Goldcrest honey and its solvent extracts : A natural product with anti-Helicobacter pylori activity

This study evaluates the anti –Helicobacter pylori activity of Goldcrest honey at 10, 20, 50 and 75% v/v concentrations as well as its extracts (n-hexane, diethyl ether, chloroform and ethyl acetate) using the agar well diffusion technique. The minimum inhibitory concentrations (MIC50) of the two most active extracts were determined by the broth microdilution assay and MICs were read by enzyme-linked immunosorbent assay (ELISA) microtitre plate reader at 620 nm. The rate of kill of H. pylori strains by the most active extract was determined by viability studies over a period of 72 h. Data were analyzed by one-way analysis of variance (ANOVA) at 95% significance level. Goldcrest honey demonstrated anti-H. pylori activity, with the most potent activity at 75% concentration, with zone diameter in the range from 12.9 to 14.4 mm. Clarithromycin recorded a zone diameter of 18.0±7.4 mm not significantly different (P>0.05) from diethyl ether extract, with a zone diameter of 19.9±10.1 mm. MIC50 values of n-hexane and diethyl ether extracts were in the range of 0.039 to 10% and 0.078 to 10% (v/v), respectively. The bactericidal effect of n-hexane extract was highest at 4 × MIC concentration, within 30 to 72 h during which 100% of bacterial cells were killed. Goldcrest honey and its solvent extracts may contain potential lead molecules with anti-H. pylori activity. Further studies are therefore needed to determine their phytochemical constituents and activity.


INTRODUCTION
Helicobacter pylori (H.pylori) has probably been part of the human gastric biota since time immemorial (Blaser, 1997).It is a human pathogen of major public health concern since it is directly associated with many diseases of the upper gastrointestinal tract including acute and chronic gastritis, non-ulcer dyspepsia, peptic ulcer disease (gastric and duodenal ulcers) and gastric cancers (McColl, 2010), which are the major causes of death worldwide.Eradication of this pathogen is now a major step in the therapeutic management of the aforementioned diseases (O'Connor et al., 2011).However, triple therapy which is currently the treatment of choice contains two antibacterials (that is, clarithromycin or amoxicillin with metronidazole), combined with a proton pump inhibitor (Malfertheiner et al., 2012).Treatment of H. pylori infection is relatively successful, with approximately 80% of patients exhibiting eradication of the organism (O'Connor et al., 2011).The emergence of H. pylori strains resistant to the recommended antibiotics, the associated side effects, inactivation of antibiotics by pH and low patient compliance; are cited as some of the major causes of treatment failure (Bytzer and O'Morain, 2005).As a result, there is a need to seek new, safe and effective anti-H.pylori regimens with highly selective antibacterial activity against the pathogen, but without the risk of resistance and untoward effects.Honey is among the attractive sources that has received attention as an alternative treatment for H. pylori infections.
Apitherapy or therapy with bee products is an age-old therapeutic practice as recorded by several ancient civilizations.The therapeutic effects of honey have been ascribed to its antimicrobial, anti-inflammatory and antioxidant properties.It has shown powerful antibacterial effects against pathogenic and non-pathogenic microorganisms, yeast and fungi even against those that developed resistance to many antibiotics (Molan and Cooper, 2000).Honey is a supersaturated sugar solution containing vitamins, minerals, proteins, amino acids and nutrients (Gheldof et al., 2002;Ansari and Alexander, 2009).Its sugars exert strong osmotic potential attracting water molecules and as such inhibit the growth of bacteria and fungi.In addition, it contains the enzyme glucose oxidase, which acts on glucose in the presence of water, producing hydrogen peroxide and gluconic acid (Ansari and Alexander, 2009).
Hydrogen peroxide is the major contributor to the antimicrobial activity of honey, and the different concentrations of this compound in different honeys result in their varying antimicrobial effects (Molan, 1992).Furthermore, antioxidants and flavonoids that may function as antibacterial agents are also present.The pH of honey is low and ranges from 3.2 to 4.5 with the most predominant proton donor being gluconic acid.In light of modern science, several important therapeutic effects of honey have been elucidated (Cooper et al., 2002) and these vary with the quality of the honey produced.The type of honey produced is dependent on the flowers blooming in different seasons in different regions and countries (Ndip et al., 2007).Consequently, there is variation in the chemical composition as well as the physical properties of honey.As a result, micro-organisms differ in their sensitivity to honeys collected from different sources, regions and countries.
At present, a number of honeys are sold with standardized levels of antibacterial activity for example, Manuka and Capillano honeys are sold as therapeutic honeys in Australia (Lusby et al., 2005).The search for other honeys from different sources with considerable antibacterial activity continues.South Africa has a large floral biodiversity with many unique plants indigenous to the region and honeys with varying antimicrobial potential are produced from these plants and sold commercially in the country.In the Eastern Cape Province of the country, the 'Xhosas' use their honeys to produce "iqhilika" (Mead), which is a cultural wine and is said to be of great medicinal value and can be used in the treatment of coughs, kidney and stomach ailments.In spite of the wide body of research on the antibacterial activity of honey against several medically important pathogens including H. pylori, in different parts of the world, there is paucity of information, on the possible antibacterial activity of South African honeys against H. pylori.This study was designed to evaluate the anti-H.pylori activity of Goldcrest honey and its solvent extracts as well as to determine the minimum inhibitory concentrations (MIC 50 ) of the two most active extracts and the rate of kill of the test strains by the most active extract.

Bacterial strains
Thirty strains of H. pylori were used which were cultured from gastric biopsy specimen obtained from patients with gastroduodenal pathologies attending the endoscopic unit of Livingstone hospital, Port Elizabeth, South Africa.This was done after informed consent was obtained as per our previously reported schemes (Ndip et al., 2008;Tanih et al., 2010).Confirmed isolates were stored in Brain Heart Infusion broth plus 20% glycerol at -80°C for subsequent bioassays.H. pylori ATCC 43526 was used as the control.

Source and dilution of honey
In this study, Goldcrest honey produced from Citrus limon and Citrus sinesis as the main floral sources was used; it was obtained from within South Africa.It was diluted with sterile distilled water to different concentrations of 10, 20, 50 and 75(%v/v); and sterilized by filtering through a 0.22 µm membrane filter (Al-Somal et al., 1994), into separate sterile bijou bottles.

Antimicrobial susceptibility testing of crude honey
The agar well diffusion method of Dastouri et al. (2008) was used to assess the antimicrobial activity of the crude honey.Brain Heart Infusion (BHI) agar (Oxoid, England) was prepared following the manufacturer's instructions, supplemented with 7% laked horse blood (Oxoid, England) and Skirrow's antibiotics (SR 0147E, Oxoid, England).A 0.5 Mc Farland standard was prepared and 5 ml aliquoted into a sterile test tube.An inoculum of each clinical isolate was prepared from subculture of the bacterial suspension.With a sterile wire loop, five colonies of the same morphological type were picked and emulsified in sterile normal saline.The turbidity of the suspension was adjusted to correspond to 0.5 Mc Farland standard containing 1.8 × 10 8 cfu/ml.The suspension was then inoculated evenly onto the surface of BHI plates in triplicate with sterile calcium alginate swab sticks.The plates were allowed to dry for 3 to 5 min.Using a sterile 6 mm diameter cork borer, five wells were cut in the agar and into each was introduced 100 µl of the different concentrations of the honey solution.Into the remaining well, clarithromycin (0.05 μg/ml) was added as the positive control.The plates were incubated at 37ºC for 2 to 5 days under microaerophilic conditions (CampyGen BR0056A, Oxoid, England) and later examined for zones of inhibition.The zones of inhibition (in millimeters) were measured, averaged and the mean values recorded.H. pylori control strain ATCC 43526 was included in all the experiments.

Solvent extraction of crude honey
This was carried out using the method of Zaghloul et al. (2001) with modifications.A 100 g of crude honey was placed in a 500 ml separating funnel, diluted with 150 ml of sterile distilled water and extracted with 150 ml of the different organic solvents (n-hexane, diethyl ether, chloroform and ethyl acetate).This was performed as three successive extractions using 50 ml of solvent each time.The shaking time for each extraction process was 15 min, after which the mixture was allowed to stand to permit the solvent layer to separate.The three successive layers were collected, mixed and concentrated by evaporation under reduced pressure using a rotary evaporator (Steroglass, Strike 202, Italy) at 40°C for n-hexane, 30°C for diethyl ether, 50°C for chloroform and 60°C for ethyl acetate.Water contaminating solvent extract was removed by filtration over anhydrous Sodium sulphate.

Antimicrobial susceptibility testing of honey solvent extracts
The different extracts of the honey at 75% concentration were tested against the isolates using the method described for crude honey above.The respective pure solvent used for the extraction was tested side by side with its extract.Diameters of zones of inhibition of extracts were measured; averaged and mean values recorded in millimeters.

Determination of MIC
The two most active extracts of the honey were employed in broth microdilution assay to determine their MICs against the test isolates, according to the method of Bonacorsi et al. (2009).Twofold dilutions were prepared in 96-well-round-bottom microtitre plates (Greiner Bio-One, Frickenhausen) in BHI broth (Oxoid, England); the final extract concentration was 0.01 to 10%v/v.Similarly, amoxicillin (0.0012 to 1.25 mg/ml) and metronidazole (0.01 to 10 mg/ml) were two-fold diluted and tested on the same plate with the honey solvent extracts as reference antimicrobials.Control wells were also prepared with culture medium only, culture medium with honey extract and culture medium with bacterial isolate only.
The inoculum of each strain was diluted tenfold in sterile normal saline.20 µl of the bacterial suspension (108 CFU/ml) was aliquoted into each well.The final volume in each well of BHI broth, honey extract and inoculum was 120 µl.The absorbencies were read using an ELISA micro plate reader (Model 680, S/N 19138, Biorad, Japan) adjusted at 620 nm.The micro plates were sealed and incubated at 37°C under microaerophilic conditions for 2 to 3 days, agitated and the absorbencies were again read at the same wavelength.The absorbencies were compared to the values obtained before incubation to detect any increase or decrease in bacterial growth.The lowest concentration of the extract resulting in inhibition of bacterial growth by 50% was taken as the MIC 50 .

Time-kill assay of extract
Assay for the rate of kill of H. pylori isolates by the most active extract of Goldcrest honey was determined in accordance with the method of Akinpelu et al. (2008) with modifications.Each isolate was subcultured on CBA (Oxoid, England) plates and incubated at 37°C under microaerophilic conditions for 2 to 3 days.Growth of each isolate was transferred into BHI broth (Oxoid, England) and incubated overnight under the same growth conditions.The turbidity of an 18 h old broth culture of the test isolate was standardized to contain approximately 1.8 × 10 8 cfu/ml.A 0.5 ml volume of the standardized suspension was added to 4.5 ml of different concentrations of the extracts (1/2 MIC, MIC, 2 × MIC and 4 × MIC).These were incubated at 37°C under microaerophilic condition in an orbital shaker at 120 rpm and the killing rate was determined over a period of 72 h.Exactly 0.5 ml volume of each suspension was withdrawn at 6 h intervals and transferred to 4.5 ml of BHI broth recovery medium containing 3% "Tween 80" to neutralize the effects of the antimicrobial compound carry-overs from the test isolates.The suspension was serially diluted and 100 µl plated out for viable counts.The plates were later incubated at 37°C for 72 h.The control plates contained the bacterial cells without the extract.The emergent bacterial colonies were counted and compared to the counts of the culture control.Time-kill assays were carried out in duplicate.

Statistical analysis
Diameters of zones of inhibition were expressed as mean ±  Standard deviation (SD).One-way analysis of variance (ANOVA) test was used to determine any statistically significant difference by comparing zone diameters of the honey at various concentrations; zone diameters of clarithromycin to different solvent extracts, zone diameters of different extracts as well as the MIC values of these extracts were compared to each other at 95% significance level.

Antimicrobial susceptibility testing of crude honey
At various concentrations (10, 20, 50 and 75% v/v), the honey demonstrated antibacterial activity against H. pylori strains with zone diameters (mean ± SD) that ranged from 12.9 to 14.4 mm, with the most potent activity (20/30, 66.7%) at 75%v/v concentration (Table 1).The least antibacterial activity was observed with 20% concentration (16/30, 53.3%).Our positive control (Clarithromycin) equally demonstrated antibacterial activity against H. pylori strains exhibited by a zone diameter (mean ± SD) of 18.0 ± 7.4 mm.However, no statistically significant difference (P>0.05) was reached comparing the zone diameters of the honey at various concentrations.

Antimicrobial susceptibility testing of solvent extracts
All the solvent extracts demonstrated antibacterial activity against the test strains.The zone of inhibition (mean ± SD) ranged from 15.2 to 19.9 mm (Table 2).The most inhibitory activity was demonstrated by diethyl ether extract (22/30, 73.3%) exhibited by an inhibition zone of diameter 19.9 ± 10.1 mm while the least antibacterial effect was observed with the chloroform extract (17/30, 56.7%), exhibited by an inhibition zone of diameter 15.2 ± 8.7 mm.Both the n-hexane (19/30, 63.3%) and ethyl acetate extracts (18/30, 60%) were equally active with inhibition zones of diameter 17.9 ± 8.7 and 16.7 ± 9.3 mm, respectively.Consequently, the zone diameters of the extracts did not reach statistical significance (P>0.05).

MIC determination
The MICs of the two most active extracts were determined by an ELISA microtitre plate reader at 50% bacterial growth inhibition using broth microdilution.Nhexane and diethyl ether extracts were the two most active extracts, with MIC 50 values in the range of 0.039 to 10%v/v and 0.078 to 10%v/v, respectively (Table 3).Contrastingly, MIC50 values of the reference antibiotics were in the range; 0.001 to 1.25 mg/ml (amoxicillin) and 2.5 to 10 mg/ml (metronidazole).However, no statistically significant difference (P>0.05) was reached comparing the MIC values of the solvent extracts to each other by One-way ANOVA test at 95% significance level.

Time-kill assay
The bactericidal effect of n-hexane extract on test strain was determined over a period of time ranging between 6 to 72 h.As shown in Figure 1, the lag phase of H. pylori strain PE 252C was between 6 and 12 h as there was no growth on the control plate as well.At a concentration of 40%v/v (4 × MIC), 100% bacterial cells were killed within 30 to 72 h.Overall, there was growth of bacterial cells at the concentrations of ½ MIC, MIC, 2 × MIC within 12 to 72 h.

DISCUSSION
In the preliminary screening, it was revealed that the test strains were sensitive to honey concentration as low as 10%v/v.This result corroborates the finding of Tajik et al. (2008) who reported the antimicrobial efficacy of natural Urmia honey against Gram negative and Gram positive bacteria.There are many reports about the antibacterial properties of natural honey.Its antimicrobial activity has been ascribed to high content of reducing sugars, low pH, low water activity (Aw), low protein content, lysozyme, and hydrogen peroxide and non-peroxide components (phytochemicals) as investigated by several authors (Molan and Cooper, 2000).Honey is known to contain phytochemicals including tetracycline derivatives, peroxides, amylase, fatty acids, phenols, ascorbic acid, flavonoids, streptomycin, sulfathiazole, terpens, benzyl alcohol and benzoic acids (Heering et al., 1998).They may be of plant origin and could be extracted with organic solvents.The amount of these could be small or diluted in the honey but become concentrated and exhibit more activity after extraction (Yao et al., 2004).These components could act as natural anti-oxidants as they are reported to scavenge for free superoxide and other reactive oxygen metabolites liberated during respiratory burst in H. pylori induced mucosal damage (Li et al., 2001).However, Obaseiki-Ebor et al. (1983) isolated volatile substances from honey with antibacterial activity.
Other researchers found the non-peroxide components of honey extractable by organic solvents (Oka et al., 1987).In this study, organic solvents of varying polarities namely n-hexane, diethyl ether, chloroform and ethyl acetate were used to extract the antimicrobials in honey.
All the solvent extracts demonstrated antibacterial activity against the test strains.Most of the strains were sensitive to diethyl ether extract (22/30, 73.3%).This is in contrast with the finding of Khalil et al. (2001) that showed a better activity of chloroform extract against their isolates.Furthermore, the solvent extracts demonstrated a greater inhibitory effect than the crude honey as evidenced by an increase in the zone diameter of inhibition and percentage susceptibilities of the test strains.This may suggest that the putative antimicrobial components might have been concentrated in the honey after extraction.In addition, five out of seven test strains that were resistant to the positive control (23/30, 76.7%) were remarkably sensitive to the extracts producing appreciable zones.Moreover, no statistically significant difference (P>0.05) was recorded comparing the zone diameters (mean ± SD) of the extracts to the positive control.Suggesting these extracts may contain antimicrobial agents whose therapeutic potential are comparable to clarithromycin (positive control), which is recognized as a key antibiotic against H. pylori infections (Ndip et al., 2008).
In the assay to determine the MIC 50 values of diethyl ether and n-hexane extracts against the test strains, we employed amoxicillin and metronidazole that have been reported as the most sensitive and resistant antibiotics respectively against H. pylori in our environment (Tanih et al., 2010).The antibiotics served as positive controls to test the validity of the assay as well as to control the experiment.This is in accordance with the work of Kñazovicka et al. (2009).The assay indicated that the MIC 50 values varied with extracts and the test strains.This variation could be attributed to different putative components present in the different extracts.This result is in contrast with the work of Aljadi and Yusoff (2003).The bactericidal activity of the n-hexane extract was determined using viability studies.The periods 6 to 12 h could be considered as the lag phase of the isolate since there was no growth on the negative control plate as well as on plates inoculated with sample from extract constituted broth.The bactericidal activity was highest at 4 × MIC concentration within 30 to 72 h.This may suggest that with a further increase in the extract concentration, better results would be obtained.There was recurrent growth of bacterial cells from 48 to 72 h after the cells were being killed at 42 h at the lowest concentration (1/2 MIC).In addition, at MIC, there was growth of bacterial cells from 12 to 72 h showing they were  inhibited as expected but very little growth of bacterial cells at 2 × MIC from 12 to 48, 60 to 72 h, suggesting the n-hexane extract at these concentrations could be bacteriostatic.
In conclusion, it is worth mentioning that Goldcrest honey and its solvent extracts presented anti-H.pylori activity.It is a promising observation thus indicating compounds of antibacterial potential could be isolated by fractionation from the extracts and harnessed in a bid to provide potential lead molecules with anti-H.pylori activity.Moreover, the honey could complement/supplement standard triple therapy in order to alleviate unpleasant side effects posed by antibiotics as well as improve patients compliance -which are pivotal factors affecting the effectiveness of treatment regimen (Bytzer and O'Morain, 2005).Further studies are therefore needed to determine their phytochemical constituents and activity.

Figure 1 .
Figure 1.Bactericidal activity of n-hexane extract of Goldcrest honey against PE 252C.

pylori isolates Diameter of zones of inhibition (mm)
aMean zone diameter after triplicate assay.

Table 2 .
Antibacterial activity of solvent extracts of Goldcrest honey against H. pylori isolates. H.

pylori isolates Diameter of zones of inhibition (mm) a Solvent extracts of Goldcrest honey
aMean zone diameter after triplicate assay

Table 3 .
MIC values of solvent extracts and classical antibiotics at 50% bacterial growth inhibition. H.

pylori isolates MIC 50 values in different concentrations Solvent extracts of Goldcrest (%v/v)
MIC 50 values after triplicate assay; -: MIC 50 values not within susceptible range.