Toxicity of Australian essential oil Backhousia citriodora (Lemon myrtle). Part 1. Antimicrobial activity and in vitro cytotoxicity

Food and Chemical Toxicology 40 (2002) 535 543 Research section Toxicity of Australian essential oil Backhousia citriodora (Lemon myrtle). Part 1. Antimicrobial activity and in vitro cytotoxicity A.J. Hayes*,B. Markovic Chemical Safety and Applied Toxicology (CSAT) Laboratories, School of Safety Science, The University of New South Wales, Sydney, 2052, Australia Accepted 17 September 2001 www.elsevier.com/locate/foodchemtox Abstract The antimicrobial and toxicological properties of the Australian essential oil,lemon myrtle,(backhousia citriodora) were investigated. Lemon myrtle oil was shown to possess significant antimicrobial activity against the organisms Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, Candida albicans,methicillin-resistant S. aureus (MRSA), Aspergillus niger, Klebsiella pneumoniae and Propionibacterium acnes comparable to its major component-citral. An in vitro toxicological study based on the MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) cytotoxicity assay was performed. In vitro cytotoxicity testing indicated that both lemon myrtle oil and citral had a very toxic effect against human cell lines: HepG2 (a hepatocarcinoma-derived cell line); F1-73 (a fibroblast cell line derived from normal skin) and primary cell cultures of human skin fibroblasts. Cytotoxicity IC 50 (50% inhibitory concentration) values ranged from 0.008 to 0.014% (w/v) at 4 h to 0.003 0.012% (w/v) at 24 h of exposure. The no-observed-adverse-effect level (NOAEL) for lemon myrtle oil was calculated as 0.5 mg/l at 24 h exposure and the RfD (reference dose) was determined as 0.01 mg/l. A product containing 1% lemon myrtle oil was found to be low in toxicity and could potentially be used in the formulation of topical antimicrobial products. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Backhousia citriodora; Lemon myrtle; Citral; In vitro cytotoxicity; Essential oils; Antimicrobial activity 1. Introduction Essential oils are used extensively in pharmacy,medicine,food,beverages,cosmetics,perfumery and aromatherapy. The increased usage of essential oils worldwide has raised a number of concerns in relation Abbreviations: ADI,allowable daily intake; ATCC,American Type Culture Collection; DMEM,Dulbecco s modified Eagle s medium; FDA,Food and Drug Administration; GCMS,gas chromatography/mass spectrometry; IC 50,50% inhibitory concentration; MIC, minimum inhibitory concentration; MTS,(3-(4,5-dimethylthiazol-2- yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2h-tetrazolium); NOAEL,no-observed-adverse-effect level; NCTC,National Culture Type Collection; RfD,reference dose; RPMI-1640,Roswell Park Memorial Institute Medium (tissue culture); UNSW,University of New South Wales; UWSH,University of Western Sydney Hawkesbury; ZOI,zone of inhibition. * Corresponding author. Tel.: +61-2-9385 4200; fax: +61-2-9385-6190. E-mail address: amanda.hayes@unsw.edu.au (A.J. Hayes). to adverse health effects which need to be addressed (Woolf,1999; Thompson and Wilkinson,2000). Australian endemic plants produce a wide range of essential oils such as,tea tree oil (oil of Melaleuca alternifolia),which is used in consumer health products including topical antiseptics,mouthwashes and acne treatments (Beer,1987; Altman,1989; Bassett et al., 1990; Saller et al., 1998). The broad spectrum antimicrobial activity,chemistry and in vitro cytotoxicity of tea tree oil has been well documented (Swords and Hunter,1978; Southwell et al., 1993; Carson and Riley, 1994a; Soderberg et al.,1996; Hayes et al.,1997). The plant genus Backhousia is endemic to eastern Australia and belongs to the family Myrtaceae (Brophy et al.,1995). B. citriodora was first described by the German firm Schimmel and Co. in 1888 from a specimen found in the rainforests near Imbill,southeastern Queensland. The most common chemotype B. citriodora F. Mueller (oil of lemon myrtle) is a highly aromatic shrub which contains predominantly citral 0278-6915/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0278-6915(01)00103-X

536 A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 (3,7-dimethyl-2-7-octadienal) with two main isomeric aldehydes: neral and geranial (Penfold et al.,1951; Brophy et al.,1995). Within the last decade,the essential oil,foliage and fruits of lemon myrtle have found wide usage as ingredients for use in perfumes,food flavourings and herbal teas (Taylor,1996). The standardisation of this essential oil has recently occurred with the release (Standards Australia,July 2001) of the Australian Standard 4941-2001 for the oil of B. citriodora,citral type (lemon myrtle oil). This standard stipulates a minimum concentration of 85% citral in the oil. Within this standard there is no reference to antimicrobial activity or toxicological/ safety aspects of the oil. Preliminary research has demonstrated that lemon myrtle oil possesses significant germicidal activity (Penfold and Grant,1925) and some antimicrobial activity (Atkinson and Brice,1955). However,a more thorough investigation into the antimicrobial activity of the oil has yet to be performed. Citral,the major component of the oil,was shown to exhibit broad spectrum antimicrobial activity (Stevens et al.,1970; Moleyar and Narasimham,1987,1992). In order to determine the potential antimicrobial properties of lemon myrtle oil, an investigation using microbiological screening against a range of Gram positive and Gram negative bacteria, yeast and mold was undertaken using the MIC (minimum inhibitory concentration) agar dilution and disc diffusion assays. The complete toxicity profile of lemon myrtle oil has yet to be determined. The dermal LD 50 value of citral in rabbits was reported as 2.25 g/kg (Moreno,1974) and the oral LD 50 value for rats was reported as 4.96 g/kg with depression followed by death within 4 h to 4 days (Jenner et al.,1964). In view of future consumer usage of lemon myrtle oil there is considerable interest in assessing potential health and safety risks. In vitro cytotoxicity has been readily accepted as a routine method of non-animal-based toxicity testing (Anderson and Russell,1995). In this study the in vitro cytotoxicity of lemon myrtle oil and its major component citral were reported using the MTS in vitro cytotoxicity assay against a range of human cells. The characteristics of exposure to a substance and the spectrum of responses can be described in a correlative relationship referred to as the dose response relationship (Loomis and Hayes,1996). In risk assessment it is important to estimate exposures to humans which are below the experimentally observable range of responses in any test assay. The NOAEL can be used as a basis of risk assessment and/or for calculating the reference dose (RfD) of test compounds (Rees and Hattis,1994). This study reports on the extrapolation of NOAEL and RfD values for lemon myrtle oil in an attempt to determine the toxicity to human cells and to generate potential consumer safety data. 2. Materials and methods 2.1. Test compounds The lemon myrtle oil tested was supplied by Mr. John Prince (Currumbin Valley,Queensland,Australia). Tea tree oil (Melaleuca alternifolia) was supplied by Main Camp Tea Tree Oil (Ballina,NSW,Australia). Citral standard was supplied by Haarmann and Reimer (Germany). Baseline standards for the in vitro cytotoxicity assay were mercuric chloride and aspirin (acetylsalicylic acid) purchased from Sigma (USA). All solvents and reagents were of analytical grade or higher. 2.2. Essential oil composition The chemical composition of each essential oil was determined by chiral phase gas chromatography mass spectrometry (GCMS) using a Hewlett-Packard 5890 Series II GC coupled to the ion source of a Hewlett- Packard HP-5971A Mass Selective Detector (MSD). The chromatograph was equipped with a J&W b-cyclodextrin capillary column (30 m0.25 mm ID0.25 mm film thickness) and data were acquired under the following conditions: initial temperature 70 C for 2 min; program rate 5 C/min; final temperature 200 C for 10 min; injector and detector temperatures 200 C; carrier gas N 2 at 10 psi and a split ratio of 1:20. The MS was run in the electron ionisation mode with the ion source at 190 C,using an ionisation energy of 70eV and a scan rate of 0.9 scans/sec confirmed by a mass spectral library database and Kovats indices (R 1 ) calculated against external hydrocarbon standards. 2.3. Test microorganisms The strains of bacteria used in this study were Staphylococcus aureus (NCTC 4163),methicillin-resistant S. aureus (clinical isolate), Propionibacterium acnes (UNSW 642200), Escherichia coli (NCTC 8196), Pseudomonas aeruginosa (NCTC 6750) and Klebsiella pneumoniae (UNSW isolate). Strains of fungi used were Candida albicans (ATCC 10231) and Aspergillus niger (UWSH isolate). 2.4. Bacterial culture methods All bacteria (excluding P. acnes) were subcultured from Nutrient agar (Oxoid) stock cultures stored at 2.5 C into Iso-sensitest broth (Oxoid) incubated at 37 C for 24 h. The resulting broth cultures were twice passaged at 16 18 h in fresh Iso-sensitest broth and used as the inoculum for the MIC assays. P. acnes were subcultured from brain heart infusion agar (Oxoid)+1% glucose (BDH Chemicals) stock cultures stored at 2.5 C into brain heart infusion broth (Oxoid)+1% glucose

A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 537 incubated at 37 C for 48 120 h under anaerobic conditions using the BBL anaerobic system: gas jar,gas pak and disposable anaerobic indicator (Becton Dickinson). Subculturing was repeated at least twice into fresh broth+1% glucose under the same conditions and used as inoculum in the disc diffusion assays. 2.5. Fungi culture methods The same method (as per bacterial cultures) was used for the yeast culture (C. albicans) except that nutrient agar was replaced with malt extract agar (Oxoid). The mold stock culture (A. niger) was subcultured from malt extract agar stock cultures stored at 2.5 C into Sabouraud dextrose agar (Oxoid) plates and incubated at 37 C until a thick mycelium was formed (usually 48 h). The mycelium was introduced into 0.1% bacteriological peptone solution using an inoculating loop. The resulting suspension was shaken vigorously and used as the inoculum for microbiological analysis. 2.6. Standardisation of all inocula Optical density was measured on all prepared inocula in a Milton Roy Spectronic 20D. All inocula were standardised at a wavelength of 600 nm (420 nm for C. albicans) by the addition of appropriate sterile broth. Cell densities were estimated from standard growth curves. 2.7. Internal standards and controls Medium was always prepared on the day of use. Appropriate controls were set up which involved: (1) a negative control involving the presence of test material but absence of organism; and (2) a positive control involving the absence of test material but presence of the organism. Microbiological methods are labour intensive and only a limited number of test substances can be tested on any one day. In each microbiological test,tea tree oil was used as a control to firstly compare the activity of lemon myrtle oil or standards and secondly to determine the reproducibility of the test procedure. 2.8. MIC agar dilution assay Test materials were added aseptically to 15-ml aliquots of appropriate sterile molten agar containing polyoxyethylenesorbitan monolaurate (Tween 20; Sigma 0.5%,v/v),at the appropriate volumes to produce the required concentration range of test material (0.03 2.0%,v/v). The resulting agar solutions were vortexed at high speed for 15 s or until completely dispersed,immediately poured into sterile petri dishes then allowed to set for 30 min. The plates were then spot inoculated by pipetting the desired test organism (3 ml) from the prepared inoculum onto the plates. Inoculated plates were left to stand until the inoculum had set (approximately 1 h) and then incubated at either 37 C for 24 h (all bacteria and C. albicans) or27 C for 24 48 h(a. niger). Following the incubation period,plates were observed and recorded for presence (+) or absence ( ) of growth. From these results the MIC was recorded as the lowest concentration of test material where absence of growth was recorded. Each test was repeated in triplicate on at least two separate occasions. 2.9. Disc diffusion assay A modification of the disc diffusion assay was chosen to determine the antimicrobial activity of test samples against the anaerobic organism P. acnes. Aliquots (15 ml) of molten brain heart infusion agar (Oxoid)+1% glucose were poured into sterile petri dishes and allowed to set to form an agar base. A prepared inoculum (100 ml) was added to sterile molten agar (5 ml) in a M c Cartney bottle,mixed and poured over the surface of the agar base and left to set (approximately 30 min). Using a sterile hole-punch,a hole (13 mm) was cut from the centre of the agar plate. A 13-mm sterile paper disc (Schleicher & Schvell,West Germany) was then placed in the resulting reservoir and the disc was impregnated with a quantity of test material (20 ml) using a micropipette. The plates were left to stand for approximately 30 min. Plates were then inverted and incubated at 37 C for 48 96 h under anaerobic conditions as previously described. Following incubation,zoi (zones of inhibition) were measured (mm) and recorded as the meansd (standard deviation). Each test was performed in triplicate on at least two separate occasions. 2.10. Statistical analysis Reproducibility of the MIC and disc diffusion results was determined using Grubbs Test for outlying results. All results were determined to be within the 95% confidence level for reproducibility. 2.11. Primary cell cultures and cell lines Primary cell cultures of human skin fibroblasts were obtained from fresh skin biopsies taken from the arm of healthy individuals (Cytogenetics Department,The New Children s Hospital,Westmead,Australia). Permanent cell lines included those of HepG2 cells derived from a human hepatocarcinoma (ATCC No. HB8065) and F1-73 cells,an epithelial cell line derived from a human skin biopsy (ATCC No. CRL 7922). 2.12. Culture methods All cultures were maintained in a colour-free medium composed of 50% Dulbecco s modified Eagle s medium

538 A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 (DMEM,Sigma); 50% RPMI-1640 Medium (Roswell Park Memorial Institute Medium-tissue culture,sigma); supplemented with 5% fetal bovine serum (Life Technologies) or 5% newborn calf serum (Trace); and containing 0.5% of antibiotic mix: 200 mm l-glutamine, 10,000 units penicillin and 10 mg streptomycin per ml (Sigma). Cultures were maintained at 37 C in a humidified 5% CO 2 incubator. The adherent cells were removed from the tissue culture surface by adding Trypsin/EDTA (Life Technologies) once the cells had almost reached confluence. Cells were in the exponential growth phase at the time of testing,with experiments conducted at concentrations of 250,000 500,000 cells per ml. The viability of the cells exceeded 95% on all occasions as determined by the Trypan Blue (trypan blue solution 0.4%; Sigma) dye exclusion method. 2.13. In vitro cytotoxicity testing The MTS cytotoxicity/proliferation assay (Promega, 1996) was used to measure the toxicity of test materials by determining the number of viable cells in culture. Two exposure periods of 4 and 24 h were chosen for determining the in vitro cytotoxicity of the test material. 2.14. Baseline standards and controls Baseline standards of mercuric chloride (Sigma M- 6529) and aspirin (acetylsalicylic acid; Sigma A-5376) were included so that toxicity ranking of test samples against a known highly toxic substance and a moderately toxic substance could be performed,respectively,as well as,providing internal standards between experiments. Aspirin is available as an over-the-counter drug and is regarded as a moderately safe compound with extensive toxicity data available on its use. Aspirin has an oral LD 50 value of approximately 1.3 g/kg in mice (Martindale,1982). Mercuric chloride is described as a violent poison and highly toxic,with 1 2 g often resulting in human fatalities (Martindale,1982). It has an oral LD 50 value of 1 mg/kg in rats (Richardson,1992). These baseline standards were also used to determine the reproducibility within each of the experimental tests undertaken. Two internal controls were set up for each experiment at both exposure periods: (1) an IC 0 consisting of cells only; and (2) an IC 100 consisting of media only. Background absorbances due to the non-specific reaction between test compounds and the MTS reagent was deducted from exposed cell values as described below. In general,this background absorbance was only observed at high concentrations of some test materials. 2.15. Sample preparation Test material included lemon myrtle oil,citral,tea tree oil,1% lemon myrtle oil product and baseline standards of mercuric chloride and aspirin. Each test material was dissolved in 96% ethanol and filtered sterilised (0.22 mm) to give a stock solution with a final concentration of 10 50% (w/v). This stock solution was stored in a sterile amber glass vial with a teflon seal. Serial dilutions of each test material were prepared in cell culture medium from this stock solution to give test concentrations in the range of 0.0001 10%. Aliquots (900 ml) of each of these serial dilutions were dispensed into four separate sterile round-bottomed culture tubes (Falcon). A suspension of cells (100 ml) was added to two of these tubes with a concentration of 2.5 5.010 6 cells per ml representing the 4 and 24 h exposure periods. The remaining two tubes were designed as blanks (no cell suspension was added) to determine any background absorbance of the test material dilutions at both the exposure periods. The 100 ml suspension of cells was substituted with an additional 100 ml of sterile culture medium. All other experimental conditions were identical to the dilutions containing cells. 2.16. Cytotoxicity procedure At time zero for the 4 h exposure period and 4 h prior to the end of the 24 h exposure period,tubes were removed from the incubator and the desired quantities of MTS (Promega) and PMS (phenazine methosulfate, Sigma) reagents (ratio 20:1) were added to each tube in accordance with the manufacturer s instruction (Promega,1996). Tubes were then incubated at 37 C for the remaining 4 h of the exposure period. After the 4 h reaction period,aliquots (4100 ml) of the coloured reaction product (formazan) were pipetted into 96-well flat-bottomed microtiter tissue culture plates from each of the tubes. The absorbance was measured in a Multiskan MS (Labsystems,Finland) microtitre platereader at 492 nm. Dose response curves were plotted for test materials and controls after correction by subtracting the background absorbance from that of the blanks. IC 50 values (50% inhibitory concentration) were extrapolated graphically from plotted absorbance data. IC 50 values (%,w/v) were expressed as the mean and the % S.D. of a minimum of three repeat experiments performed on each test compound. 2.17. Risk assessment calculations of NOAEL and RfD from in vitro cytotoxicity data Using the data obtained from the IC 50 values and the corresponding dose response curves,noael values were determined for lemon myrtle oil and controls of aspirin and mercuric chloride against each of the cell lines and cultures tested. The RfD was then calculated from the extrapolation of these NOAEL values with the inclusion of additional safety factors (Rees and Hattis, 1994; Faustman and Omenn,1996). These safety factors

A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 539 included a modifying factor which was used to account for differences in routes of administration due to the use of human cells and their extrapolation to the human body as a whole entity (5 was used),and a human variability factor that would account for differences in sensitivity among the human population (10 was used). Based on the multiplication of these safety factors (total 50),RfDs were calculated for the most sensitive cell lines at both the 4 and 24 h exposure periods. 3. Results 3.1. Chemical composition of lemon myrtle oil Percentage compositions of major components of lemon myrtle oil and tea tree oil (as determined by chiral phase GCMS) are summarised in Table 1. Lemon myrtle oil had a total citral concentration of 96.6% (consisting of neral,geranial,iso-citral and trans-citral) while tea tree oil had a terpinen-4-ol concentration of 42.8% and a 1,8-cineole concentration of 4.6% The 100% citral standard used as a control had a neral concentration of 36.0% and a geranial concentration of 64.0%. 3.2. Antimicrobial activity of lemon myrtle oil MIC results indicated that lemon myrtle oil and 100% citral showed comparable activity against organisms tested. S. aureus was the only organism tested that showed greater susceptibility to citral (Table 2). These MIC results also indicated that lemon myrtle oil was considerably more active than tea tree oil against all organisms tested. Both lemon myrtle oil and citral inhibited P. aeruginosa at 2.0% (v/v),while tea tree oil did not inhibit P. aeruginosa at the highest concentration tested (2.0%,v/v). The lowest minimum inhibitory concentration of lemon myrtle oil was 40.03% (v/v) against C. albicans and E. coli. ZOI results indicated that lemon myrtle oil was effective at inhibiting P. acnes given by the largest ZOI (66.23.1 mm) followed by 100% citral (58.5 2.8 mm) and then tea tree oil (22.51.2 mm). Table 1 Major components and composition (as %,v/v) of lemon myrtle oil and tea tree oil Kovats index (R 1 ) Compound Lemon myrtle oil Tea tree oil 992/996 ()-a-pinene 1.8 1059 a-terpinene 8.6 1082 p-cymene 3.0 1022 b-myrcene Trace 1059 a-terpinene 8.6 1082 p-cymene 3.0 1089 5-Methyl-6-en-2-one 0.5 1104 g-terpinene 18.2 1120 1,8-Cineole 4.6 1130 a-terpinolene 3.0 1247/1250 ()-Linalool Trace 1260 Citronellal 0.5 1277 Cis-iso citral 1.7 1294 Trans-iso citral 2.6 1327/1331 ()-Terpinen-4-ol - 42.8 1372/1375 ()-a-terpineol 3.5 1379 Neral 40.9 1417 Geranial 51.4 1428 Trans-geraniol 0.7 Table 2 MIC values (as %,v/v) for lemon myrtle oil and tea tree oil Organisms Lemon myrtle oil Tea tree oil Citral S. aureus 0.05 0.2 0.03 E. coli 0.03 0.2 0.03 P. aeruginosa 2.0 >2.0 2.0 C. albicans 0.03 0.2 0.03 A. niger 0.1 0.4 0.1 MRSA 0.2 0.3 0.2 K. pneumoniae 0.2 0.3 0.2

540 A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 3.3. Antimicrobial activity of lemon myrtle oil and tea tree oil mixtures This study focused on the antimicrobial activity of mixtures of lemon myrtle and tea tree oils. MIC results (Table 3) showed that as the concentration of lemon myrtle oil was increased in the overall oil composition, the antimicrobial activity increased accordingly. Major improvements in antimicrobial activity were evident when the concentration of lemon myrtle oil was increased to 20% or higher,oil numbers 4 7 (i.e. a ratio of 1:5,lemon myrtle: tea tree oil). Minor differences in the antimicrobial activity of oil number 6 (20% tea tree oil:80% lemon myrtle oil) and oil number 7 (100% lemon myrtle oil) were observed. P. aeruginosa was the only organism in the experiment that showed greater susceptibility to 100% lemon myrtle oil. 3.4. In vitro cytotoxicity of lemon myrtle oil, tea tree oil and citral using human cells The MTS in vitro cytotoxicity assay was used to compare the toxicity of lemon myrtle oil,citral and tea tree oil. IC 50 values calculated at 4 and 24 h exposure periods are summarised in Tables 4 and 5. Human skin cells (F1-73) at both exposure periods were the more resilient against all test compounds. The overall general rating for sensitivity of the primary cultures and cell lines tested to the test compounds was HepG2>skin fibroblasts > F1-73. The mean IC 50 range of lemon myrtle oil in the primary cell culture and cell lines tested was 0.008 0.014% (w/v) at 4 h to 0.004 0.010% (w/v) at 24 h. These results were similar to those obtained for citral where the IC 50 range at 4 h was 0.009 0.016% (w/v) and 0.004 0.011% Table 3 MIC values (as %,v/v) of tea tree oil and lemon myrtle oil blends % Composition Organisms Oil no. Tea tree oil Lemon myrtle oil S. aureus E. coli P. aeruginosa C. albicans 1 100 0 0.2 0.2 >10 0.2 2 95 5 0.2 0.2 5 10 0.2 3 90 10 0.2 0.2 5 10 0.2 4 80 20 0.2 0.2 5 10 0.1 5 50 50 0.1 0.1 5 10 0.03 6 20 80 0.05 0.05 3.0 0.03 7 0 100 0.05 0.03 2.0 0.03 Table 4 IC 50 values (as %,v/v) for tea tree oil,lemon myrtle oil,lemon myrtle oil product (1%),citral and baseline standards following an incubation period of 4 h using skin and liver derived cells Cell line Tea tree oil Lemon myrtle Citral Lemon myrtle product (1%) Aspirin Mercuric chloride F1-73 0.05645 0.01440 0.01655 0.6523 0.2165 0.00170 a Skin Fibroblasts 0.02460 0.01235 0.01140 0.2025 0.2450 0.000975 a HepG2 0.04345 0.00840 0.00935 Not tested 0.1545 0.00165 a *Indicates estimated IC 50 values extrapolated graphically due to absorbance values below test limits. a IC 50 values given represent the mean and the SD as percentages of a minimum of three repeat experiments with six measurements per dilution performed on each test material. Table 5 IC 50 values (as %,v/v) for tea tree oil,lemon myrtle oil,lemon myrtle oil product (1%) citral and baseline standards following an incubation period of 24 h against skin and liver derived cells Cell line Tea tree oil Lemon myrtle Citral Lemon myrtle product (1%) Aspirin Mercuric chloride F1-73 0.03250 0.01055 0.01165 0.2520 0.2255 0.00150 a Skin Fibroblasts 0.02055 0.00750 0.00550 0.1461 0.1965 0.000885 a HepG2 0.01970 0.00450 0.00455 Not tested 0.1450 0.000675 a IC 50 values given represent the mean and the SD as percentages of a minimum of three repeat experiments with six measurements per dilution performed on each test material. a Indicates estimated IC 50 values extrapolated graphically due to absorbance values below test limits.

A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 541 Table 6 NOAEL values (mg/l) for lemon myrtle oil,citral,aspirin and mercuric chloride against human cell lines following an incubation period of 4h HepG2 Skin fibroblasts F1-73 Lemon myrtle oil 5 5 5 Citral 5 5 5 Aspirin 10 10 10 Mercuric chloride 0.1 0.1 0.1 Table 7 NOAEL values (mg/l) for lemon myrtle oil,citral,aspirin and mercuric chloride against human cell lines following an incubation period of 24 h HepG2 Skin fibroblasts F1-73 Lemon myrtle oil 0.5 0.5 0.5 Citral 0.5 0.5 0.5 Aspirin 1 10 10 Mercuric chloride 0.1 0.1 0.1 (w/v) at 24 h. While tea tree oil had an IC 50 range of 0.024 0.056% (w/v) at 4 h and 0.019 0.032% (w/v). IC 50 values indicated that lemon myrtle oil and citral were more cytotoxic than tea tree oil against cell lines HepG2,F1-73 and skin fibroblasts at both exposure periods. All test compounds showed an increase in toxicity (a decrease in IC 50 values) as exposure time increased. With respect to controls,the overall rating for toxicity of test material was mercuric chloride > lemon myrtle oilfficitral > tea tree oil > aspirin. 3.5. In vitro cytotoxicity of the 1% lemon myrtle oil product using human skin cells Results of Tables 4 and 5 indicate that a 1% lemon myrtle oil product was considerably less toxic than pure lemon myrtle essential oil against human skin cells. The mean IC 50 value range for the 1% lemon myrtle oil product was 0.20 0.65% (w/v) at 4 h and 0.14 0.25% (w/v) at 24 h exposure compared to the neat lemon myrtle oil which had a mean IC 50 value range of 0.012 0.014% (w/v) at 4 h and 0.007 0.010% (w/v) at 24 h. With respect to controls,the overall rating for toxicity against skin fibroblasts was mercuric chloride > lemon myrtle oil>1% lemon myrtle oil product>aspirin and against skin cells F1-73 was mercuric chloride > lemon myrtle oil > aspirin> 1% lemon myrtle oil product. 3.6. Risk assessment calculations for lemon myrtle oil NOAELs for lemon myrtle oil,citral and controls against HepG2,F1-73 and skin fibroblasts were calculated (Tables 6 and 7). Results indicated that the lowest obtained NOAEL for lemon myrtle oil at 4 h exposure was 5 mg/l given by HepG2 cells,while at 24 h exposure all cell types gave a NOAEL of 0.5 mg/l. These results were comparable to those obtained for citral. With respect to controls,mercuric chloride gave the lowest NOAEL of 0.1 mg/l against all cell types at 4 and 24 h exposure and aspirin gave the highest NOAEL ranging from 1.0 to 10.0 mg/l at 4 to 24 h exposure. The RfD values for 4 and 24 h (Table 8) ranged from 0.1 to 0.01 mg/l for lemon myrtle oil and citral against the most sensitive cell lines. Mercuric chloride gave the lowest RfD value for both 4 and 24 h exposure of Table 8 RfD values (mg/l) for lemon myrtle oil,citral,aspirin and mercuric chloride against the most sensitive cell lines at 4 and 24 h exposure 0.002 mg/l and aspirin gave the highest RfD value of 0.2 0.02 mg/l at 4 to 24 h exposure. 4. Discussion 4 h 24 h Lemon myrtle oil 0.1 0.01 Citral 0.1 0.01 Aspirin 0.2 0.02 Mercuric chloride 0.002 0.002 The results of microbiological tests indicate that lemon myrtle oil was significantly more active than tea tree oil against the broad range of microorganisms tested. Citral,the major component of lemon myrtle oil, was shown to have comparable antimicrobial activity to that of pure lemon myrtle oil. Thus,it can be deduced that the minor components of lemon myrtle oil such as 5-methyl-6-en-2-one and citranellal had little effect on the overall antimicrobial activity of the oil. Lemon myrtle oil inhibited all organisms (excluding P. aeruginosa) at concentrations of 40.2%,v/v). The strain of P. aeruginosa used in this study was an extremely resistant strain of Pseudomonas,indicated by the highest MIC values. In some studies P. aeruginosa has been shown to be more resistant to tea tree oil at concentrations as high as 8% (Carson and Riley,1994b),while lemon myrtle oil and citral inhibited P. aeruginosa at 2.0% (v/ v) in our study. Overall results indicate that lemon myrtle oil has potential usage as an antiseptic or antimicrobial agent in the treatment of skin microorganisms associated with cuts,bites,acne and tinea. For lemon myrtle oil to be used as an effective topical antimicrobial,a toxicity safety assessment is required. Skin sensitisation has been reported for the topical application of 100% citral at concentrations greater than 1%. However,researchers have shown that the addition of terpenes such as a-pinene and d-limonene to citral in the ratio of 1:4 prevents a sensitisation reaction occurring (Opdyke,1976). Results of this current study

542 A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 suggest that the inclusion of tea tree oil blended with lemon myrtle oil (ratio 1:4) would improve the microbiological activity of the tea tree oil in the formulation substantially,and potentially reduce any sensitising effects of lemon myrtle oil. This quenching effect has also been reported for lemongrass oil which contains approximately 85% citral (Opdyke,1976). Further studies in this area are needed. The MTS in vitro cytotoxicity assay was used in this study to determine and compare the toxicity of lemon myrtle oil,citral and tea tree oil against a selection of human cell lines and primary cultures at 4 and 24 h exposure. As the potential use of lemon myrtle oil may be for topical application,skin fibroblasts and the human skin cell line F1-73 were chosen for this study. HepG2,the liver derived cell line,was also used as it has been shown to be more susceptible to damage by essential oils such as tea tree oil (Hayes et al., 1997). On comparison with the numerical toxicity rating scheme (Gosselin et al.,1984),the mean IC 50 values obtained in this study for 100% lemon myrtle and citral would be rated as very toxic (rating of 4) while tea tree oil would be rated as moderately toxic (rating of 3). Using the estimate of 1 litre of cells being equivalent to 1 kg of body weight,the NOAEL value for lemon myrtle oil at 24 h with respect to HepG2 cells estimates that the average adult (ffi70 kg) could ingest 35 mg of lemon myrtle oil and the average child (ffi10 kg) 5 mg without any observed adverse effects. Owing to the high toxicity of lemon myrtle oil as indicated by these in vitro cytotoxicity results,it is not advisable that high concentrations of the oil be applied directly to human skin. A product containing 1% lemon myrtle oil was tested in this study and found to be substantially less toxic to primary skin fibroblasts and skin cells than neat lemon myrtle oil at both 4 and 24 h of exposure. As no published toxicity data on lemon myrtle oil could be sourced,animal toxicity data on citral was used as an estimate for comparing in vivo and in vitro toxicity data of lemon myrtle oil. The oral LD 50, of citral in vivo of rats has been previously reported as 0.50% (w/w),while the maximum non-lethal dose in mice was 0.09% (w/w) orally and 0.03% (w/w) intravenously (Jenner et al.,1964). The extrapolation of these data may not be representative of true human toxicity. The IC 50 results obtained in this study (Tables 4 and 5) suggested that the in vitro cytotoxicity testing using human cells showed greater sensitivity to citral than in vivo toxicity using rats (greater than 100-fold) or mice (greater than 10-fold). Safety estimates (RfD values) for citral calculated in this study (as shown in Table 6) are 5 50 times lower in dose than the allowable daily intake (ADI) value of 0.5 mg/kg established for citral by the FDA (Food and Drug Administration,USA). There are several limitations in the extrapolation of in vitro experimentation data for use in safety assessments. In vitro toxicity experimentation cannot duplicate exactly the biodynamics of the human body. The increased sensitivity of the in vitro experiments compared to previous animal in vivo studies was probably due to the accumulation of metabolites in the cell culture system and the lack of biotransformation and excretion pathways for their elimination in vitro,as would normally occur in vivo. Because of these differences,in vitro experimentation may at times overestimate the toxicity of test substances. However,this increased sensitivity may also be due to physiological differences between laboratory animals and humans. For example,human skin has been found to be more reactive to irritants than rabbit skin and shows a much greater variety of inflammatory responses (Higmet et al.,1990). In conclusion,we have demonstrated that lemon myrtle essential oil (oil of B. citriodora) and its major component citral,possess significant antimicrobial activity against a range of Gram positive and Gram negative bacteria,yeast and mold given by an MIC range of 0.03 2.0% (v/v). Preliminary toxicity studies using an in vitro cytotoxicity assay suggest that neat lemon myrtle oil and citral are toxic to human liver derived cells (HepG2),skin cells (F1-73) and skin fibroblasts. A product containing 1% lemon myrtle oil was significantly less toxic to human skin cells and skin fibroblasts. More studies are needed to assess the toxicity of lemon myrtle oil such as the mechanisms by which the oil may penetrate the human skin barrier and the potential effects it may have at the cellular level. This information would not only be a benefit for formulating potential products containing lemon myrtle oil but also could help to elucidate the mechanisms of essential oil toxicity in general. Acknowledgements We would like to thank Blackmores Ltd for part funding of the project and Dr. Z.H. Wu. Cytogenetics, Westmead Hospital,Sydney,for supplying the human cell cultures. References Altman,P.M.,1989. Australian tea tree oil a natural antiseptic. Australian Journal of Biotechnology 3,247 248. Anderson,D.,Russell,T.,1995. The Status of Alternative Methods in Toxicology. The Royal Society of Chemistry,Cambridge. Atkinson,N.,Brice,H.E.,1955. Antibacterial substances produced by flowering plants 2. The antibacterial action of essential oils from some Australian plants. Australian Journal of Experimental Biology 33,547 554. Bassett,I.B.,Pannowitz,D.L.,Barnetson,R.S.C.,1990. A comparative study of tea tree oil versus benzoyl peroxide in the treatment of acne. The Medical Journal of Australia 153,455 458.

A.J. Hayes, B. Markovic / Food and Chemical Toxicology 40 (2002) 535 543 543 Beer,C.,1987. Australian tea tree oil. Nature and Health 6,3 7. Brophy,J.J.,Goldsack,R.J.,Fookes,C.J.R.,Forster,P.I.,1995. Leaf oils of the genus Backhousia (Myrtaceae). Journal of Essential Oil Research 7,237 254. Carson,C.F.,Riley,T.V.,1994a. The antimicrobial activity of tea tree oil. The Medical Journal of Australia 160,236. Carson,C.F.,Riley,T.V.,1994b. Susceptibility of Propionibacterium acnes to the essential oil of Melaleuca alternifolia. Letters in Applied Microbiology 19,24 25. Faustman,E.M.,Omenn,G.S.,1996. Risk assessment. In: Klaassen, C.D. (Ed.),Casarett and Doull s Toxicology. The Basic Science of Poisons,5th Edition. McGraw-Hill,New York,pp. 75 88. Gosselin,R.E.,Smith,R.P.,Hodge,H.C.,1984. Clinical Commercial Toxicology of Commercial Products. Williams and Wilkins,Baltimore,MD. Hayes,A.J.,Leach,D.N.,Markham,J.L.,Markovic,B.,1997. In vitro cytotoxicity of Australian tea tree oil using human cell lines. Journal of Essential Oil Research 9,575 582. Higmet,S.,Dorko,J.D.,Herzog,N.J.,Barrow,C.S.,1990. A comparison of skin irritation in rabbits versus humans. Toxicologist 19, 152 156. Jenner,P.M.,Hagan,E.C.,Taylor,J.M.,Cook,E.L.,Fitzhugh,O.G., 1964. Food flavourings and compounds of related structure I. Acute oral toxicity. Food and Cosmetics Toxicology 2,327 343. Loomis,T.A.,Hayes,A.W.,1996. Loomis s Essentials of Toxicology, 4th Edition. Academic Press,CA. Martindale,1982. The Extra Pharmacopoeia,28th Edition. Pharmaceutical Press,London. Moleyar,V.,Narasimham,P.,1987. Mode of antifungal action of essential oil components citral and camphor. Indian Journal of Experimental Biology 25,781 784. Moleyar,V.,Narasimham,P.,1992. Antibacterial activity of essential oil components. International Journal of Food Microbiology 16, 337 342. Moreno,O.M. Report to RIFM. 31 January 1974. Opdyke,D.L.J.,1976. Inhibition of sensitization reactions induced by certain aldehydes. Food and Cosmetics Toxicology 14,197 198. Penfold,A.R.,Grant,R.,1925. The germicidal values of some Australian essential oils and their pure constituents. Journal and Proceedings of the Royal Society of New South Wales 59,346 350. Penfold,A.R.,Morrison,F.R.,Willis,J.L.,McKern,H.G.,Spies, M.C.,1951. The occurrence of a physiological form of Backhousia citriodora F Muell. and its essential oil. Journal of the Proceedings of the Royal Chemical Society of New South Wales 85,123 126. Promega,1996. CellTiter 96 TM AQ ueous Non-radioactive Cell Proliferation Assay. Technical Bulletin No. 169,revised. Promega Corporation,Madison,pp. 1 99. Rees,D.C.,Hattis,D.,1994. Developing quantitative strategies for animal to human extrapolation. In: Hayes,A.D. (Ed.),Principles and Methods of Toxicology,3rd Edition. Raven Press,New York. Richardson,M.,1992. The Dictionary of Substances and their Effects. The Royal Society of Chemistry,Cambridge. Saller,R.,Berger,T.,Reichling,J.,Harkenthal,M.,1998. Pharmaceutical and medicinal aspects of Australian tea tree oil. Phytomedicine 5,489 495. Soderberg,T.A.,Johansson,A.,Gref,R.,1996. Toxic effects of some conifer resin acids and tea tree oil and human epithelial and fibroblast cells. Toxicology 107,99 109. Southwell,I.A.,Hayes,A.J.,Markham,J.L.,Leach,D.N.,1993. The search for optimally bioactive Australian tea tree oil. Acta Horticulturae 334,265 275. Stevens,K.L.,Jurd,L.,King,A.D.,Mihara,K.,1970. The antimicrobial activity of citral. Experientia 27,600 602. Standards Australia,2001. Oil of Backhousia citriodora,citral Type (lemon myrtle oil). AS 4941-2001. Standards Australia International,Sydney. Swords,G.,Hunter,G.L.K.,1978. Composition of Australian tea tree oil. American Chemical Society 26,734 737. Taylor,R.,1996. Lemon myrtle,the essential oil. Rural-Research 172, 18 19. Thompson,K.F.,Wilkinson,S.M.,2000. Allergic contact dermatitis to plant extracts in patients with cosmetic dermatitis. British Journal of Dermatology 142,84 88. Woolf,A.,1999. Essential oil poisoning. Clinical Toxicology 37,721 727.