EDTA. The samples were incubated for 1 h at 37°C. The numbers of surviving bacteria were determined by plating aliquots on MHA.

Cytotoxicity on eukaryotic cells.

The cytotoxicity assay was prepared as previously described (20). Briefly, mouse L929 fibroblasts were cultured in complete RPMI 1640 supplemented with 5% heat-inactivated fetal bovine serum (FBS). To assess the cytotoxicity of EN4 over time, 10⁴ cells/well were seeded in a flat-bottom 96-well plate in RPMI 1640 supplemented with 3.75% FBS, followed by incubation for 4 h until the fibroblasts adhered. After this time, EN4 solutions in 3.75% FBS-RPMI were transferred to wells to final concentrations of 25, 50, and 100 μg/ml. Fibroblasts incubated with 1% DMSO-PBS instead of EN4 served as an untreated control. At the indicated time points, the activity of the lactate dehydrogenase (LDH) was assessed in cell supernatants and corresponding lysates (total LDH) using the CytoTox 96 nonradioactive cytotoxicity assay (Promega) according to the manufacturer's instructions. Cytotoxicity was defined as LDH release and was expressed as a percentage of the total LDH.

For determination of the influence of FBS on the cytotoxicity of EN4, fibroblasts were seeded and treated with EN4 at 100 μg/ml in RPMI supplemented with 3.75, 5, or 10% FBS, and then the cytotoxicity was assessed as described above.

Activity in vivo.

The in vivo activity of EN4 against S. aureus SA113 was investigated in a mouse model of foreign body infection as described previously (21) with the approval of the Kantonale Veterinaeramt Basel-Stadt (permit 1710). Experiments were conducted according to the regulations of Swiss veterinary law. Female C57BL/6 mice (Harlan Laboratories, Switzerland), 9 to 11 weeks old, kept under specific-pathogen-free conditions in the animal house of the Department of Biomedicine, University Hospital Basel, were anesthetized via intraperitoneal injection of 65 mg of ketamine (Ketalar; Pfizer, Zürich, Switzerland)/kg and 13 mg of xylazinum (Xylasol; Graeub, Bern, Switzerland)/kg, followed by the subcutaneous implantation of sterile Teflon tissue cages (Angst + Pfister AG, Zürich, Switzerland). After surgery, the mice were treated with 0.05 mg of buprenorphine (Temgesic; Essex Chemie, Lucerne, Switzerland)/kg to treat postoperative pain. Upon complete wound healing (2 weeks), the mice were anesthetized with isoflurane (Isofuran; Abbott, Wiesbaden, Germany), and the cages were tested for sterility by plating percutaneously aspirated tissue cage fluid (TCF) on Columbia sheep blood agar plates. To simulate a perioperative infection, the cages were injected with 100 and 250 μg of EN4 (10 mice per group) or 30 μg of DAP (5 mice per group) and then immediately infected with 4 × 10³ CFU of SA113. Untreated control cages were injected with 1% DMSO-PBS (8 mice per group). After 24 and 48 h, the TCF was collected into ultracentrifuge tubes containing 10 μl of 1.5% EDTA in 0.45% saline (pH 7.3) to avoid clotting, and the cages were reinjected with the respective substances. TCF was used to determine the numbers of planktonic SA113 by plating appropriate dilutions on blood agar plates and to assess the viability of leukocytes by trypan blue staining. Aspiration of TCF was repeated after 72 h, and the mice were sacrificed. Tissue cages were explanted under aseptic conditions, washed twice with 0.9% saline in order to remove planktonic bacteria, and cultured in TSB. The cages were vortexed at 0, 24, and 48 h to increase the possible regrowth of adherent bacteria. After 48 h of incubation at 37°C, the bacterial presence in the cultures was determined by plating on blood agar plates. Detection of the growth of SA113 was defined as treatment failure. The efficacy of treatment against adherent bacteria was expressed as the cure rate, defined as the percentage of cages without growth in the individual treatment group.

Inhibition of EN4 activity.

An overnight culture of SA113 was transferred to glass tubes (2 × 10⁶ CFU/ml) containing TSB supplemented with EN4 (100 μg/ml) in the presence of mouse TCF (2.5, 5, and 75%), pooled normal human serum (NHS; 1.25, 2.5, and 75%), human serum albumin (HSA; Blutspendedienst SRK, Bern, Switzerland) (0.125 and 0.25%), or 300 μg of human fibrinogen (Hyphen BioMed, Allschwil, Switzerland)/ml (representing 10% of the normal blood level). The samples were incubated for 6 h at 37°C. Then, 0.5-ml portions of the suspensions were transferred to microcentrifuge tubes and washed once with 0.9% saline, and the numbers of surviving bacteria were determined by plating. The results were calculated as the log₁₀ CFU/ml reduction from the initial inoculum. None of the supplements influenced the viability of SA113 in the absence of EN4 (data not shown).

Statistical analysis.

Data were analyzed by GraphPad Prism 5.0a program (GraphPad Software). The Mann-Whitney U test or one-way analysis of variance (ANOVA) were used to determine the statistical significance (P) of differences in the in vitro assays. The statistical tests used for each experiment are specified in the figure legends.

RESULTS

EN4 acts in a rapid and dose-dependent manner against Gram-positive bacteria in vitro.

The MICs and MBCs of EN4 were determined for different strains of Gram-positive and Gram-negative bacteria, as well as for M. tuberculosis and Candida species. The results are summarized in Table 2 and Table 3. Both Streptococcus pyogenes and Streptococcus pneumoniae showed the highest susceptibility to EN4, with MICs of 6.25 and 1.6 μg/ml and MBCs of 12.5 and 3.12 μg/ml, respectively. The MIC and MBC of EN4 for Enterococcus faecalis were both 50 μg/ml. As previously reported (12), Gram-negative bacteria were not susceptible to EN4 at concentrations up to 200 μg/ml. Interestingly, EN4 was active against rifampin-susceptible and -resistant M. tuberculosis with an MIC of 100 μg/ml. C. albicans was not influenced by EN4, whereas C. glabrata and C. krusei were susceptible with an EN4 MIC of 200 μg/ml, the highest concentration used. EN4 showed a bactericidal effect against various logarithmic-phase staphylococci. The MIC_log values for S. epidermidis 1457 and its isogenic, biofilm-deficient Δica mutant were 25 and 50 μg/ml, respectively, and the MBC_log was 100 μg/ml. EN4 was similarly active against MRSA strain WSPPA and its isogenic, methicillin-susceptible mutant ME230 with an MIC_log of 25 μg/ml and an MBC_log of 50 μg/ml. SA113 capable of producing PIA-mediated biofilm under aerobic conditions was susceptible to EN4, with an MIC_log and an MBC_log of 25 and 50 μg/ml, respectively.

Table 2.

In vitro susceptibility of Gram-positive and Gram-negative bacteria and fungi to EN4 determined as MICs and MBCs

Species	Strain	Concn (μg/ml)
		MIC	MBC
Streptococcus pyogenes	ATCC 19615	6.25	12.5
Streptococcus pneumoniae	TIGR4 (ATCC BAA-334)	1.6	3.12
Enterococcus faecalis	ATCC 19433	50	50
Pseudomonas aeruginosa	PA01	>200	>200
Escherichia coli	ATCC 25922	>200	>200
Mycobacterium tuberculosis	ATCC 27294	100	–a
	ATCC 35838	100
Candida albicans	SC5314	>200
Candida glabrata	t608919	200
Candida krusei	ATCC 6258	200

^aSusceptibility was determined only as the MIC.

Table 3.

In vitro susceptibility to EN4 of staphylococci in the logarithmic (MIC_log and MBC_log) and stationary (MBC_stat) growth phases

Species	Strain	Concn (μg/ml)
		MIClog	MBClog	MBCstat
Staphylococcus aureus	MRSA WSPPA	25	50	100
	MSSA ME230	25	50	100
	MSSA SA113	25	50	50
Staphylococcus epidermidis	1457	25	100	100
	1457 Δica	50	100	100

To investigate whether EN4 could be used to treat IAI, further experiments were focused selectively on staphylococci. In time-kill studies EN4 showed a dose-dependent activity against WSPPA with a bactericidal effect (≥3-log₁₀ CFU/ml reduction) within 5 min at MBC_log (Fig. 1). Similar kinetics was observed for other staphylococci (data not shown). This is in line with previous findings indicating that Eremophila duttoni extract eradicated MRSA within 1 h (23).

Fig 1.

Time-kill studies of WSPPA exposed to titrated concentrations of EN4. Dotted line represents a bactericidal effect, i.e., a 3-log₁₀ CFU/ml reduction from the initial inoculum. The values shown are means of three independent experiments ± the standard deviations (SDs).

Taken together, these results show that EN4 acts specifically against a range of Gram-positive bacteria and M. tuberculosis with a rapid and dose-dependent kinetics for staphylococci.

EN4 acts against staphylococci independently of growth phase and PIA-mediated biofilm.

IAI are mainly caused by staphylococci growing in biofilms. Therefore, the effect of EN4 on staphylococci in the stationary growth phase, which is believed to reflect the status of biofilm-embedded bacteria (24), was studied. Interestingly, the transition from the logarithmic to the stationary growth phase did not decrease bacterial susceptibility to EN4, as evidenced by comparable MBC values for logarithmically (MBC_log) and stationary (MBC_stat) growing staphylococci (Table 3). Next, the impact of EN4 on biofilm was investigated using CV staining. EN4 at concentrations of ≥25 μg/ml significantly decreased an abundant biofilm of S. epidermidis 1457 (Fig. 2). The S. epidermidis 1457 Δica mutant deficient in PIA used as a control gave an overall low CV signal that was not influenced by EN4 (Fig. 2). Since CV does not distinguish between live and dead bacteria due to its affinity to the negatively charged molecules present both at the bacterial surface and in the biofilm matrix (25), the viability of adherent staphylococci was additionally assessed. Of note, the bactericidal activity of EN4 on S. epidermidis 1457 and its Δica mutant was similar. Treatment with EN4 at ≥100 μg/ml resulted in 3-log₁₀ CFU/ml reduction of adherent bacteria for both strains (Fig. 2B). Similar results were obtained for other staphylococci (WSPPA, ME230, and SA113; data not shown).

Fig 2.

Effect of EN4 on adherent S. epidermidis 1457 and its Δica mutant. Inocula of both strains (10³ CFU/well in 96-well plates) were incubated for 18 h at 37°C, followed by the addition of EN4 at titrated concentrations. After 24 h of exposure, the influence of EN4 on the biofilm and the viability of staphylococci was determined by CV staining (A) and plating of detached bacteria (B), respectively. The values shown are means of three independent experiments prepared in triplicates ± the SDs. Significant CV signal reduction for S. epidermidis 1457 compared to results for the untreated control is indicated (*, P < 0.05 [Mann-Whitney U test]). In panel B, dotted lines represent a 3-log₁₀ CFU/ml reduction in S. epidermidis 1457 (⧫) and Δica (♢) from untreated samples.

These results show that EN4 is similarly active against staphylococci in the logarithmic and stationary growth phases, as well as embedded in PIA-mediated biofilm, making it promising for the treatment of IAI.

EN4 inhibits synthesis of DNA, RNA, protein, and peptidoglycan.

To understand the mechanism of action of EN4, macromolecular biosynthesis assays were performed. The incorporation of ³H-labeled precursors of peptidoglycan, RNA, DNA, and proteins by WSPPA was assessed in the presence of EN4, control antibiotics, and the antiseptic CHX at 4× the MIC. The control antibiotics vancomycin, actinomycin D, ciprofloxacin, and chloramphenicol reduced biosynthesis of their specific targets, peptidoglycan, RNA, DNA, and proteins, respectively (Fig. 3). Moreover, actinomycin D consequently reduced also the production of DNA and proteins. In contrast, EN4 inhibited all biosynthetic pathways. This was comparable to the mode of action of CHX and points toward a multitarget antiseptic rather than an antibiotic mechanism of action of EN4.

Fig 3.

Inhibition of biosynthesis of macromolecules by EN4. The incorporation of [³H]N-acetylglucosamine (A), [³H]uridine (B), [³H]thymidine (C), and [³H]leucine (D) by WSPPA treated for 1 h with EN4, vancomycin (VAN), actinomycin D (ActD), ciprofloxacin (CIP), chloramphenicol (CHL), or chlorhexidine (CHX) at 4× the MIC was expressed as the percentage of untreated control (peptidoglycan, 3,444 ± 1,212 cpm; RNA, 115,538 ± 19,533 cpm; DNA, 13,862 ± 762 cpm; protein, 7,065 ± 323 cpm). The values shown are means of at least two independent experiments prepared in duplicates ± the SDs. Dotted lines represent 100% incorporation. Significant reduction of biosynthesis compared to results for the untreated control is indicated (*, P < 0.05; **, P < 0.01; ***, P < 0.001 as determined by one-way ANOVA [Kruskal-Wallis test]) with a Dunn post test).

EN4 affects bacterial membrane integrity causing ATP leakage.

To further reveal the mechanism of action of EN4, its effect on bacterial membrane integrity was studied. Flow cytometric analysis with propidium iodide and SYTO9 staining showed that the membrane integrity of S. epidermidis 1457 was similarly affected by treatment with bactericidal concentrations of EN4 and CHX (Fig. 4A). This is in agreement with a previous report on membranolytic properties of the extract of Eremophila duttoni leaves, which also contain active serrulatanes (26). Exposure to EN4 and CHX also led to ultrastructural changes of WSPPA, as shown by transmission electron microscopy. Membrane invagination and mesosome-like structures were observed after EN4 treatment; such features were previously attributed to membranolytic activity (27) (Fig. 4B). CHX mostly affected the bacterial cytoplasm causing formation of multiple intracellular granules, which possibly mirrored precipitation of cytoplasmic molecules (28).

Fig 4.

Effect of EN4 on bacterial membrane integrity. (A) S. epidermidis 1457 was incubated for 10 min with EN4 or CHX at 100 μg/ml or 1% DMSO-PBS and subsequently double stained with propidium iodide (PI) and SYTO9 and analyzed by flow cytometry. The results of one representative experiment of three performed are shown. (B) Transmission electron microscopy images of ultrathin sections of WSPPA treated for 1 h with 1% DMSO-PBS as a control, EN4 at 100 or 200 μg/ml, or 250 μg of CHX/ml. Arrowheads indicate ultrastructural changes; the bars represent 200 nm. (C) WSPPA was treated for 10 min with EN4 or CHX at 100 μg/ml, followed by centrifugation and investigation of supernatants for the presence of ATP using luciferase reaction. Nisin (NIS) and ciprofloxacin (CIP) were used as a positive and a negative control, respectively, and 1% DMSO-PBS was used as an untreated control (Ctrl). The values shown are means ± the SDs of areas under the concentration-time curve calculated for the first 30 min of luciferase reaction from at least three independent experiments prepared in duplicates. Significant ATP leakage compared to results for the untreated control is indicated (*, P < 0.05; **, P < 0.01 as determined by one-way ANOVA [Kruskal-Wallis test] with a Dunn post test).

To elucidate the physiological consequences of membrane disintegration, ATP leakage from WSPPA was measured. The results indicate that the membrane disruption by EN4 is accompanied by a concentration-dependent ATP leakage (Fig. 4C). The release of ATP was also observed for CHX and nisin, known for their membrane activity, but not for ciprofloxacin, in accordance with its intracellular mode of action on DNA gyrase and topoisomerase IV (29). Hence, EN4 was found to affect membrane integrity and to cause ATP leakage.

The tolerance of Gram-negative bacteria to EN4 can be reversed by destabilization of the outer membrane.

Since the findings thus far point toward a multitarget mode of action, we wondered why EN4 selectively acts against Gram-positive bacteria. One of the causes of higher antibiotic tolerance of Gram-negative bacteria is the presence of the outer membrane, which poses an efficient permeability barrier. This barrier, however, can be breached by EDTA (30). EDTA chelates Ca²⁺ and Mg²⁺ that are involved in noncovalent binding of lipopolysaccharide molecules, and treatment with EDTA results in an increased permeability of the outer membrane (30). The induced susceptibility of Gram-negative bacteria to antibiotic/EDTA cotreatment has been reported for many antimicrobials, such as nisin (22). Therefore, to assess whether the outer membrane of Gram-negative bacteria is responsible for the lack of susceptibility to EN4, P. aeruginosa PA01 was exposed to EN4 and EDTA. Nisin was used as a control (Fig. 5). EDTA induced the activity of EN4 against P. aeruginosa PA01 up to a bactericidal level. Similar results were obtained for nisin. This effect was observed to a lesser extent for E. coli ATCC 25922 (data not shown), which can be explained by the fact that the impermeability of the E. coli membrane is less dependent on divalent cations (31). Thus, the tolerance of Gram-negative bacteria to EN4 is mediated by the outer membrane and can be reversed in the presence of EDTA.

Fig 5.

Effect of EN4-EDTA cotreatment on P. aeruginosa PA01. P. aeruginosa PA01 (10⁶ CFU/ml) was incubated for 1 h in the presence of 200 μg of EN4/ml with or without 5 mM EDTA. Bacteria treated with 100 μg of nisin/ml with or without 1 mM EDTA served as control. The results were calculated as the log₁₀ CFU/ml differences between initial and final numbers of bacteria. The dotted line represents a 3-log₁₀ CFU/ml reduction from the initial inoculum. The values shown are means of three independent experiments ± the SDs.

EN4 induces the cytotoxicity of eukaryotic cells in vitro in a dose- and time-dependent manner.

The cytotoxicity of an antimicrobial compound indicating the adverse potential is important if the compound is to be used to prevent or treat IAI. Therefore, we quantified the cytotoxicity of EN4 on eukaryotic cells by LDH release. EN4 at 25 μg/ml corresponding to MICs of S. aureus did not affect the viability of mouse L929 fibroblasts (Fig. 6). Cytotoxicity was observed within the range of MBCs and was time dependent. After 8 h of exposure to EN4 at 50 μg/ml, the cytotoxicity reached 50%, whereas 1 h to 2 h of contact with EN4 at 100 μg/ml had already resulted in a cytotoxicity of ca. 100%. These results indicate that mouse fibroblasts and staphylococci are similarly susceptible to EN4 and that EN4 therefore has a small therapeutic window in vitro.

Fig 6.

Cytotoxicity of EN4 on mouse L929 fibroblasts. L929 cells were treated with EN4 at antistaphylococcal concentrations, and the cytotoxic effect was assessed over time by measuring the released lactate dehydrogenase (LDH). The results are expressed as the percentage of total LDH obtained from completely lysed cells. The values shown are means of at least three independent experiments prepared in triplicates ± the SDs. The dotted lines represent inhibitory concentrations of EN4 causing the death of 50 or 90% of cells (IC₅₀ and IC₉₀, respectively).

EN4 exhibits neither antimicrobial nor cytotoxic activity in vivo.

Since EN4 showed favorable in vitro activity against biofilm-forming staphylococci but concomitantly was cytotoxic on eukaryotic cells, we performed in vivo studies to test its efficacy in a previously established tissue cage model in mice. Tissue cages were infected with 4 × 10³ CFU of SA113, an inoculum mimicking a perioperative infection. Injection of 4× and 10× the MICs of EN4 or 30 μg of DAP per cage was followed by an immediate infection to check whether EN4 is able to prevent bacterial adhesion and consequently clears the infection. EN4 did not decrease the number of planktonic bacteria in the TCF (Fig. 7A), whereas 30 μg of DAP per cage reduced the number of planktonic bacteria by ∼2 log₁₀ CFU/ml, as previously reported (32). Reinjection of EN4 at days 1 and 2 did not improve its antimicrobial effect, as opposed to DAP that cleared planktonic bacteria in 80 and 60% cages at days 2 and 3, respectively. The cure rate that corresponds to the regrowth of adherent SA113 from explanted cages was 0% for all cages (data not shown). This indicates a failure of EN4 and also, as previously described (32), of DAP in the eradication of adherent bacteria in vivo. In addition, the viability of leukocytes present in TCF was not affected by EN4 (Fig. 7B). In summary, despite its in vitro satisfactory activity, EN4 did not exhibit any antibacterial or cytotoxic properties in vivo, indicating an inhibition of its activity.

Fig 7.

In vivo activity of EN4 in a mouse tissue cage model. The numbers of planktonic SA113 (A) and the viability of leukocytes (B) present in tissue cage fluid from cages treated with 100 or 250 μg of EN4 or 30 μg of daptomycin (DAP) and infected with 4 × 10³ CFU of SA113 were investigated. Open circles indicate control mice injected with 1% DMSO-PBS. The values shown are means ± the SDs.

The activity of EN4 is inhibited by interaction with albumin.

Pharmacokinetic and pharmacodynamic studies revealed that the in vivo efficacy of drugs strongly depends on their interactions with plasma proteins and lipids. These interactions allow a prolonged activity but may also decrease the treatment efficacy (33). Therefore, to elucidate whether interactions with proteins play a role in the inhibition of EN4 observed in vivo, the survival of SA113 was determined in the presence of EN4 at 100 μg/ml and various physiological fluids and proteins. TCF and NHS decreased the antibacterial activity of EN4 in a concentration-dependent manner. Similar levels of inhibition were observed for 5% TCF and for 2.5% NHS, in agreement with previous findings that TCF contains around half of the amount of proteins present in serum (34). Both TCF and NHS at 75% resulted in complete inhibition of EN4 (Fig. 8A). Furthermore, a similar effect was observed for HSA at 0.25%, but not for fibrinogen, indicating that binding to albumin is responsible for the inhibition of EN4. Under the same conditions, ciprofloxacin, which is bound by plasma proteins up to 28% (35), was not inhibited, and at a concentration of 4× the MIC reduced the number of SA113 by 3 log₁₀ CFU/ml (data not shown).

Fig 8.

Inhibition of activity of EN4 by interaction with albumin. (A) Survival of SA113 (2 × 10⁶ CFU/ml) incubated for 6 h in the presence (gray bars) and absence (white bar) of EN4 at 100 μg/ml in tryptic soy broth (TSB) supplemented with increasing concentrations of tissue cage fluid (TCF), normal human serum (NHS), human serum albumin (HSA), and human fibrinogen (hFib). The results were calculated as log₁₀ CFU/ml differences between initial and final numbers of SA113. The positive values depict growth; the negative values depict killing. Means ± the SDs of three independent experiments are shown. The dotted line represents a 3-log₁₀ CFU/ml reduction from the initial inoculum. (B) Cytotoxic effect of EN4 at 100 μg/ml on L929 in the presence of 3.75, 5, or 10% FBS investigated over time as described above. The means ± the SDs of three independent experiments prepared in triplicates are shown. The dotted lines represent the inhibitory concentrations of EN4 causing the death of 50 or 90% of cells (IC₅₀ and IC₉₀, respectively).

Consequently, to assess whether the eukaryotic cytotoxicity of EN4, which is observed in vitro but not in vivo, is influenced by albumin, mouse L929 fibroblasts were exposed to EN4 at 100 μg/ml in the presence of titrated concentrations of FBS, and the cytotoxic effect was measured. Whereas EN4 at 100 μg/ml in the presence of 3.75% FBS, which was used in the previous cytotoxicity experiment, induced rapid cell death, this effect was completely abolished by 10% FBS (Fig. 8B). These findings indicate that the activity of EN4 is inhibited in vivo by interactions with albumin.

DISCUSSION

Over the past several decades, antibiotic pressure has induced selection for resistant bacteria to every clinically used antibiotic. In the search for new antibacterial drugs, natural products derived from traditional medicinal plants are currently used as scaffolds for drug development.

The aim of the present study was to investigate EN4, a new antimicrobial compound extracted from Eremophila neglecta. EN4 belongs to a class of derivatives of diterpenes, one of the largest plant-derived families of secondary metabolites with antistaphylococcal activity (36). Due to this activity, the underlying premise of the present study was the potential use of EN4 in the prophylaxis or treatment of IAI. This is, to our knowledge, the first detailed report on an antistaphylococcal activity and the mode of action of an active compound purified from an extract from Eremophila plant species.

The activity of EN4 was investigated primarily against staphylococci, since these organisms are the main cause of IAI (2). EN4 showed a rapid, dose-dependent bactericidal effect on logarithmically growing staphylococci regardless of methicillin resistance. Furthermore, its activity was not affected by transition of staphylococci to the stationary phase. This is of great advantage for the treatment of IAI since agents bactericidal for logarithmically growing pathogens may fail to act on slow-growing or nongrowing bacterial populations (37).

Chronic infections with biofilms, such as IAI, pose a challenge for antimicrobials. To date, the only antibiotic with proven in vivo activity against staphylococcal biofilm is rifampin, which due to a rapid resistance development has to be used in combination therapy for IAI (8). In this context, EN4 reduced biofilm of S. epidermidis 1457 by ≥3 log₁₀ CFU/ml in vitro, irrespective of PIA. The mechanism of its antibiofilm activity might be an efficient penetration of the biofilm matrix, which allows accessing and exerting growth-phase-independent properties against sessile bacteria. This is in contrast to DAP. Its activity is decreased against stationary-phase S. aureus and biofilm in vivo and in vitro (21, 32) despite its ability to penetrate the biofilm structure (38). Recent results on antibiotics such as telavancin (39) or oritavancin (40) have demonstrated activity in vitro against stationary-phase bacteria or biofilms; however, they still require in vivo evaluation. These in vitro properties make EN4 a promising compound despite its elevated MIC and MBC values that are often found for plant-derived products. Indeed, it has been suggested, due to the lack of antimicrobials active against nongrowing bacteria, to prioritize such compounds even if the MICs are high (41).

The studies on the mode of action revealed that EN4 affects bacterial membrane integrity, resulting in a rapid leakage of ATP. This finding can in part explain its efficacy against stationary and biofilm bacteria, since the membrane is essential for both metabolically active and inactive microorganisms. Many antibiotics target bacterial molecules that are accessible only in metabolically active bacteria, hindering the eradication of quiescent pathogens (6, 7). Therefore, the importance of membrane-active compounds in the eradication of persistent organisms has recently gained more attention (6). Furthermore, the membranolytic mode of action may prevent the development of resistance (41). This was confirmed for EN4 by several rounds of passages of SA113 in subinhibitory concentrations of EN4, where no decrease in susceptibility was observed (data not shown).

The membranolytic multitarget activity is also one of the attributes of antiseptics. In fact, EN4 exhibits properties similar to CHX, affecting bacterial membrane integrity and macromolecular biosynthesis, and thereby suggesting an antiseptic-like character. Both EN4 and CHX (42) are active against Gram-positive bacteria. EN4 also acts on mycobacteria but, in contrast to CHX, not against Gram-negative bacteria. Gram-negative bacteria are inherently less susceptible to antimicrobials due to the architecture of the cell wall (30). EN4 acts on Gram-negative bacteria only upon destabilization of the outer membrane, which further confirms that the membrane is a target of EN4.

The activity of some antiseptics such as CHX is decreased in the presence of serum (43). EN4 did not influence SA113 in the mouse tissue cage model due to inhibition by albumin. Interestingly, another diterpene compound has also failed in vivo (44). However, this finding has not been further investigated. The tendency for protein binding and poor tissue distribution are recognized as an issue of membrane-active agents due to their general lipophilic character (6). This is well exemplified by daptomycin, which undergoes organ-specific inhibition by pulmonary surfactants (45). Furthermore, in contrast to other membrane-active antibiotics such as daptomycin, oritavancin (46), or telavancin (47), the cost of membranolytic activity of EN4 is cytotoxicity on eukaryotic cells, which was also inhibited by serum, suggesting that these properties may be located at similar sites of the chemical structure of EN4. Nevertheless, chemical synthesis and structural modifications will be required to decrease its impairment by albumin and reduce cytotoxicity, while retaining antimicrobial properties.

In summary, the present study is a comprehensive investigation of the activity and mode of action of a plant-derived antimicrobial that impartially evaluates its possible future applications. Due to the good in vitro effect against stationary growing and biofilm-embedded staphylococci and at the same time poor in vivo activity, EN4 could serve as a fundamentally new pharmacophore scaffold for the development of a novel class of membrane-active compounds to treat biofilm-mediated infections.