Antimicrobial Activity of Two Garlic Species (Allium Sativum and A. Tuberosum) Against Staphylococci Infection. In Vivo Study in Rats

Purpose: This study observed the effect of garlic extracts and amoxicillin against an induced staphylococcal infection model. MIC and MBC were also obtained for aqueous extracts of Allium sativum (Asa) and Allium tuberosum (Atu) against Staphylococcus aureus penicillin-sensitive (PSSA - ATCC 25923) and MRSA (ATCC 33592). Methods: Granulation tissues were induced in the back of 205 rats. After 14 days, 0.5 mL of 108 CFU/mL of PSSA or MRSA were injected inside tissues. After 24h, animals were divided: G1 (Control) – 0.5 mL of NaCl 0.9%; G2 – Asa 100 mg/kg or 400mg/kg; G3 – Atu 100 mg/kg or 400 mg/kg; G4 – amoxicillin suspension 50 mg/kg, considering PSSA infection; and G5 (Control) - 0.5 mL of NaCl 0.9%; G6 – Asa 400mg/kg; G7 – amoxicillin 50 mg/kg; and G8 - Asa 400 mg/kg + amoxicillin 50 mg/kg for MRSA. All treatments were administered P.O. every 6h. Animals were killed at 0, 6, 12 and 24h. Samples were spread on salt-mannitol agar. Colonies were counted after 18 h at 37 °C. Atu was not able to inhibit or kill PSSA and MRSA. Considering Asa, MIC and MBC against PSSA were 2 mg/mL and 4 mg/mL, respectively; and 16 mg/mL and 64 mg/mL against MRSA. Results: No effect was observed in vivo for control, Asa 100 mg/kg and Atu 100 mg/kg, while amoxicillin, Atu 400 mg/kg and Asa 400 mg/kg decreased PSSA counts in all-time points. No effect of any group against MRSA was observed at any time. Conclusion: Thus, A. sativum and A. tuberosum were able to reduce PSSA infection, but not MRSA infection.


Introduction
Among all staphylococci, Staphylococcus aureus is the most important cause of infectious diseases in humans. This microorganism is part of the commensal microbiota, causing opportunistic infections under appropriate conditions. 1 Besides, intrinsic virulence factors, life-threatening staphylococcal infections occur due to its ability to develop antibiotic resistance. 2 The antimicrobial resistance against penicillins is remarkable in the most known S. aureus strains. The minimum inhibitory concentration (MIC) of amoxicillin against penicillin-sensitive S. aureus strains is usually below to 0.5 µg/mL. 3 The beta-lactams' MIC against resistant strains could be hundred times higher than the usual MIC of sensitive strains. Thus, there is a clear need for novel antimicrobials or new methods to improve the efficacy of old antibiotics. Natural products are the most important source of innovative contributions to antimicrobial therapeutic. 4 The genus Allium is often used and studied in medicine as antibacterial and anti-viral agents among many other properties. 5 There are over 600 species of Allium found throughout the world, most of them in Asia. Most species have characteristic aroma and odor; some are used as ornamentals or spices. Garlic is often used in many forms (oils, powder, etc.), being also present in many pre-packaged foods, soups and bread used in a daily diet. 6 Among the species of Allium spp., A. tuberosum ("Chinese chive") is an important ingredient in Asian cuisine, being used as a medicinal herb for many disorders and diseases. 7 Both pressed juice and essential oil showed inhibition properties against a broad range of gram-positive and gram-negative microorganisms as well as many fungi. 8 Allium sativum contains several biologically active chemical constituents, such as alliin, allicin, alliinase, S-allylcysteine and other organosulfur compounds. 9 Alliin is a natural sulfoxide constituent of fresh garlic. It is derivate from the amino acid cysteine, being converted to allicin by the enzyme aliinase when bulbs are mashed. Allicin is a precursor of sulphur compounds, responsible for the odor and some of the pharmacological properties of garlic. Once exposed to the atmospheric air, allicin is converted into diallydissulphinate, which has antimicrobial activity. 10,11 Garlic extract and some of its products, such as diallyl sulphide and diallyl disulphide showed protective action against MRSA infection in mice. 12 The additive or complimentary effect of garlic or its products on some antibiotics against bacterial infections was previously demonstrated in vitro 13 and in vivo. 14 The aim of this study was to observe the effect of garlic extracts alone or in combination with amoxicillin on the staphylococcal infection, which was induced with penicillin-sensitive or penicillin-resistant strains, in rats.

Preparation of garlic extracts
Garlic bulbs (Allium sativum) and Chinese chive (Allium tuberosum) were supplied by Research Center for Chemistry, Biology and Agriculture -CPQBA-UNICAMP (Campinas, SP, Brazil). The dry peel around the bulbs was removed and they were used to prepare A. sativum extract. A. tuberosum extract were prepared using only the plant leaves. Bulbs are the most common part of A. sativum used by population and in most of studies. Considering A. tuberosum the few studies in literature used its leaves as a primary source for extracts. Both aqueous extracts were prepared by using 100 g of fresh rootless bulbs of A. sativum or 100 g of A. tuberosum leaves and 100 mL of distilled/deionized water. Both plants were ground in a blender for 10 min and the resulting extracts were filtered in qualitative paper filters, and sterilized through 0.2 µm membrane filters by using a vacuum pump. All residues were weighed, and the concentration of the final solution of both extracts was considered 25% (w/v) or 250 mg/mL, as previously described. 15,16

Analysis of extracts compounds
The extracts were analyzed using a gas chromatograph (Hewlett-Packard -HP-5890) with a mass-selective detector (Hewlett-Packard -HP-5971), equipped with DB-1 capillary column (25 m x 0.2 mm x 0.33 µm). Carrier gas was helium at a flow rate of 1.0 mL/min. The injection temperature was 240 °C and the detector was set at 300 °C. The column temperature program was 60 °C (2 min) to 300 °C at 4 °C/min. The injection volume was 1.0 µL and the detector was a quadropole system with ionization energy of 70 eV. Both extracts were assayed under the same conditions, using the NIST98 electronic library for GC/MS.

Bacteria
Both Staphylococcus aureus ATCC 25923 (penicillin sensitive -PSSA) and ATCC 33592 (methicillin resistant -MRSA) were used in the in vitro and in vivo models. Both strains were kept in TSA medium.

In vitro study Minimal inhibitory (MIC) and bactericidal (MBC) concentrations
Eleven culture tubes containing 5 mL of Mueller-Hinton broth were used for each strain. Two-fold progressive concentrations of both extracts were added in the first 10 tubes. The last tube had microorganisms only (positive control). An extra tube was used as negative control (without microorganisms or extracts All tubes received 250 µL of 10 6 CFU/mL of either PSSA or MRSA. Tubes were incubated in aerobic conditions at 37 °C. Bacterial concentration was adjusted by using a spectrophotometer. After 18 h, the culture medium was observed and the turbidity of medium was measured and compared with the negative-control tubes. The first tube without bacterial growth was considered as MIC. Amounts of 5 µL from the tubes without bacterial growth were spread on Petri dishes containing BHI agar. The first concentration without any bacterial growth on agar was considered as MBC.

In vivo study
This study was approved by the Ethical Committee of Animal Experimentation of Biology Institute-UNICAMP (protocol # 1457-1). Two assays were carried out. The first one was designed to compare both garlic extracts against the PSSA infection. The second one was designed to compare the effect of A. sativum, which was the most efficient formulation in assay 1, and amoxicillin against MRSA infection.

First assay
140 adult-male Wistar rats (Rattus norvegicus-albinus), 60 days of age and weighing 175 g ± 25 g, were obtained from CEMIB-UNICAMP (Centro de Bioterismo -ICLAS Monitoring/Reference Centre, Campinas, Brazil). Granulation tissue was induced by insertion of four sterilized polyurethane sponge discs (density = 35 kg/m 3 ) subcutaneously in the back of each animal, according to a method previously described. 17 These sponge discs (Proespuma Com. & Ind. Ltd., Sao Paulo, Brazil) were 12 mm in diameter and 5 mm thick, weighing 12 ± 1 mg, being all of them were previously sterilized. Animals were anesthetized by an injection of ketamine (40-87 mg/kg/I.P.) and xylasine (5-13 mg/kg/I.P.). After careful antisepsis with povidone-iodine, the back skin was incised and the sponge discs were carefully distributed considering a distance of approximately 2 cm among them, two in tail direction and two in head direction. Meperidine chlorhydrate (2 mg/kg/I.M.) was used to control post-operative pain. After 14 days of sponge insertion, a well-delimited granulation tissue was formed around and inside the sponge disks. This tissue was used as a scaffold for the staphylococcal infection. In order to establish the staphylococcal infection, a careful antisepsis was carried out in the back of all animals and 0.5 mL of a suspension of PSSA (10 8 CFU/mL) was injected into the four granulation tissues. After 24 hours of infection, the animals were submitted to the following treatments: 1) Control -20 animals received 0.5 mL of 0.9% NaCl. These animals were killed (n = 5 per group) right after the administration (t=0) and after 6, 12 and 24 hours after saline administration.

4.
A. sativum + amoxicillin -15 animals received both A. sativum extract 400 mg/kg and amoxicillin suspension 50 mg/kg. Both extract and amoxicillin suspension were administered P.O. every 6 hours. Animals were killed after 6, 12 and 24 hours after the first administration.

Granulation tissue removal
All infected granulation tissues were surgically removed after general anesthesia and blood removal by cutting carotid plexus. Tissue samples were individually disposed into assay tubes with 10 mL of 0.9% NaCl. Tubes were weighed before and after the tissue insertion. Samples were dispersed using an ultrasonic system (Vibra Cell 400W, Sonics & Materials Inc, Danbury, CT, USA) and 10 µL of the resulting suspension was spread on salt mannitol agar and incubated at 37 °C. Eighteen hours after incubation, the colonies were counted by using a manual colony counter.

Statistical analysis
Results were submitted to the Bartlett's test to assess variance homogeneity, and the Kolmogorov & Smirnov test to observe data distribution. Data of the first assay were transformed and analyzed by two-way ANOVA and Tukey's test adjusted for multiple comparisons (post hoc). Second assay data were submitted to Kruskal-Wallis' and Dunn's tests. GraphPad Prism 6.0 was used to analyze all data, considering 5% of a significant level. Figure 1 and (Table 1) show respectively the gas chromatography profile and the probable chemical composition of both A. tuberosum and A. sativum aqueous and hydroalcoholic extracts. The most abundant compound of the A. tuberosum aqueous extract was not identified by gas chromatography, being only 8% of its compounds identified. However, A. tuberosum hydroalcoholic extract had phytol as the most abundant compound (48.4%) followed by other acyclic compounds (23.4%) and only 1.8% of sulphides. A. sativum aqueous extract showed 82% of sulphur compounds, being disulphides the most prevalent ones. The hydroalcoholic extract of A. sativum showed trisulphides (31%) as the most abundant compound.

MIC and MBC
A. tuberosum was not able to inhibit or kill both resistant and sensitive staphylococci strains in MIC and MBC tests. A. sativum and amoxicillin inhibited (MIC) the penicillin-sensitive strain respectively at 2 mg/mL and 0.1 µg/mL. They killed (MBC) this strain at 4 mg/mL and 2 µg/mL, respectively. However, it was necessary 2048 µg/mL of amoxicillin to inhibit the MRSA strain and 8192 µg/mL to kill it. A. sativum was able to inhibit MRSA at 16 mg/mL, being able to kill it at 64 mg/mL.   Figure 2 shows the effect of treatments in the number of bacteria in all-time points. There are no significant differences (p>0.05) among periods for control, A. sativum 100 mg/kg and A. tuberosum 100 mg/kg. Amoxicillin decreased the CFU counts after 6 and 12 hours, but no differences were found between 24 hours and 6 or 12 hours. A. tuberosum 400 mg/kg decreased the bacterial counts starting from 6 to 24 hours and A. sativum showed a decrease after 6 and 24 hours. Comparison among groups in each period showed no significant differences among control, A. sativum 100 mg/kg and A. tuberosum 100 mg/kg, except for 6 hours period, when both aqueous solutions significantly decreased the CFU in comparison with control. Amoxicillin and A. sativum 400 mg/kg decreased CFU in comparison with control, A. sativum 100 mg/kg and A. tuberosum 100 mg/kg in each period, but they did not differ from A. tuberosum 400 mg/kg. This formulation also decreases CFU in comparison with control, A. sativum 100 mg/kg and A. tuberosum 100 mg/kg in all periods, except for 6 hours.  Figure 3 shows the effect of treatments against the resistant bacteria in all-time points. CFU increased after 12 h for control and amoxicillin + A. sativum, and after 6 hours for A. sativum 400 mg/kg and amoxicillin. Comparison among groups in the 6 hours period showed greater number of CFU for the amoxicillin group than the others groups, which did not differ one from each other. After 12 hours, control showed bigger number of CFU than amoxicillin alone or combined with A. sativum. No significant differences were observed in this period among A. sativum and the other groups. No significant differences were observed at the 24 hours period.

Discussion
Among many other models, the use of PVC or polyurethane sponges to induce a well-delimited infection was previously described for research on antistaphylococci substances. This model allowed a welldelimited staphylococcal infection development, since one of the most important phenomena for the infection establishment is bacterial surface adherence. The validity of the model was previously demonstrated for acute staphylococcal infection. 17 Most of the important infections results from bacterial adherence on some tissue surface, and the first line of the host defense against the bacterial invasion also request a surface to exert its defensive function. 18 Thus, the surface provided by the granulomatous tissue was suitable to both staphylococcal infection development and host defensive events. It is remarkable that amoxicillin was not able to kill all the PSSA microorganisms, and, thus, it did not eradicate the staphylococcal infection. This result is similar to the ones previously obtained using the same model. The main reason for this phenomenon is probably due to the penicillin mechanism of action. It is well stablished that bacteria growing in tissues surrounded by immune cells show slow reproduction, which decreases the penicillin activity. 17 As anticipated, the chemical composition differed between the two garlic species. The main chemical composition of A. tuberosum is thiosulphinates, 19 and various steroidal saponins, 20 which were observed in both A. tuberosum hydroalcoholic and aqueous extracts in the present study.
A. sativum chemical composition was similar to the one previously described by others. 9,21,22 Discrepancies on chemical composition of A. sativum can occur due to the method to obtain the plant extract. In the present study, A. sativum extract was obtained as previously described, 15,23 and it provided antimicrobial activity in both in vivo and in vitro assays. In addition, we used raw garlic extracts since these extracts showed higher antimicrobial activity than some isolated compounds, such as allicin. 12,24 Besides, progressive concentrations of a raw extract of A. sativum (100 to 1000 µg/mL) strongly inhibited many microorganisms. 25 The MIC observed for A. sativum in the present study (2 mg/mL) against PSSA was lower than MIC observed by, 26  In the present study, the in vitro anti-staphylococcal activity of A. tuberosum was not observed against both PSSA and MRSA strains. However, it was able to decrease the PSSA strain counts in the in vivo model. Thus, it is possible that the anti-staphylococcal activity of A. tuberosum was indirect. As observed in many antimicrobial models, in vitro assays do not provide the variability observed in the in vivo assays, especially considering immunological aspects. 28 Differences between in vitro and in vivo models in antimicrobial research are well recognized and future studies are necessary to address the effects of A. tuberosum on host immune response.
A. sativum is a well-recognized anti-staphylococcal agent against the biofilm formation in burn wounds, 29 and against both PSSA and MRSA in an in vitro MIC model, 30 as observed in the present study. In addition, diallyl sulphide and diallyl disulphide from a raw A. sativum extract showed protective action against MRSA-infection in both diabetic, 31 and non-diabetic mice. 12 Raw garlic extracts (100% and 50% concentrations) administered orally, exhibited comparable effect with both diallyl sulphide and diallyl disulphide, being safer than the two sulphides. 12 However, in the present study A. sativum did not show antibacterial activity against the MRSA strain, despite of the activity against the PSSA strain. The animal model and the garlic concentration used in the present study could have affected the results.

Conclusion
We concluded that both garlic species showed constitution comparable to other similar strains and they have molecules related to antimicrobial properties. Both species were able to reduce staphylococcal infection, despite only A. sativum showed in vitro antistaphylococci activity. No additive or complementary effect was observed by adding A. sativum extract with amoxicillin against the strains studied.