Phytomedicine 22 (2015) 223–230
Contents lists available at ScienceDirect
Phytomedicine journal homepage: www.elsevier.com/locate/phymed
Interaction between lichen secondary metabolites and antibiotics against clinical isolates methicillin-resistant Staphylococcus aureus strains Pierangelo Bellio a, Bernardetta Segatore a, Alisia Mancini a, Letizia Di Pietro a, Carlo Bottoni a, Alessia Sabatini a, Fabrizia Brisdelli a, Marisa Piovano b, Marcello Nicoletti c, Gianfranco Amicosante a, Mariagrazia Perilli a, Giuseppe Celenza a,∗ a
Department of Biotechnological and Applied Clinical Sciences, University of l’Aquila, L’Aquila, Italy Department of Chemistry, Universidad Técnica F. Santa María, Casilla 110 V, Valparaíso, 6, Chile c Department of Enviromental Biology, University Sapienza, Rome, Italy b
a r t i c l e
i n f o
Article history: Received 16 June 2014 Revised 6 October 2014 Accepted 14 December 2014
Keywords: Lichen secondary metabolites MRSA Checkerboard assay Antimicrobial activity
a b s t r a c t The in vitro antimicrobial activities of five compounds isolated from lichens, collected in several Southern regions of Chile (including the Chilean Antarctic Territory), were evaluated alone and in combination with five therapeutically available antibiotics, using checkerboard microdilution assay against methicillin-resistant clinical isolates strains of Staphylococcus aureus. MIC90 , MIC50 , as well as MBC90 and MBC50 , for the lichen compounds were evaluated. The MIC90 was ranging from 32 μg/ml for perlatolic acid to 128 μg/ml for α -collatolic acid. MBC90 was ranging from onefold up to twofold the MIC90 for each compound. A synergistic action was observed in combination with gentamicin, whilst antagonism was observed for some lichen compounds in combination with levofloxacin. All combinations with erythromycin were indifferent, whilst variability was observed for clindamycin and oxacillin combinations. Data from checkerboard assay were analysed and interpreted using the fractional inhibitory concentration index and the response surface approach using the E model. Discrepancies were found between both methods for some combinations. These could mainly be explained by the failure of FIC approach, being too much subjective and sensitive to experimental errors. These findings suggest, however, that the natural compounds from lichens are good candidates for the individuation of novel templates for the development of new antimicrobial agents or combinations of drugs for chemotherapy. © 2015 Elsevier GmbH. All rights reserved.
Introduction Lichens are symbiotic organisms derived by the close cellular union of a fungal (mycobiont) and an algal and/or cyanobacterial (phytobiont) partner, comprising about 20,000 known species. They produce a variety of secondary compounds that typically arise from the fungal component secondary metabolism, many of them, exclusive of the lichen production. Chemotaxonomic studies have shown that the most unique lichen metabolites belong to the chemical classes of depsides, depsidones and dibenzofurans. The almost 800 known lichen secondary metabolites can be classified according to the classic biosynthetic pathways: the poliketidic path (monocyclic phenols, depsides, depsidones, depsones, dibenzofurans, xanthones, naphtaquinones anthraquinones, macrocyclic lactones, aliphatic acids, etc.), the mevalonic acid path (steroids, carotenoids, ∗ Corresponding author at: Department of Biotechnological and Applied Clinical Sciences, University of l’Aquila, Via Vetoio, 1, 67100 l’Aquila, Italy. Tel.: +39 0862433444. E-mail address: [email protected] (G. Celenza).
http://dx.doi.org/10.1016/j.phymed.2014.12.005 0944-7113/© 2015 Elsevier GmbH. All rights reserved.
etc.) and the shikimic acid path (amino acid derivatives, cyclopeptides, etc.) (Huneck 1999). Although medicinal plants have been used for centuries as sources of therapeutic agents worldwide, they cannot be classified as pure and efficient antimicrobial agents. However, in spite of the fact that plant-derived antibacterial compounds show a general low degree of activity, most plants, indeed, are successful in fighting infections (Hemaiswarya et al. 2008). Plants, in fact adopt “synergy” as their peculiar different paradigm to fight pathogenic microorganisms. Several studies have demonstrated that a number of natural products, which failed as antimicrobials, are able to dramatically increase the effectiveness of chemotherapeutic agents against Gram-negative and Gram-positive bacteria (Gibbons and Udo 2000; Tegos et al. 2002; Stavri et al. 2007; Hemaiswarya et al. 2008; Celenza et al. 2012; Segatore et al. 2012). Antimicrobial resistance has emerged among pathogenic bacteria since the beginning of the antibiotic era. Resistance potentially extends to the entire repertoire of available therapeutic agents. Nowadays, bacteria expressing multidrug resistant (MDR), extensively drug
224
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
resistant (XDR) and pandrug resistant (PDR) phenotypes are amongst the most important cause of infections in nosocomial and community settings and new drugs are urgently needed. As a result of its intrinsic ability to overcome antibiotic chemotherapy, Staphylococcus aureus continuously expands its ecological niche. It is resistant to many adverse environmental conditions, so that MRSA strains are mainly associated with hospital acquired infections (HA-MRSA). The rate of mortality of septicemia caused by VISA raised from 30% for MRSA, to almost 80% (Hiramatsu et al. 1997; Burnie et al. 2000; Fridkin et al. 2003). Thus, the emergence of resistant S. aureus bacteria has serious consequences both in terms of therapeutic failures and impact on Health Care System. To meet the growing challenge of S. aureus, the identification of novel targets for small molecules is one of the most important approach to face the problem (García-Lara et al. 2005; Wright and Sutherland 2007; Gibbons 2008; Silver 2011). To overcome antibiotic-mediated resistance, a valuable alternative would be the use of combination of drugs. Thus, substances that can increase susceptibility to currently licensed agents would be a very attractive and valuable option (Wagner and Ulrich-Merzenich 2009). In this paper the authors analyse five selected compounds from Chilean lichens for their antimicrobial activity against MRSA clinical isolated strains, tested alone and in combination with five therapeutically available antibiotics. Data from checkerboard assay were interpreted by Loewe additivity-based model and Bliss independencebased model. Materials and methods Organisms Twenty methicillin-resistant S. aureus strains were used in this study. The organisms were collected during a 4 years period, from 2006 to 2010, at the University Hospital “San Salvatore” of l’Aquila, Italy. They were isolated from hospitalized patients, from wounds, surgical wounds, vascular and urinary catheters, blood, respiratory tract. They were identified as MRSA organisms by Phoenix System (Becton Dickinson). The methicillin-resistant S. aureus from the American Type Culture Collection, ATCC 43300 was used as control. Four strains, namely, AQ004, AQ006, AQ007 and AQ012 clinical isolates and the reference strain ATCC 43300 were used for the drug interaction assay. All those strains were resistant to clindamycin, erythromycin, gentamicin, levofloxacin, oxacillin, with the exception of ATCC 43300 that was sensible to levofloxacin. Antibiotics All tested antibiotics, clindamycin (CLI), erythromycin (ERY), gentamicin (GEN), levofloxacin (LVX), oxacillin (OXA), were from Sigma– Aldrich (Milan, Italy). Secondary metabolites from lichens The lichen secondary metabolites used in this study are:
α -collatolic acid (COL), epiforellic acid (EPI), lobaric acid (LOB), per-
latolic acid (PER) and (+)-protolichesterinic acid (PRO), whose structures are reported in Fig. 1 and Table 1. These compounds were isolated and structurally determined as previously reported (Fiedler et al. 1986; Piovano et al. 1989). The degree of purity for the compounds was >98% as determined by thin layer chromatography (TLC) and 1 H NMR analyses. In vitro susceptibility tests The antimicrobial susceptibility pattern of the organisms used in this study was determined in accordance with the CLSI guidelines
Fig. 1. Chemical structures of lichen secondary metabolites used in this study.
(CLSI 2010) by microdilution test performed in a 96 microwell plates with an inoculum of 5 × 105 CFU/ml. Bactericidal tests were performed as previously described by Pearson et al. (1980) and Taylor et al. (1983).
Checkerboard microdilution assay The in vitro interactions between the antibiotics and the compounds from lichens were investigated by a two-dimensional checkerboard microdilution assay, using a 96-well microtitration plates as previously described (Segatore et al. 2012; Celenza et al. 2012). Briefly, in each well of the microplate 25 μl of microbial growth medium were added. An aliquot of 25 μl of a fourfold concentrated antibiotic was added to column 12. Then a twofold dilution was made from column 12 to column 2. A 25 μl aliquot of each drug concentration of the compound was added to rows A to G. Row H contained only the antibiotic whilst column 1 only the compound. Well H1 was the drug free well used as growth control. Finally, 50 μl of a saline solution (0.9% of NaCl) containing bacteria were added to each well of the microplate in order to obtain a final inoculum of 5 × 105 CFU/ml. The microtitre plates were incubated at 37 °C for 18 h. The growth in each well was quantified spectrophotometrically at 595 nm by a microplate reader iMark, BioRad (Milan, Italy). The percentage of growth in each well was calculated as:
ODdrug combination well − ODbackground ODdrug free well − ODbackground where the background was obtained from the microorganism-free plates, processed as the inoculated plates. The MICs for each combination of drugs were defined as the concentration of drug that reduced growth by 80% compared to that of organisms grown in the absence of drug. All experiments were performed in triplicate.
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
225
Table 1 Selected constituents tested for antimicrobial activity, their lichen species and Chilean regions of collection. Compound
Species
Chilean geographic origin
Reference
α -Collatolic acid Epiphorellic acid Lobaric acid Perlatolic acid Protolichesterinic acid
Lecanora atra (Hudson) Acharius Cornicularia epiphorella (Nyl.) Du Rietz Stereocaulon alpinum Laurer ex Funck Stereocaulon sp. Cornicularia aculeata (Schreb.) Ach.
Robert Island, Shetland del Sur, Antarctica Conguillío National Park, Región de la Araucanía Ardley Cove, King George Island, Shetland del Sur, Antarctica Parque Nacional Puyehue, Región de Los Lagos, Chile Ardley Cove, King George Island, Shetland del Sur, Antarctica
Sweeney and Zurenko (2003) Gibbons (2008) Sweeney and Zurenko (2003) Not published Sweeney and Zurenko (2003)
Table 2 MIC and MBC calculated for all S. aureus strains.
Drug interaction models In order to assess the nature of the in vitro interactions between the lichen compounds and antibiotics against each S. aureus, the data obtained from the checkerboard assay were analysed by nonparametric models based on the Loewe additivity model (LA) and the Bliss independence (BI) theory (Greco et al. 1995). Loewe additivity-based model The Loewe additivity model, based on the idea that an agent cannot interact with itself, is expressed by the following equation:
D1 D2 1= + IDX,1 IDX,2 where IDX,1 and IDX,2 are the concentrations of the drugs to result in X% inhibition for each respective drug alone. D1 and D2 are concentrations of each drug in the mixture that yield X% inhibition. The interaction index as define by Berenbaum (1977) is expressed by the equation:
I=
D1 D2 + IDX,1 IDX,2
When I > 1, Loewe antagonism is claimed, when I < 1, Loewe synergism is claimed. The interaction index as define by Berenbaum, can be adapted to calculate the fractional inhibitory concentration index (FICI), that is the mathematical expression of the effect of the combination of antibacterial agents expressed as:
FIC = FICA + FICB =
MICAB MICBA + MICA MICB
where MICA and MICB are the MICs of drugs A and B when acting alone and MICAB and MICBA are the MICs of drugs A and B when acting in combination. Among all FICs calculated for each microplate, the FICI was determined as the lowest FIC ( FICmin ) when synergy is supposed, or the highest FIC ( FICmax ) when antagonism is evident. Since in MIC determination, the variation in a single result places a MIC value in a three-dilution range (±1 dilution), therefore, the reproducibility errors in MIC checkerboard assays are considerable. For that reasons, the interpretation of FICI data should be done taking into consideration values well below or above the theoretical cut-off (1.0) defined by Berenbaum. Synergy was, therefore, defined when FICI ࣘ 0.5, while antagonism was defined when FICI > 4. A FIC index between 0.5 and 4 (0.5 < FICI ࣘ 4) was considered indifferent (Odds 2003). Bliss independence-based model In the Bliss models, the combined effects of the drugs calculated from the effect of the individual drugs, are compared with those obtained experimentally. The BI theory is described by the equation: Ei = EA × EB , where Ei is the calculated percentage of growth based
Compound
COL EPI LOB PER PRO
MIC (μg/ml)
MBC (μg/ml)
Range
MIC50
MIC90
Range
MBC50
MBC90
32–128 8–32 32–128 4–64 4–64
128 32 32 16 32
128 32 64 32 64
128–512 16–128 32–256 16–64 32–128
256 64 64 32 64
512 128 128 64 128
on the theoretical non-interactive combination of drug A and B, and EA and EB are the experimental percentages of growth of each drug acting alone. The experimental dose–response surface (Fig. 2A) is subtracted from the calculated theoretical surface to reveal any significant deviation from the zero-plane. The interaction is described by the difference (E) between the predicted and measured percentages of growth with drugs at various concentrations (E = Epredicted – Emeasured ). To determine the significance of differences between the experimental and calculated additive effects, the upper and lower 95% confidence limits of the experimental data were compared to the calculated additive effects. If the lower confidence limit of a point was greater than the calculated additivity, the observed synergy was considered to be significant. Similarly, if the upper confidence limit was lower than the calculated additivity, the observed antagonism was considered to be significant (Deminie et al. 1996; Prichard and Shipman 1990; Prichard et al. 1991). The E values calculated on a point-by-point basis were subsequently plotted on the z axis (Fig. 2B). Points of the difference surface above zero (positive) indicate synergy, below zero (negative) antagonism. In order to summarize the interaction surface, the Bliss synergy and antagonism differences and all their combinations were added up to yield a summary measure, respectively of Bliss synergy ( SYN) and Bliss antagonism ( ANT). Interactions 200% were considered strong (Meletiadis et al. 2005). Results In vitro susceptibility test and interaction of drugs The in vitro antibacterial effects of lichen compounds is reported in Table 2. For all the 20 MRSA clinical isolates and the ATCC4300, MIC50 and MIC90 were calculated, as well as MBC50 and MBC90 . The minimum inhibitory activity required to inhibit the growth of 90% of the organisms was ranging from 32 μg/ml for PER to 128 μg/ml for COL. The MBC90 measured was one- or twofold higher than the calculated MIC90 . Amongst the compounds the most active was PER with MIC50 and MIC90 values 16 μg/ml and 32 μg/ml, respectively. All 20 S. aureus clinical isolates and the reference strain ATCC 43300 were tested for their susceptibility to clindamycin, erythromycin, gentamicin, levofloxacin and oxacillin by microdilution test (data not shown). Amongst these strains, four (namely AQ004, AQ006, AQ007 and AQ012) and the reference strain ATCC 4330
226
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
Fig. 2. The three-dimensional plot of the experimental percentage of growth between gentamicin and protolichesterinic acid against the methicillin-resistant Staphylococcus aureus ATCC43300 (A). The three-dimensional plot of the difference between the predicted percentage of growth and the experimental percentage of growth based on the E model ( E = Epredicted – Emeasured ) between gentamicin and protolichesterinic acid against the methicillin-resistant Staphylococcus aureus ATCC43300 (B).
were chosen for the interaction assay, since they were resistant to all tested antibiotics (Table 3), with the exception of the control strains which was sensible to levofloxacin. The MICs values of the lichen compounds versus the selected strains are reported in Table 4. Clindamycin, erythromycin, gentamicin, levofloxacin and oxacillin were chosen for the drug-interaction assay, since they belong to different classes of antimicrobial agents, respectively, lincosamides, macrolides, aminoglycosides, quinolones and β-lactams. Tables 5–9 summarize the results of the broth microdilution checkerboard analysis interpreted by the FICI and E methods of the five MRSA strains for the combination of lichen compounds and
antibiotics. Results for each combination of antibiotics and lichen compounds are reported below. Clindamycin According to FICI interpretation, synergism was found only in S. aureus AQ006 in combination with LOB (FICI = 0.3125), PER (FICI = 0.3125) and PRO (FICI = 0.3125). Indifference (FICI > 0.5) was observed in all other combinations. E model interpretation for clindamycin confirmed synergism for those combinations. Moreover, most of the combinations, which FICI was approximately 0.5 and reported as indifference, were interpreted by the E model as synergism (Table 5).
Table 3 MIC of antibiotics for the strains of S. aureus used in the checkerboard assay. Strain
Median MIC (μg/ml) (range) Antibiotic
ATCC43300 AQ004 AQ006 AQ007 AQ012
CLI
ERY
GEN
LVX
OXA
8192 8192 (64–128) 128 (32–64) 32 16,384
(512–1024) 1024 (512–1024) 1024 512 (16–64) 64 (1024–2048) 1024
(64–256) 256 (16–32) 32 (128–256) 128 128 (32–64) 64
ࣘ0.5 (16–64) 16 (8–32) 8 (16–64) 16 32
(8–16) 8 (128–256) 128 2048 (256–512) 256 (256–512) 512
P. Bellio et al. / Phytomedicine 22 (2015) 223–230 Table 4 MIC of lichen compounds for the strains of S. aureus used in the checkerboard assay. Strain
Median MIC (μg/ml) (range)
Table 6 In vitro interaction between erythromycin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela .
Compound
ATCC43300 AQ004 AQ006 AQ007 AQ012
Compound Strain
COL
EPI
LOB
PER
PRO
(16–32) 32 (32–64) 32 (32–128) 128 (64–128) 128 (32–128) 64
(8–16) 8 (8–16) 16 (16–32) 16 (16–32) 32 (8–32) 16
32 64 (64–128) 64 64 (64–128) 64
(8–32) 32 (32–64) 32 (8–32) 16 (4–32) 16 (8–32) 16
(8–16) 16 (4–16) 16 (32–64) 32 32 (16–32) 32
Erythromycin For the combination of erythromycin with lichen compounds, no synergism was found. In all combinations, FICI and E model were in accordance defining indifference (Table 6). Gentamicin For the gentamicin combination, FICI synergism was observed in all lichen compounds and strains. The lowest FICIs were observed in PRO combination, with FIC indexes ranging from 0.1563 for ATCC 43300 and AQ007, to 0.2813 for AQ012 (Table 7). Levofloxacin For the combination of levofloxacin with lichen compounds, no synergism was found. FIC index for each combination was reported as indifference. E model confirmed FICI interpretation only in the combination with EPI and PRO. For all other combinations, antagonism was reported (Table 8). Oxacillin According to FICI interpretation, synergism was found in only in two MRSA strains in oxacillin combination with COL (ATCC 4330,
227
E modelb
FICI Median (range)
INT SYN (n) ANT (n)
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5625–0.625) 0.625 (0.5625–0.625) 0.625 (0.5625–0.625) 0.5625 (0.625–0.75) 0.75 (0.625–1) 0.75
IND IND IND IND IND
64.5 (9) 99.8 (13) 23.8 (4) 13.2 (2) 13.1 (2)
−99.2 (9) −83.6 (11) −96.8 (19) −86.9 (2) −94.8 (18)
IND IND IND IND IND
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.625–075) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (0.625–0.75) 0.75 (1.25–1.5) 1.25
IND IND IND IND IND
68.3 (11) 56.5 (9) 45.7 (7) 46.3 (6) 20.7 (5)
−97.7 (22) −94.6 (21) −90.4 (21) −92.8 (17) −84.6 (15)
IND IND IND IND IND
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.25 (0.625–0.75) 0.75 (0.5625–0.625) 0.625 1 (1.25–1.5) 1.25
IND IND IND IND IND
29.6 (5) 13.5 (3) 42.0 (7) 41.7 (6) 11.4 (2)
−28.4 (5) −68.1 (13) −89.3 (19) −13.1 (2) −18.6 (4)
IND IND IND IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.5 (0.75–1) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (1.0625–1.125) 1.125
IND IND IND IND IND
17.5 (3) 80.4 (21) 49.8 (9) 14.8 (5) 58.7 (10)
−66.9 (11) −36.1 (7) −63.6 (13) −33.9 (7) −74.0 (15)
IND IND IND IND IND
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.25 (0.625–075) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (0.625–0.75) 0.75
IND IND IND IND IND
74.0 (13) 93.2 (20) 99.8 (21) 88.6 (17) 38.4 (7)
−44.4 (9) −98.0 (25) −99.2 (26) −29.9 (8) −58.1 (10)
IND IND IND IND IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
Table 5 In vitro interaction between clindamycin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
INT
E modelb
FICI Median (range)
INT
SYN (n)
ANT (n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0625–1.125) 1.25 (1.0625–1.125) 1.125 (0.625–0.75) 0.625 1 (1.0625–1.125) 1.125
IND IND IND IND IND
62.7 (28) 42.7 (29) 98.1 (26) 87.5 (18) 43.1 (12)
−30.7 (30) −38.1 (35) −96.8 (43) −67.6 (20) −89.3 (18)
IND IND IND IND IND
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5156–0.5313) 0.5156 (0.5625–0.625) 0.625 (0.5313–05625) 0.5313 (0.5313–0.5625) 0.5625 (1.0313–1.0625) 1.0625
IND IND IND IND IND
1285.9 (30) 119.4 (20) 179.0 (23) 780.4 (41) 73.1 (15)
−85.9 (8) −82.5 (11) −46.0 (7) −43.2 (9) −91.5 (25)
SYN SYN SYN SYN IND
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0625–1.125) 1.0625 (0.5020–0.0539) 0.502 (0.2813–0.3125) 0.3125 (0.625–0.75) 0.75 (1.0625–1.125) 1.125
IND IND SYN IND IND
69.8 (36) 199.6 (51) 401.2 (23) 82.3 (16) 83.4 (15)
−92.2 (11) −40.0 (13) −65.4 (34) −65.7 (14) −94.0 (19)
IND SYN SYN IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0039–1.0078) 1.0039 (0.5–0.501) 0.501 (0.2813–0.3125) 0.3125 (0.5313–0.625) 0.625 (0.5–0.5002) 0.5002
IND IND SYN IND IND
22.8 (31) 164.1 (46) 1057.2 (41) 88.9 (44) 692,5 (30)
−99.6 (27) −50.1 (25) −52.3 (24) −41 (26) −71.4 (25)
IND SYN SYN IND SYN
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5078–0.5313) 0.5313 (0.5–0.501) 0.501 (0.3125–0.375) 0.3125 (0.5313–0.5625) 0.5625 (1.0625–1.125) 1.0625
IND IND SYN IND IND
149.2 (50) 154.6 (56) 1340.3 (38) 795.1 (40) 75.2 (17)
−4.5 (1) −32.6 (6) −65.6 (25) 95.1 (14) −79.9 (16)
SYN SYN SYN SYN IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
228
P. Bellio et al. / Phytomedicine 22 (2015) 223–230 Table 7 In vitro interaction between gentamicin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
E modelb
FICI Median (range)
INT
SYN (n)
ANT (n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.375–0.5) 0.5 (0.1875–0.25) 0.1875 (0.1875–0.250) 0.25 (0.375–0.5) 0.375 0.5
SYN SYN SYN SYN SYN
1182.1 (44) 1819.2 (52) 843.0 (37) 1145.0 (47) 473.0 (37)
−66.4 (13) −95.2 (20) −58.1 (12) −63.7 (12) −91.8 (15)
SYN SYN SYN SYN SYN
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
0.3125 (0.375–0.5) 0.5 (0.25–0.2578) 0.25 (0.25–0.2578) 0.25 (0.3125–0.375) 0.375
SYN SYN SYN SYN SYN
929.8 (29) 2192.0 (53) 1631.1 (46) 1237.8 (45) 564.4 (30)
−76.4 (16) −26.0 (3) −97.6 (18) −99.3 (21) −94.1 (10)
SYN SYN SYN SYN SYN
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.3125–0.375) 0.375 0.375 (0.5–0.5002) 0.5 (0.5–0.501) 0.5 (0.5–0.502) 0.5
SYN SYN SYN SYN SYN
659.3 (21) 2639.7 (51) 734.2 (43) 345.2 (20) 335.0 (23)
−89.4 (15) −66.5 (13) −91.9 (15) −68.3 (12) −76.2 (5)
SYN SYN SYN SYN SYN
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.2813–0.375) 0.375 (0.3125–0.375) 0.375 (0.375–0.5) 0.375 (0.375–0.5) 0.375 (0.375–0.5) 0.5
SYN SYN SYN SYN SYN
360.9 (21) 534.7 (32) 1058.5 (37) 598.7 (32) 595.7 (35)
−57.1 (10) −88.5 (15) −74.1 (16) −81.2 (16) −67.0 (12)
SYN SYN SYN SYN SYN
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.1563–0.1875) 0.1563 (0.1875–0.25) 0.1875 (0.25–0.2578) 0.25 (0.1563–0.1875) 0.1563 (0.2813–0.25) 0.2813
SYN SYN SYN SYN SYN
1657.2 (44) 3312.7 (55) 1173.4 (48) 2875.3 (54) 664.0 (35)
−87.4 (19) −58.6 (6) −44.7 (9) −35.1 (5) −97.3 (8)
SYN SYN SYN SYN SYN
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
Table 8 In vitro interaction between levofloxacin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound Strain
COL
EPI
LOB
PER
PRO
E modelb
FICI Median (range)
INT SYN (n) ANT (n)
INT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (1.5–2) 1.5 1.5 2 (1.125–1.5) 1.5
ND IND 25.3 (5) IND 24.9 (3) IND 5.3 (2) IND 29.5 (6)
ND −635.4 (59) −459.8 (45) −1601.8 (60) −502.0 (52)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (1.0625–1.5) 1.5 (1.0625–1.25) 1.25 1.5 1.5
IND IND IND IND
ND ND 72.3 (15) −45.1 (9) 98.1 (21) −62.3 (12) 56.8 (12) −90.9 (17) 64.3 (14) −87.9 (16)
IND IND IND IND
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND 2.5 2.5 (2–2.5) 2.5 2.25
IND IND IND IND
ND ND 99.7 (19) −485.4 (54) 98.6 (20) −507.2 (41) 0 (0) −998.5 (61) 92.9 (12) −643.4 (40)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND 2.5 2.5 2.5 2.25
IND IND IND IND
ND ND 43 (8) −971.7 (36) 14.5 (4) −483.0 (39) 57.4 (20) −408.1 (28) 75.1 (27) −272.8 (25)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (2–2.125) 2 (1.25–1.5) 1.5 (1–1.5) 1.5 1
IND IND IND IND
ND ND 82.1 (19) −96.0 (10) 57.2 (10) −98.3 (15) 38.3 (5) −78.1 (39) 97.1 (19) −68.0 (12)
IND IND IND IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. ND, not detected. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
FICI = 0.5; AQ004, FICI = 0.3125), one with EPI (ATCC 4330, FICI = 0.5) and four with PRO (ATCC 43300, AQ004, AQ007 and AQ012, FICI = 0.5). Indifference (FICI > 0.5) was observed in all other combinations. E model interpretation for oxacillin confirmed synergism for those combinations. Moreover, most of the combinations, which FICI was approximately 0.5 and reported as indifference, were interpreted by the E model as synergism (Table 9).
Discussion The aim of this paper is to investigate the potential antimicrobial activity of several selected natural compounds from lichens alone, and in combination with commercially available antibiotics. The MIC values calculated for these compounds versus MRSA clinical isolates were comparable with those generally found in many antibiotics used in clinical therapy. Since, the MBCs calculated were close to the MICs, we hypothesize that those compounds, at least against MRSA, act as bactericidal. Synergy is ideally defined as the interaction of two or more substances, to produce a combined effect greater than the sum of their separate effects. In addition, it is important to emphasize that even the models used to analyse drug–drug interaction might, somehow, affect the interpretation of the data. There are many published methods for assessing drug interactions, most of them are summarized and compared in the review of Greco (Greco et al. 1995). For instance we used two models, the FICI model based on Loewe additivity theory and E model based on Bliss independence theory. The percentage of agreement in the interpretation of the FIC index and the response surface approach was highly variable amongst the combinations. It was ranging from 0% to 100% for the levofloxacin combinations (Table 8) to 100% for the erythromycin combinations (Table 6).
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
229
Table 9 In vitro interaction between oxacillin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
E modelb
FICI Median (range)
INT
SYN (n)
ANT(n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.5002) 0.5 (0.3125–0.375) 0.3125 (0.5313–0.5625) 0.5625 (0.625–0.75) 0.75 (0.625–0.75) 0.75
SYN SYN IND IND IND
645.8 (25) 1075.3 (42) 238.0 (19) 132.7 (11) 131.1 (13)
−99.2 (9) −83.6 (6) −48.9 (9) −86.9 (19) −94.8 (22)
SYN SYN SYN SYN SYN
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.501) 0.5 (0.5625–0.6250) 0.625 0.625 (0.5078–0.5156) 0.5078 (0.5078–0.5156) 0.5078
SYN IND IND IND IND
665.7 (30) 495.5 (20) 819.0 (43) 1129.6 (34) 804.1 (32)
−88.0 (17) −99.1 (22) −62.0 (19) −82.4 (12) −47.3 (11)
SYN SYN SYN SYN SYN
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1–1.0039) 1 (1–1.002) 1.002 (0.5313–0.5625) 0.5625 (0.5625–0.625) 0.625 (0.5625–0.625) 0.625
IND IND IND IND IND
92.5 (18) 89.1 (15) 274.3 (27) 99.2 (17) 87.2 (13)
−77.6 (14) −85.0 (12) −94.1 (13) −67.2 (8) −80.5 (10)
IND IND SYN IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
0.75 1 (1.0039–1.25) 1.0039 1 (0.625–0.75) 0.75
IND IND IND IND IND
97.1 (23) 89.2 (16) 79.8 (18) 89.8 (17) 63.7 (9)
−75.2 (12) −67.3 (12) −98.9 (21) −54.6 (9) −92.2 (11)
IND IND IND IND IND
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.5002) 0.5 (0.5–0.5002) 0.5 (0.5039–0.5078) 0.5039 (0.5–0.502) 0.5 (0.5–0.5002) 0.5
SYN SYN IND SYN SYN
426.6 (18) 406.8 (37) 380.2 (31) 819.8 (27) 1047.5 (35)
−31.5 (6) −74.9 (15) −99.2 (10) −76.7 (15) −23.4 (4)
SYN SYN SYN SYN SYN
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
The discrepancy between the results of the two methods is the natural consequence of the high variability intrinsic in the FICI model (Cappelletty and Rybak 1996; Mackay et al. 2000; Sun et al. 2008; TeDorsthorst et al. 2002). The results obtained by this method are strongly dependent on the MIC endpoints and the cutoff values used to define synergism and antagonism. For instance, most of the combinations interpreted as indifference by FIC index model, which value is more than 0.5 but less than 1, are interpreted as synergic using the E model. The same is found in levofloxacin combinations, in which, the interpretation of FIC index model as indifference (FICI > 1 but
Contents lists available at ScienceDirect
Phytomedicine journal homepage: www.elsevier.com/locate/phymed
Interaction between lichen secondary metabolites and antibiotics against clinical isolates methicillin-resistant Staphylococcus aureus strains Pierangelo Bellio a, Bernardetta Segatore a, Alisia Mancini a, Letizia Di Pietro a, Carlo Bottoni a, Alessia Sabatini a, Fabrizia Brisdelli a, Marisa Piovano b, Marcello Nicoletti c, Gianfranco Amicosante a, Mariagrazia Perilli a, Giuseppe Celenza a,∗ a
Department of Biotechnological and Applied Clinical Sciences, University of l’Aquila, L’Aquila, Italy Department of Chemistry, Universidad Técnica F. Santa María, Casilla 110 V, Valparaíso, 6, Chile c Department of Enviromental Biology, University Sapienza, Rome, Italy b
a r t i c l e
i n f o
Article history: Received 16 June 2014 Revised 6 October 2014 Accepted 14 December 2014
Keywords: Lichen secondary metabolites MRSA Checkerboard assay Antimicrobial activity
a b s t r a c t The in vitro antimicrobial activities of five compounds isolated from lichens, collected in several Southern regions of Chile (including the Chilean Antarctic Territory), were evaluated alone and in combination with five therapeutically available antibiotics, using checkerboard microdilution assay against methicillin-resistant clinical isolates strains of Staphylococcus aureus. MIC90 , MIC50 , as well as MBC90 and MBC50 , for the lichen compounds were evaluated. The MIC90 was ranging from 32 μg/ml for perlatolic acid to 128 μg/ml for α -collatolic acid. MBC90 was ranging from onefold up to twofold the MIC90 for each compound. A synergistic action was observed in combination with gentamicin, whilst antagonism was observed for some lichen compounds in combination with levofloxacin. All combinations with erythromycin were indifferent, whilst variability was observed for clindamycin and oxacillin combinations. Data from checkerboard assay were analysed and interpreted using the fractional inhibitory concentration index and the response surface approach using the E model. Discrepancies were found between both methods for some combinations. These could mainly be explained by the failure of FIC approach, being too much subjective and sensitive to experimental errors. These findings suggest, however, that the natural compounds from lichens are good candidates for the individuation of novel templates for the development of new antimicrobial agents or combinations of drugs for chemotherapy. © 2015 Elsevier GmbH. All rights reserved.
Introduction Lichens are symbiotic organisms derived by the close cellular union of a fungal (mycobiont) and an algal and/or cyanobacterial (phytobiont) partner, comprising about 20,000 known species. They produce a variety of secondary compounds that typically arise from the fungal component secondary metabolism, many of them, exclusive of the lichen production. Chemotaxonomic studies have shown that the most unique lichen metabolites belong to the chemical classes of depsides, depsidones and dibenzofurans. The almost 800 known lichen secondary metabolites can be classified according to the classic biosynthetic pathways: the poliketidic path (monocyclic phenols, depsides, depsidones, depsones, dibenzofurans, xanthones, naphtaquinones anthraquinones, macrocyclic lactones, aliphatic acids, etc.), the mevalonic acid path (steroids, carotenoids, ∗ Corresponding author at: Department of Biotechnological and Applied Clinical Sciences, University of l’Aquila, Via Vetoio, 1, 67100 l’Aquila, Italy. Tel.: +39 0862433444. E-mail address: [email protected] (G. Celenza).
http://dx.doi.org/10.1016/j.phymed.2014.12.005 0944-7113/© 2015 Elsevier GmbH. All rights reserved.
etc.) and the shikimic acid path (amino acid derivatives, cyclopeptides, etc.) (Huneck 1999). Although medicinal plants have been used for centuries as sources of therapeutic agents worldwide, they cannot be classified as pure and efficient antimicrobial agents. However, in spite of the fact that plant-derived antibacterial compounds show a general low degree of activity, most plants, indeed, are successful in fighting infections (Hemaiswarya et al. 2008). Plants, in fact adopt “synergy” as their peculiar different paradigm to fight pathogenic microorganisms. Several studies have demonstrated that a number of natural products, which failed as antimicrobials, are able to dramatically increase the effectiveness of chemotherapeutic agents against Gram-negative and Gram-positive bacteria (Gibbons and Udo 2000; Tegos et al. 2002; Stavri et al. 2007; Hemaiswarya et al. 2008; Celenza et al. 2012; Segatore et al. 2012). Antimicrobial resistance has emerged among pathogenic bacteria since the beginning of the antibiotic era. Resistance potentially extends to the entire repertoire of available therapeutic agents. Nowadays, bacteria expressing multidrug resistant (MDR), extensively drug
224
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
resistant (XDR) and pandrug resistant (PDR) phenotypes are amongst the most important cause of infections in nosocomial and community settings and new drugs are urgently needed. As a result of its intrinsic ability to overcome antibiotic chemotherapy, Staphylococcus aureus continuously expands its ecological niche. It is resistant to many adverse environmental conditions, so that MRSA strains are mainly associated with hospital acquired infections (HA-MRSA). The rate of mortality of septicemia caused by VISA raised from 30% for MRSA, to almost 80% (Hiramatsu et al. 1997; Burnie et al. 2000; Fridkin et al. 2003). Thus, the emergence of resistant S. aureus bacteria has serious consequences both in terms of therapeutic failures and impact on Health Care System. To meet the growing challenge of S. aureus, the identification of novel targets for small molecules is one of the most important approach to face the problem (García-Lara et al. 2005; Wright and Sutherland 2007; Gibbons 2008; Silver 2011). To overcome antibiotic-mediated resistance, a valuable alternative would be the use of combination of drugs. Thus, substances that can increase susceptibility to currently licensed agents would be a very attractive and valuable option (Wagner and Ulrich-Merzenich 2009). In this paper the authors analyse five selected compounds from Chilean lichens for their antimicrobial activity against MRSA clinical isolated strains, tested alone and in combination with five therapeutically available antibiotics. Data from checkerboard assay were interpreted by Loewe additivity-based model and Bliss independencebased model. Materials and methods Organisms Twenty methicillin-resistant S. aureus strains were used in this study. The organisms were collected during a 4 years period, from 2006 to 2010, at the University Hospital “San Salvatore” of l’Aquila, Italy. They were isolated from hospitalized patients, from wounds, surgical wounds, vascular and urinary catheters, blood, respiratory tract. They were identified as MRSA organisms by Phoenix System (Becton Dickinson). The methicillin-resistant S. aureus from the American Type Culture Collection, ATCC 43300 was used as control. Four strains, namely, AQ004, AQ006, AQ007 and AQ012 clinical isolates and the reference strain ATCC 43300 were used for the drug interaction assay. All those strains were resistant to clindamycin, erythromycin, gentamicin, levofloxacin, oxacillin, with the exception of ATCC 43300 that was sensible to levofloxacin. Antibiotics All tested antibiotics, clindamycin (CLI), erythromycin (ERY), gentamicin (GEN), levofloxacin (LVX), oxacillin (OXA), were from Sigma– Aldrich (Milan, Italy). Secondary metabolites from lichens The lichen secondary metabolites used in this study are:
α -collatolic acid (COL), epiforellic acid (EPI), lobaric acid (LOB), per-
latolic acid (PER) and (+)-protolichesterinic acid (PRO), whose structures are reported in Fig. 1 and Table 1. These compounds were isolated and structurally determined as previously reported (Fiedler et al. 1986; Piovano et al. 1989). The degree of purity for the compounds was >98% as determined by thin layer chromatography (TLC) and 1 H NMR analyses. In vitro susceptibility tests The antimicrobial susceptibility pattern of the organisms used in this study was determined in accordance with the CLSI guidelines
Fig. 1. Chemical structures of lichen secondary metabolites used in this study.
(CLSI 2010) by microdilution test performed in a 96 microwell plates with an inoculum of 5 × 105 CFU/ml. Bactericidal tests were performed as previously described by Pearson et al. (1980) and Taylor et al. (1983).
Checkerboard microdilution assay The in vitro interactions between the antibiotics and the compounds from lichens were investigated by a two-dimensional checkerboard microdilution assay, using a 96-well microtitration plates as previously described (Segatore et al. 2012; Celenza et al. 2012). Briefly, in each well of the microplate 25 μl of microbial growth medium were added. An aliquot of 25 μl of a fourfold concentrated antibiotic was added to column 12. Then a twofold dilution was made from column 12 to column 2. A 25 μl aliquot of each drug concentration of the compound was added to rows A to G. Row H contained only the antibiotic whilst column 1 only the compound. Well H1 was the drug free well used as growth control. Finally, 50 μl of a saline solution (0.9% of NaCl) containing bacteria were added to each well of the microplate in order to obtain a final inoculum of 5 × 105 CFU/ml. The microtitre plates were incubated at 37 °C for 18 h. The growth in each well was quantified spectrophotometrically at 595 nm by a microplate reader iMark, BioRad (Milan, Italy). The percentage of growth in each well was calculated as:
ODdrug combination well − ODbackground ODdrug free well − ODbackground where the background was obtained from the microorganism-free plates, processed as the inoculated plates. The MICs for each combination of drugs were defined as the concentration of drug that reduced growth by 80% compared to that of organisms grown in the absence of drug. All experiments were performed in triplicate.
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
225
Table 1 Selected constituents tested for antimicrobial activity, their lichen species and Chilean regions of collection. Compound
Species
Chilean geographic origin
Reference
α -Collatolic acid Epiphorellic acid Lobaric acid Perlatolic acid Protolichesterinic acid
Lecanora atra (Hudson) Acharius Cornicularia epiphorella (Nyl.) Du Rietz Stereocaulon alpinum Laurer ex Funck Stereocaulon sp. Cornicularia aculeata (Schreb.) Ach.
Robert Island, Shetland del Sur, Antarctica Conguillío National Park, Región de la Araucanía Ardley Cove, King George Island, Shetland del Sur, Antarctica Parque Nacional Puyehue, Región de Los Lagos, Chile Ardley Cove, King George Island, Shetland del Sur, Antarctica
Sweeney and Zurenko (2003) Gibbons (2008) Sweeney and Zurenko (2003) Not published Sweeney and Zurenko (2003)
Table 2 MIC and MBC calculated for all S. aureus strains.
Drug interaction models In order to assess the nature of the in vitro interactions between the lichen compounds and antibiotics against each S. aureus, the data obtained from the checkerboard assay were analysed by nonparametric models based on the Loewe additivity model (LA) and the Bliss independence (BI) theory (Greco et al. 1995). Loewe additivity-based model The Loewe additivity model, based on the idea that an agent cannot interact with itself, is expressed by the following equation:
D1 D2 1= + IDX,1 IDX,2 where IDX,1 and IDX,2 are the concentrations of the drugs to result in X% inhibition for each respective drug alone. D1 and D2 are concentrations of each drug in the mixture that yield X% inhibition. The interaction index as define by Berenbaum (1977) is expressed by the equation:
I=
D1 D2 + IDX,1 IDX,2
When I > 1, Loewe antagonism is claimed, when I < 1, Loewe synergism is claimed. The interaction index as define by Berenbaum, can be adapted to calculate the fractional inhibitory concentration index (FICI), that is the mathematical expression of the effect of the combination of antibacterial agents expressed as:
FIC = FICA + FICB =
MICAB MICBA + MICA MICB
where MICA and MICB are the MICs of drugs A and B when acting alone and MICAB and MICBA are the MICs of drugs A and B when acting in combination. Among all FICs calculated for each microplate, the FICI was determined as the lowest FIC ( FICmin ) when synergy is supposed, or the highest FIC ( FICmax ) when antagonism is evident. Since in MIC determination, the variation in a single result places a MIC value in a three-dilution range (±1 dilution), therefore, the reproducibility errors in MIC checkerboard assays are considerable. For that reasons, the interpretation of FICI data should be done taking into consideration values well below or above the theoretical cut-off (1.0) defined by Berenbaum. Synergy was, therefore, defined when FICI ࣘ 0.5, while antagonism was defined when FICI > 4. A FIC index between 0.5 and 4 (0.5 < FICI ࣘ 4) was considered indifferent (Odds 2003). Bliss independence-based model In the Bliss models, the combined effects of the drugs calculated from the effect of the individual drugs, are compared with those obtained experimentally. The BI theory is described by the equation: Ei = EA × EB , where Ei is the calculated percentage of growth based
Compound
COL EPI LOB PER PRO
MIC (μg/ml)
MBC (μg/ml)
Range
MIC50
MIC90
Range
MBC50
MBC90
32–128 8–32 32–128 4–64 4–64
128 32 32 16 32
128 32 64 32 64
128–512 16–128 32–256 16–64 32–128
256 64 64 32 64
512 128 128 64 128
on the theoretical non-interactive combination of drug A and B, and EA and EB are the experimental percentages of growth of each drug acting alone. The experimental dose–response surface (Fig. 2A) is subtracted from the calculated theoretical surface to reveal any significant deviation from the zero-plane. The interaction is described by the difference (E) between the predicted and measured percentages of growth with drugs at various concentrations (E = Epredicted – Emeasured ). To determine the significance of differences between the experimental and calculated additive effects, the upper and lower 95% confidence limits of the experimental data were compared to the calculated additive effects. If the lower confidence limit of a point was greater than the calculated additivity, the observed synergy was considered to be significant. Similarly, if the upper confidence limit was lower than the calculated additivity, the observed antagonism was considered to be significant (Deminie et al. 1996; Prichard and Shipman 1990; Prichard et al. 1991). The E values calculated on a point-by-point basis were subsequently plotted on the z axis (Fig. 2B). Points of the difference surface above zero (positive) indicate synergy, below zero (negative) antagonism. In order to summarize the interaction surface, the Bliss synergy and antagonism differences and all their combinations were added up to yield a summary measure, respectively of Bliss synergy ( SYN) and Bliss antagonism ( ANT). Interactions 200% were considered strong (Meletiadis et al. 2005). Results In vitro susceptibility test and interaction of drugs The in vitro antibacterial effects of lichen compounds is reported in Table 2. For all the 20 MRSA clinical isolates and the ATCC4300, MIC50 and MIC90 were calculated, as well as MBC50 and MBC90 . The minimum inhibitory activity required to inhibit the growth of 90% of the organisms was ranging from 32 μg/ml for PER to 128 μg/ml for COL. The MBC90 measured was one- or twofold higher than the calculated MIC90 . Amongst the compounds the most active was PER with MIC50 and MIC90 values 16 μg/ml and 32 μg/ml, respectively. All 20 S. aureus clinical isolates and the reference strain ATCC 43300 were tested for their susceptibility to clindamycin, erythromycin, gentamicin, levofloxacin and oxacillin by microdilution test (data not shown). Amongst these strains, four (namely AQ004, AQ006, AQ007 and AQ012) and the reference strain ATCC 4330
226
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
Fig. 2. The three-dimensional plot of the experimental percentage of growth between gentamicin and protolichesterinic acid against the methicillin-resistant Staphylococcus aureus ATCC43300 (A). The three-dimensional plot of the difference between the predicted percentage of growth and the experimental percentage of growth based on the E model ( E = Epredicted – Emeasured ) between gentamicin and protolichesterinic acid against the methicillin-resistant Staphylococcus aureus ATCC43300 (B).
were chosen for the interaction assay, since they were resistant to all tested antibiotics (Table 3), with the exception of the control strains which was sensible to levofloxacin. The MICs values of the lichen compounds versus the selected strains are reported in Table 4. Clindamycin, erythromycin, gentamicin, levofloxacin and oxacillin were chosen for the drug-interaction assay, since they belong to different classes of antimicrobial agents, respectively, lincosamides, macrolides, aminoglycosides, quinolones and β-lactams. Tables 5–9 summarize the results of the broth microdilution checkerboard analysis interpreted by the FICI and E methods of the five MRSA strains for the combination of lichen compounds and
antibiotics. Results for each combination of antibiotics and lichen compounds are reported below. Clindamycin According to FICI interpretation, synergism was found only in S. aureus AQ006 in combination with LOB (FICI = 0.3125), PER (FICI = 0.3125) and PRO (FICI = 0.3125). Indifference (FICI > 0.5) was observed in all other combinations. E model interpretation for clindamycin confirmed synergism for those combinations. Moreover, most of the combinations, which FICI was approximately 0.5 and reported as indifference, were interpreted by the E model as synergism (Table 5).
Table 3 MIC of antibiotics for the strains of S. aureus used in the checkerboard assay. Strain
Median MIC (μg/ml) (range) Antibiotic
ATCC43300 AQ004 AQ006 AQ007 AQ012
CLI
ERY
GEN
LVX
OXA
8192 8192 (64–128) 128 (32–64) 32 16,384
(512–1024) 1024 (512–1024) 1024 512 (16–64) 64 (1024–2048) 1024
(64–256) 256 (16–32) 32 (128–256) 128 128 (32–64) 64
ࣘ0.5 (16–64) 16 (8–32) 8 (16–64) 16 32
(8–16) 8 (128–256) 128 2048 (256–512) 256 (256–512) 512
P. Bellio et al. / Phytomedicine 22 (2015) 223–230 Table 4 MIC of lichen compounds for the strains of S. aureus used in the checkerboard assay. Strain
Median MIC (μg/ml) (range)
Table 6 In vitro interaction between erythromycin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela .
Compound
ATCC43300 AQ004 AQ006 AQ007 AQ012
Compound Strain
COL
EPI
LOB
PER
PRO
(16–32) 32 (32–64) 32 (32–128) 128 (64–128) 128 (32–128) 64
(8–16) 8 (8–16) 16 (16–32) 16 (16–32) 32 (8–32) 16
32 64 (64–128) 64 64 (64–128) 64
(8–32) 32 (32–64) 32 (8–32) 16 (4–32) 16 (8–32) 16
(8–16) 16 (4–16) 16 (32–64) 32 32 (16–32) 32
Erythromycin For the combination of erythromycin with lichen compounds, no synergism was found. In all combinations, FICI and E model were in accordance defining indifference (Table 6). Gentamicin For the gentamicin combination, FICI synergism was observed in all lichen compounds and strains. The lowest FICIs were observed in PRO combination, with FIC indexes ranging from 0.1563 for ATCC 43300 and AQ007, to 0.2813 for AQ012 (Table 7). Levofloxacin For the combination of levofloxacin with lichen compounds, no synergism was found. FIC index for each combination was reported as indifference. E model confirmed FICI interpretation only in the combination with EPI and PRO. For all other combinations, antagonism was reported (Table 8). Oxacillin According to FICI interpretation, synergism was found in only in two MRSA strains in oxacillin combination with COL (ATCC 4330,
227
E modelb
FICI Median (range)
INT SYN (n) ANT (n)
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5625–0.625) 0.625 (0.5625–0.625) 0.625 (0.5625–0.625) 0.5625 (0.625–0.75) 0.75 (0.625–1) 0.75
IND IND IND IND IND
64.5 (9) 99.8 (13) 23.8 (4) 13.2 (2) 13.1 (2)
−99.2 (9) −83.6 (11) −96.8 (19) −86.9 (2) −94.8 (18)
IND IND IND IND IND
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.625–075) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (0.625–0.75) 0.75 (1.25–1.5) 1.25
IND IND IND IND IND
68.3 (11) 56.5 (9) 45.7 (7) 46.3 (6) 20.7 (5)
−97.7 (22) −94.6 (21) −90.4 (21) −92.8 (17) −84.6 (15)
IND IND IND IND IND
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.25 (0.625–0.75) 0.75 (0.5625–0.625) 0.625 1 (1.25–1.5) 1.25
IND IND IND IND IND
29.6 (5) 13.5 (3) 42.0 (7) 41.7 (6) 11.4 (2)
−28.4 (5) −68.1 (13) −89.3 (19) −13.1 (2) −18.6 (4)
IND IND IND IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.5 (0.75–1) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (1.0625–1.125) 1.125
IND IND IND IND IND
17.5 (3) 80.4 (21) 49.8 (9) 14.8 (5) 58.7 (10)
−66.9 (11) −36.1 (7) −63.6 (13) −33.9 (7) −74.0 (15)
IND IND IND IND IND
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.25–1.5) 1.25 (0.625–075) 0.75 (0.5625–0.625) 0.625 (0.625–0.75) 0.75 (0.625–0.75) 0.75
IND IND IND IND IND
74.0 (13) 93.2 (20) 99.8 (21) 88.6 (17) 38.4 (7)
−44.4 (9) −98.0 (25) −99.2 (26) −29.9 (8) −58.1 (10)
IND IND IND IND IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
Table 5 In vitro interaction between clindamycin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
INT
E modelb
FICI Median (range)
INT
SYN (n)
ANT (n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0625–1.125) 1.25 (1.0625–1.125) 1.125 (0.625–0.75) 0.625 1 (1.0625–1.125) 1.125
IND IND IND IND IND
62.7 (28) 42.7 (29) 98.1 (26) 87.5 (18) 43.1 (12)
−30.7 (30) −38.1 (35) −96.8 (43) −67.6 (20) −89.3 (18)
IND IND IND IND IND
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5156–0.5313) 0.5156 (0.5625–0.625) 0.625 (0.5313–05625) 0.5313 (0.5313–0.5625) 0.5625 (1.0313–1.0625) 1.0625
IND IND IND IND IND
1285.9 (30) 119.4 (20) 179.0 (23) 780.4 (41) 73.1 (15)
−85.9 (8) −82.5 (11) −46.0 (7) −43.2 (9) −91.5 (25)
SYN SYN SYN SYN IND
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0625–1.125) 1.0625 (0.5020–0.0539) 0.502 (0.2813–0.3125) 0.3125 (0.625–0.75) 0.75 (1.0625–1.125) 1.125
IND IND SYN IND IND
69.8 (36) 199.6 (51) 401.2 (23) 82.3 (16) 83.4 (15)
−92.2 (11) −40.0 (13) −65.4 (34) −65.7 (14) −94.0 (19)
IND SYN SYN IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1.0039–1.0078) 1.0039 (0.5–0.501) 0.501 (0.2813–0.3125) 0.3125 (0.5313–0.625) 0.625 (0.5–0.5002) 0.5002
IND IND SYN IND IND
22.8 (31) 164.1 (46) 1057.2 (41) 88.9 (44) 692,5 (30)
−99.6 (27) −50.1 (25) −52.3 (24) −41 (26) −71.4 (25)
IND SYN SYN IND SYN
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5078–0.5313) 0.5313 (0.5–0.501) 0.501 (0.3125–0.375) 0.3125 (0.5313–0.5625) 0.5625 (1.0625–1.125) 1.0625
IND IND SYN IND IND
149.2 (50) 154.6 (56) 1340.3 (38) 795.1 (40) 75.2 (17)
−4.5 (1) −32.6 (6) −65.6 (25) 95.1 (14) −79.9 (16)
SYN SYN SYN SYN IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
228
P. Bellio et al. / Phytomedicine 22 (2015) 223–230 Table 7 In vitro interaction between gentamicin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
E modelb
FICI Median (range)
INT
SYN (n)
ANT (n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.375–0.5) 0.5 (0.1875–0.25) 0.1875 (0.1875–0.250) 0.25 (0.375–0.5) 0.375 0.5
SYN SYN SYN SYN SYN
1182.1 (44) 1819.2 (52) 843.0 (37) 1145.0 (47) 473.0 (37)
−66.4 (13) −95.2 (20) −58.1 (12) −63.7 (12) −91.8 (15)
SYN SYN SYN SYN SYN
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
0.3125 (0.375–0.5) 0.5 (0.25–0.2578) 0.25 (0.25–0.2578) 0.25 (0.3125–0.375) 0.375
SYN SYN SYN SYN SYN
929.8 (29) 2192.0 (53) 1631.1 (46) 1237.8 (45) 564.4 (30)
−76.4 (16) −26.0 (3) −97.6 (18) −99.3 (21) −94.1 (10)
SYN SYN SYN SYN SYN
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.3125–0.375) 0.375 0.375 (0.5–0.5002) 0.5 (0.5–0.501) 0.5 (0.5–0.502) 0.5
SYN SYN SYN SYN SYN
659.3 (21) 2639.7 (51) 734.2 (43) 345.2 (20) 335.0 (23)
−89.4 (15) −66.5 (13) −91.9 (15) −68.3 (12) −76.2 (5)
SYN SYN SYN SYN SYN
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.2813–0.375) 0.375 (0.3125–0.375) 0.375 (0.375–0.5) 0.375 (0.375–0.5) 0.375 (0.375–0.5) 0.5
SYN SYN SYN SYN SYN
360.9 (21) 534.7 (32) 1058.5 (37) 598.7 (32) 595.7 (35)
−57.1 (10) −88.5 (15) −74.1 (16) −81.2 (16) −67.0 (12)
SYN SYN SYN SYN SYN
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.1563–0.1875) 0.1563 (0.1875–0.25) 0.1875 (0.25–0.2578) 0.25 (0.1563–0.1875) 0.1563 (0.2813–0.25) 0.2813
SYN SYN SYN SYN SYN
1657.2 (44) 3312.7 (55) 1173.4 (48) 2875.3 (54) 664.0 (35)
−87.4 (19) −58.6 (6) −44.7 (9) −35.1 (5) −97.3 (8)
SYN SYN SYN SYN SYN
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
Table 8 In vitro interaction between levofloxacin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound Strain
COL
EPI
LOB
PER
PRO
E modelb
FICI Median (range)
INT SYN (n) ANT (n)
INT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (1.5–2) 1.5 1.5 2 (1.125–1.5) 1.5
ND IND 25.3 (5) IND 24.9 (3) IND 5.3 (2) IND 29.5 (6)
ND −635.4 (59) −459.8 (45) −1601.8 (60) −502.0 (52)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (1.0625–1.5) 1.5 (1.0625–1.25) 1.25 1.5 1.5
IND IND IND IND
ND ND 72.3 (15) −45.1 (9) 98.1 (21) −62.3 (12) 56.8 (12) −90.9 (17) 64.3 (14) −87.9 (16)
IND IND IND IND
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND 2.5 2.5 (2–2.5) 2.5 2.25
IND IND IND IND
ND ND 99.7 (19) −485.4 (54) 98.6 (20) −507.2 (41) 0 (0) −998.5 (61) 92.9 (12) −643.4 (40)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND 2.5 2.5 2.5 2.25
IND IND IND IND
ND ND 43 (8) −971.7 (36) 14.5 (4) −483.0 (39) 57.4 (20) −408.1 (28) 75.1 (27) −272.8 (25)
ANT ANT ANT ANT
ATCC43300 AQ004 AQ006 AQ007 AQ012
ND (2–2.125) 2 (1.25–1.5) 1.5 (1–1.5) 1.5 1
IND IND IND IND
ND ND 82.1 (19) −96.0 (10) 57.2 (10) −98.3 (15) 38.3 (5) −78.1 (39) 97.1 (19) −68.0 (12)
IND IND IND IND
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. ND, not detected. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
FICI = 0.5; AQ004, FICI = 0.3125), one with EPI (ATCC 4330, FICI = 0.5) and four with PRO (ATCC 43300, AQ004, AQ007 and AQ012, FICI = 0.5). Indifference (FICI > 0.5) was observed in all other combinations. E model interpretation for oxacillin confirmed synergism for those combinations. Moreover, most of the combinations, which FICI was approximately 0.5 and reported as indifference, were interpreted by the E model as synergism (Table 9).
Discussion The aim of this paper is to investigate the potential antimicrobial activity of several selected natural compounds from lichens alone, and in combination with commercially available antibiotics. The MIC values calculated for these compounds versus MRSA clinical isolates were comparable with those generally found in many antibiotics used in clinical therapy. Since, the MBCs calculated were close to the MICs, we hypothesize that those compounds, at least against MRSA, act as bactericidal. Synergy is ideally defined as the interaction of two or more substances, to produce a combined effect greater than the sum of their separate effects. In addition, it is important to emphasize that even the models used to analyse drug–drug interaction might, somehow, affect the interpretation of the data. There are many published methods for assessing drug interactions, most of them are summarized and compared in the review of Greco (Greco et al. 1995). For instance we used two models, the FICI model based on Loewe additivity theory and E model based on Bliss independence theory. The percentage of agreement in the interpretation of the FIC index and the response surface approach was highly variable amongst the combinations. It was ranging from 0% to 100% for the levofloxacin combinations (Table 8) to 100% for the erythromycin combinations (Table 6).
P. Bellio et al. / Phytomedicine 22 (2015) 223–230
229
Table 9 In vitro interaction between oxacillin and secondary metabolites from lichens as determined by nonparametric FICI and the E modela . Compound
Strain
E modelb
FICI Median (range)
INT
SYN (n)
ANT(n)
INT
COL
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.5002) 0.5 (0.3125–0.375) 0.3125 (0.5313–0.5625) 0.5625 (0.625–0.75) 0.75 (0.625–0.75) 0.75
SYN SYN IND IND IND
645.8 (25) 1075.3 (42) 238.0 (19) 132.7 (11) 131.1 (13)
−99.2 (9) −83.6 (6) −48.9 (9) −86.9 (19) −94.8 (22)
SYN SYN SYN SYN SYN
EPI
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.501) 0.5 (0.5625–0.6250) 0.625 0.625 (0.5078–0.5156) 0.5078 (0.5078–0.5156) 0.5078
SYN IND IND IND IND
665.7 (30) 495.5 (20) 819.0 (43) 1129.6 (34) 804.1 (32)
−88.0 (17) −99.1 (22) −62.0 (19) −82.4 (12) −47.3 (11)
SYN SYN SYN SYN SYN
LOB
ATCC43300 AQ004 AQ006 AQ007 AQ012
(1–1.0039) 1 (1–1.002) 1.002 (0.5313–0.5625) 0.5625 (0.5625–0.625) 0.625 (0.5625–0.625) 0.625
IND IND IND IND IND
92.5 (18) 89.1 (15) 274.3 (27) 99.2 (17) 87.2 (13)
−77.6 (14) −85.0 (12) −94.1 (13) −67.2 (8) −80.5 (10)
IND IND SYN IND IND
PER
ATCC43300 AQ004 AQ006 AQ007 AQ012
0.75 1 (1.0039–1.25) 1.0039 1 (0.625–0.75) 0.75
IND IND IND IND IND
97.1 (23) 89.2 (16) 79.8 (18) 89.8 (17) 63.7 (9)
−75.2 (12) −67.3 (12) −98.9 (21) −54.6 (9) −92.2 (11)
IND IND IND IND IND
PRO
ATCC43300 AQ004 AQ006 AQ007 AQ012
(0.5–0.5002) 0.5 (0.5–0.5002) 0.5 (0.5039–0.5078) 0.5039 (0.5–0.502) 0.5 (0.5–0.5002) 0.5
SYN SYN IND SYN SYN
426.6 (18) 406.8 (37) 380.2 (31) 819.8 (27) 1047.5 (35)
−31.5 (6) −74.9 (15) −99.2 (10) −76.7 (15) −23.4 (4)
SYN SYN SYN SYN SYN
a INT, interpretation; IND, indifference; SYN, synergy; ANT, antagonism. Synergy was defined as an FICI of ࣘ0.5, antagonism was defined as an FICI of >4, and indifference was defined as an FICI >0.5 and ࣘ4. b n, number of drug combinations (among the 77 drug combinations for each strain) with statistically significant synergy or antagonism.
The discrepancy between the results of the two methods is the natural consequence of the high variability intrinsic in the FICI model (Cappelletty and Rybak 1996; Mackay et al. 2000; Sun et al. 2008; TeDorsthorst et al. 2002). The results obtained by this method are strongly dependent on the MIC endpoints and the cutoff values used to define synergism and antagonism. For instance, most of the combinations interpreted as indifference by FIC index model, which value is more than 0.5 but less than 1, are interpreted as synergic using the E model. The same is found in levofloxacin combinations, in which, the interpretation of FIC index model as indifference (FICI > 1 but
