The Effect of Manuka Honey on Enterobacteria

ABSTRACT

Manuka honey (Leptospermum scoparium) produced in New Zealand has been shown to exhibit substantial antibacterial activity against a broad range of pathogens causing wound infection, and is being used in wound management with excellent results. This activity is due to both hydrogen peroxide and non-peroxide components. The research in this thesis is set out to evaluate in vitro the efficacy of manuka honey as an antibacterial agent against enterobacteria, taking into consideration some factors that may be involved in the gastrointestinal environment.

Because some gastrointestinal bacteria (Campylobacter spp., Helicobacter pylori, Lactobacillus spp. and Bifidobacterium animalis subsp. lactis) are not aerophilic, a cheap yet acceptable gas generating system alternative to the commercial gaspack counterpart was sought for use in this study. Various alternatives were compared for their performance. The spirits burn method was chosen for cultivating microaerobes and some anaerobes because of its comparable performance to that of commercial systems in terms of the growth of bacterial species, and because of the ease of use and the low cost.

In the first part of this thesis, the susceptibility of gastrointestinal bacteria against manuka honey was investigated by determining the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) using a standardized manuka honey. Throughout the research, a manuka honey with median level non-peroxide antibacterial activity (equivalent to that of 16.5% phenol) was used, except that Campylobacter spp. were assayed with a more potent manuka honey equivalent to 29.4% phenol. The measured sensitivity of bacteria showed that manuka honey is significantly more effective than artificial honey (a mixture sugars as in honey), indicating that osmolarity is not the only factor that is responsible for the antibacterial activity of the honey. It was found that some species of bacteria e. g. Campylobacter spp. are exceptionally sensitive to manuka honey (both MIC and MBC are about 1% honey solution), whereas most other gastrointestinal pathogens have MIC and MBC values in the range 5–10% honey other than Enterobacter and Pseudomonas which were in the range 10–17%. Bifidobacterium, lactobacilli and enterococci appear to be more tolerant to the honey (MIC: 9.36–14.29%; MBC: C13.3%) than most other species are. The difference in efficacy between the honey with and without hydrogen peroxide removed was also studied, and it was found that both hydrogen peroxide and the non-peroxide components contribute to the bacteriostatic and bactericidal activity of the honey.

Because oxygen is required for hydrogen peroxide to be produced in honey, the role that oxygen plays in the antibacterial activity of manuka honey was investigated by analyzing the susceptibility data obtained under both aerobic and anaerobic conditions using facultative anaerobes. Manuka honey appeared to be a more potent bacteriostatic agent against most species of bacteria in the absence of oxygen, whereas a relatively higher concentration of manuka honey solution was required to kill some bacteria under anaerobic conditions. This may partially be due to the atmosphere having also affected the metabolism, and hence the growth, of bacteria. Therefore, the activity of manuka honey would not necessary decline in the intestinal environmental atmosphere.

To investigate how long it takes for manuka honey to kill bacteria, time-to-kill studies were conducted by monitoring the survival of bacteria in manuka honey. It is found that it takes a 20% solution of manuka honey with a medium-level activity more than 6 hours to kill 90% of the cells of most of the species tested if the bacterial cells are kept in contact with the honey. This suggests that manuka honey is not rapidly bactericidal, and that it is unlikely to be possible to fully eradicate a bacterial gut infection by ingesting a small amount of manuka honey for a short period. It was found that probiotics can survive in the 20% honey solution for more than 12 hours.

The pharmacodynamics of the antibacterial activity of manuka honey were studied to investigate the survival and the re-growth of bacteria after they had been treated with honey. It was revealed that after being exposed to manuka honey for a short term (1 hour), the growth of most enteropathogens is slowed for approximately 2–4 hours before it gets back to a full rate. The assays of this postantibiotic effect also showed that the latency in the regrowth after being exposed to honey is not proportional to the MIC, MBC or time-to-kill profiles.

Finally, the efficacy of manuka honey on bacteria was studied under conditions simulating the environment in the stomach and intestines. The tested bacteria were unable to grow under the acidic conditions as in the stomach, so whether or not the honey had any antibacterial activity under these conditions could not be determined. Under the conditions simulating the intestinal environment, the results demonstrated that the antibacterial activity of manuka honey is slightly decreased in the mildly alkaline conditions of the intestine (pH 7.5). In the presence of pancreatin and bile at the same pH, the activity of manuka honey was found to decrease by more than 50%. This suggests that pancreatin and bile in the gut may negatively affect the efficacy of the antibacterial activity of manuka honey in vivo. This indicates that although ingested manuka honey may still have some antibacterial action when in the gut, the antibacterial activity would be different from that which is usually examined with sensitivity studies in vitro.

Gastroenteritis has generally been treated with oral rehydration solution (ORS) that consists of carbohydrates and electrolytes. Manuka honey could be used instead of the usual carbohydrate component of ORS and would provide additional bioactivities such as antibacterial activity and stimulation of growth of probiotics, which would make the honey rehydration solution more beneficial to patients with gastroenteritis than is the traditional ORS. After some initial investigation to find the most appropriate dosage and frequency of doses, a clinical trial may be warranted.

INTRODUCTION AND LITERATURE REVIEWS

This chapter first gives a general introduction to gastroenteritis, followed by giving four parts of literature review - the human digestive system, gastroenteritis and its associated bacteria, commonly used therapies and their limitations, and the medical use of honey. The possibility is also discussed that the gastrointestinal environment may affect the antibacterial activity of manuka honey, and therefore the in vitro activity may not reflect the actual efficacy in gastrointestinal infection.

Lastly, the intentions of this thesis, that the assessment of the sensitivity of enteropathogenic bacteria to the antibacterial activity of manuka honey, and the assessment of the effect of gastrointestinal environmental factors on the antibacterial activity of manuka honey, are introduced.

Gastroenteritis is an inflammation in the gastrointestinal tract that results in acute diarrhea. In the world, millions of people are killed by improperly treated gastroenteritis each year, and it is the leading cause of death among infants and children due to their impaired immunity. Most infectious gastrointestinal diseases are known to be food-borne, i. e. invoked by ingesting unflavored substances by the host along with food or drink water. Generally these “intruders” may not be able to cause disease as they would have to pass through a series of defensive barriers provided by the digestive system before any symptom can result.

HONEY THERAPY

This section reviews the use of honey, now considered to be an alternative medicine but which has been used since the time of the ancient Egyptian, the Hebrew kingdoms, and historically in China, India, Greece, Rome and many other nations. Emphasis in this section is placed on the antibacterial properties of honey.

The use of honey for treating gastroenteritis, peptic ulcers or gastritis can be traced back to the ancient era. The Muslim prophet Mohammed and the Roman physician Celsus used honey as a cure for diarrhoea. The use of honey on treating gastroenteritis was also recorded in ancient China. Other countries, especially Russia and Arabic countries, also have been traditionally using honey as an elixir for treating upper intestinal dyspepsia.

Honey is produced by bees from the nectars they collect from flowers. When a bee collects nectar from flowers, it secretes into it enzymes from its pharyngeal gland. The nectar is then dehydrated and matured in the honey combs as the stored dietary energy source of the bees. Various kinds of honey have been produced in the world and the property of the honey usually reflects that of the floral source. Crane produced a thorough review on honey in which she noted that the medical literature of honey may be traced back to as far as 2000 BC. Crane also noted that in England and the Soviet Union honey by itself was used as a surgical dressing for open wounds, burns, and septic infections in the mid-20th century. It is also of interest to note that a clinical trial in Switzerland revealed that honey was useful for easing the sickness after radiation treatment. In the Chinese Encyclopaedia it is recorded that honey is also used in ancient China for various diseases.

COMPOSITION AND PROPERTIES OF HONEY

Honey is a complex material but is primarily a saturated or super-saturated solution of sugars which largely consists of glucose and fructose (84%) and the high percentage of sugars makes it of high osmolarity. Although present in much lesser quantities than glucose and fructose, honey also contains other carbohydrates including disaccharides (sucrose and maltose) and oligosaccharides which seem to be vary depending on the floral source of the honey.

Besides carbohydrates honey also contains a number of enzymes. Some of the most significant enzymes in honey are glucose oxidase, amylase and invertase, which appear to originate from honeybees. Glucose oxidase has been of particular significance as this is responsible for the generation of gluconic acid and hydrogen peroxide, which are mainly responsible for the antimicrobial activity of honey. Other minor enzymes including catalase and acid phosphatase are also found in some honeys but these are likely to be derived from the pollens and nectar of plants.

Honey contains other constituents. These include vitamins A, the B group of vitamins, and vitamins C, D, and E, mineral salts, organic acids, proteins, amino acids, lipids, hydroxymethylfurfuraldehyde (HMF) and other minor substances that contribute to honey colour, aroma and flavour, although these are relatively less in significance in the daily diet because of their low concentrations. Interestingly, hydroxymethylfurfuraldehyde may be formed by the decomposition of fructose in the presence of acid and this was usually taken as the evidence of the addition of invert sugar. Research showed that even fresh honey contains a small amount of hydroxymethylfurfuraldehyde, which increases with time when the honey is stored at room temperature. The increase in hydroxymethylfurfuraldehyde level is retarded when stored in a cool environment.

THE ANTIBACTERIAL ACTIVITY OF HONEY

Honey has been used in medical treatment as an antiseptic since ancient times until antibiotics were invented. However, the nearly unlimited use of antibiotics has led to the emergence of antibiotic-resistant microorganisms that have made the treatments more difficult than ever, and this has also made humans to seek alternative therapies such as honey to treat the diseases. Numerous research studies have shown that honey is effective against a wide range of microorganisms including MRSA and VRE.

Several mechanisms of antimicrobial activity in honey are discussed below:

OSMOLATIRY AND WATER ACTIVITY

As described previously, the process by which bees make honey and the high content of sugars make it highly osmotic. Most water molecules in honey interact with the sugars and the proportion of “free water”, described as the water activity (aw), is too low for microorganisms to utilize when honey is not diluted. Generally honey is reported to have aw of 0.56–0.62 while most organisms have a minimum aw of 0.9–1.0 for growth and can not survive in the lower aw environment. To reach the aw above which most microorganisms can survive, a typical honey would need to be diluted down to about 2–12% (based on the reasoning that the concentration is proportional to −log aw. Although a few microorganisms such as osmophilic yeasts can live in honey with an unusually high water content and result in spoilage of the honey, undiluted ripened honeys generally have an aw which is too low for any species to survive.

However, osmolarity may not be useful from an antimicrobial viewpoint as the sugar component would readily be diluted by body fluid if ingested or by exudate if used as a wound dressing. Also the sugar content in honey can be rapidly absorbed in the gastrointestinal tract. Once honey is diluted, the water activity rises and the osmolarity is no longer inhibitory to the microorganisms.

ACIDITY

Honey is characteristically acidic, with an average pH between 3.2 and 4.5 which is too low for most organisms to survive as the optimal pH for most organisms is between 7.2–7.4 and the viability largely declines as the acidity rises. Although several organisms can survive in relatively acidic conditions (e. g. E. coli at pH 4.3, P. aeruginosa at pH 4.4 and Salmonella spp. at pH 4.0) the pH of undiluted honey is usually too low for the microorganisms to survive. The acidity of honey is largely due to the gluconic acid that exists in honey.

However, like the osmolarity described previously, acidity is unlikely to be a key factor that is responsible for the antimicrobial activity. The quantity of gluconic acid is quite low and the pH would be raised by dilution of honey with body fluid which contains buffers (0.17–1.17%). Research has shown that a buffered gluconolactone/gluconic acid solution that was equivalent to 25% honey solution did not reveal detectable antibacterial activity against Staphylococcus aureus in an agar diffusion assay whereas 12.5% honey solution gave much higher activity. Other research studies have shown that a remarkable antibacterial activity can be detected even after honey has been neutralized, which suggests that the level of antimicrobial activity is not linearly correlative to the pH.

PHYTOCHEMICAL COMPOUNDS

Several phytochemical compounds have been isolated from honey by many researchers. These include benzyl alcohol, pinocembrin, terpenes, 3,5 - dimethoxy - 4 - hydroxybenzoate (syringic acid), methyl 3,5-dimethoxy- 4-hydroxybenzoate (methyl syringate), 3,4,5-trimethoxybenzoic acid, 2- hydroxy-3-phenylpropionate, 2-hydroxybenzoate and 1,4-dihydroxybenzene. However, the quantity of these compounds in honey is too low to account for the significant antimicrobial activity.

HYDROGEN PEROXIDE

Hydrogen peroxide is one of the dominant antibacterial substances that exist in honey. When a bee collects nectar from flower, it secretes glucose oxidase from its hypopharyngeal gland into the nectar to assist the formation of honey from the nectar. Through the activity of this enzyme, glucose in the honey is transformed into gluconic acid and hydrogen peroxide.

Hydrogen peroxide is an effective antimicrobial against a number of microorganisms, and is commonly used as an antiseptic. It is reported that some microorganisms, mainly Lactobacillus spp., produce hydrogen peroxide for help competing against other microorganisms.

There are several drawbacks when hydrogen peroxide is used as an antiseptic. First, hydrogen peroxide is readily degraded into oxygen and water when catalase exists. As catalase exists in plasma and in body tissues, which would destroy hydrogen peroxide, the efficacy of the antiseptic may be lost in a short time. Second, hydrogen peroxide is an irritant to body tissues and the patients may feel uncomfortable. Third, reactive oxygen species (ROS) derived can do harm not only to bacterial cells but to tissue by breaking down proteins, nucleic acids and cell membrane lipids, and also by activating proteases in the wound tissues. Therefore using hydrogen peroxide as an antiseptic is unfavorable.

It is interesting to note that undiluted honey has a negligible level of hydrogen peroxide, and even if honey is diluted the concentration of hydrogen peroxide generated is still far lower than the 3% solution of hydrogen peroxide typically used as an antiseptic. An explanation to this is that the glucose oxidase in honey is inactive when the honey is undiluted. The enzyme reveals highest activity at pH 6.1 and a good activity between pH 5.5–8, low activity at below pH 5.5 and zero at pH 4. Bang reported that the concentration of hydrogen peroxide in honey increased once it was diluted, and reached a maximum level of 3.65 mM when the honey was diluted down to 50% (v/v), but this was still far lower (242-fold) than in the 3% hydrogen peroxide solution used as an antiseptic. Interestingly yet, it has been found that low levels of hydrogen peroxide is more effective when expose continuously to bacteria than when applied as a bolus, which suggests that the continuously generated hydrogen peroxide caused by glucose oxidase in honey is in fact more antibacterial than the low concentration could have suggested.

It must be noted that hydrogen peroxide in honey seems to reveal higher antibacterial activity than hydrogen peroxide alone, which suggests that some indigenous substances in honey would raise its activity. 0.1 mM ascorbic acid and metal ions were added in a hydrogen peroxide solution and found that the bactericidal potency of hydrogen peroxide increased. In another study done by Waites it was shown that hydrogen peroxide was more sporicidal when 10 mM copper was added. McCulloch also showed that 0.83 mM iron, copper, chromium, cobalt or manganese increased the potency of hydrogen peroxide 10-fold. As the antibacterial action of hydrogen peroxide is largely achieved via oxygen free radicals rather than by hydrogen peroxide itself, it is possible that the synergistically enhanced antimicrobial potency of hydrogen peroxide being reported is due to the catalytic action of the ions that potentialized the production of the damagingly reactive hydroxyl free radical species.

However, the gastrointestinal environment may be unfavorable to the antibacterial activity in honey that is due to hydrogen peroxide. Oxygen is needed for hydrogen peroxide to be produced by glucose oxidase activity, whereas oxygen is not available in the intestines. Although glucose oxidase is stable against digestion by protease activity in the gut and intestine, it may be denatured by the low pH in the stomach. That is, it is possible that the actual antibacterial activity of honey in the gut might not be as significant as that observed at the bench because hydrogen peroxide production could have been at least impaired (although hydrogen peroxide having been accumulated in diluted honey before it entered the gut could be of some effect).

NON-PEROXIDE COMPONENT

Several researchers have shown that beside hydrogen peroxide there exists non-peroxide antibacterial activity in some honeys. In these studies the authors noticed that some honeys exhibited antibacterial activity even if the honeys were heated to inactivate glucose oxidase or treated with catalase to destroy the hydrogen peroxide in the honeys.

Specifically in some manuka honey (Leptospermum scoparium) produced in New Zealand there exists a substantial antimicrobial activity that is not destroyed by catalase. Much research has been done on the non-peroxide antibacterial activity of manuka honey and it has been found that the activity is more stable than is the antibacterial activity due to hydrogen peroxide (not destroyed by light and heat). It actually increases with time at room temperature. However, it is rapidly inactivated in an alkaline environment. As this unusual non-peroxide activity only exists in manuka honey, Professor Peter Molan at the University of Waikato termed this activity as the Unique Manuka Factor (UMF). The non-peroxide activity of manuka honey has revealed significant efficacy against wide range of organisms, including the antibiotic-resistant organisms MRSA and VRE that are otherwise difficult to eradicate.

It should be noted that this non-peroxide antibacterial activity exists only in some manuka honey. Several possibilities have been suggested to explain the variation in the activity of manuka honey, and these theories have recently been thoroughly examined by Stephens. After investigating several possible biological factors (animal, plant and fungal associations) and non-biological factors (location of sites and climate), he concluded that some manuka honey had been diluted by nectar collected from other flora species by honeybees. In some manuka honey samples it was estimated by measurement of thixotropy that they contained less than 30% Leptospermum nectar, which renders the non-peroxide antibacterial activity too low to be measured in the agar diffusion assay of Allen.

METHYLGLYOXAL IN MANUKA HONEY

While this project was progressing, Mavric published the proposal that methylglyoxal (MGO) is the substance responsible for the nonperoxide antibacterial activity of manuka honey. At the same time Adams used HPLC to isolate the non-peroxide antibacterial activity in manuka honey and proved that it was methylglyoxal.

Surprisingly, Adams found that methylglyoxal does not dominate in freshly produced manuka honey, nor does it exist in the nectar of manuka flower at a detectable level. Adams et al. reported that the nectar of manuka flower contained a high level of dihydroxyacetone (DHA), and storage at 37°C led to a decrease in the level of the dihydroxyacetone in the honey and a related increase in that of methylglyoxal. This finding, that the methylglyoxal in manuka honey is formed with time from the dihydroxyacetone in the nectar of manuka flower, is in agreement with the observation that the non-peroxide antibacterial activity of manuka honey continuously increases during storage.

While the finding that methylglyoxal is the major antibacterial factor of manuka honey is of interest, it may be still too early to draw conclusion that methylglyoxal alone is the only factor that is responsible for the significant non-peroxide activity in manuka honey. Molan pointed out, using the data published by Adams that methylglyoxal alone did not account fully for the antimicrobial activity that a manuka honey generally has. To illustrate that methylglyoxal does not fully explain the non-peroxide antimicrobial activity in manuka honey, Molan also compared the antimicrobial activity of methylglyoxal in honey with that of methylglyoxal in water, and demonstrated that the antibacterial activity of the former was more than 3–4 times higher than that of the latter. These results clearly explain that methylglyoxal alone does not account for the antimicrobial activity of manuka honey, and also suggest that some non-antimicrobial components that exist in the honey must have acted as synergists with methylglyoxal to provide the substantial antibacterial activity of manuka honey. Additionally, in accordance with the regression analyses amongst the scatter plots of methylglyoxal vs antibacterial activity given by Adams, Atrott and Henle, there exists non-peroxide antibacterial activity equivalent to that of around 7.5–10% phenol that is not accounted for by methylglyoxal alone if the linear regression plot is extrapolated back to zero methylglyoxal. As honey has a very complex composition, and also there are various interactions among the components that may influence its activity, further research is needed to be done on the antimicrobial components in manuka honey to understand the mechanism(s) of the honey on microorganisms.

REPORTED USE OF MANUKA HONEY FOR TREATING GASTROENTERITIS

Honey has been used for the treatment of veterinary diarrhea. An 8% (v/v) solution of honey was reported to be effective for the treatment of chronic diarrhea in horses. It has also been reported that the number of ulcers caused by aspirin, a non-steroidal anti-inflammatory drug (NSAID), in 10 rats was significantly decreased in the group treated with floral honey (3 cf. 10), whereas it was less significantly decreased if treated with honey from sugar-fed bees (8 cf. 10) or increased in those given saline (15 cf. 10). Ali carried out a similar study, in which the healing rate of honey against ulcers caused by another non-steroidal anti-inflammatory drug, indomethacin, in the rats was reported to be 70%. The same author also reported that honey prevented ulceration from being caused by indomethacin. Badawy reported that mice infected with E. coli O157:H7 or S. typhimurium had a lower mortality in the group injected with 7 month old Egyptian clover honey than in the control group (E. coli: 0% cf. 86.6%; S. typhimurium: 40% cf. 93.3%), whereas the reduction in the mortality was less significant in the honey being stored over a long term. However, it must be noted that Badawy et al. did not include oral rehydration solution as a control group in their animal trial, therefore it is not known if the reduction in the mortality is due to the antibacterial activity of honey or the effect of rehydration of the honey.

A clinical trial in humans with a relatively large sample size was reported. In this study 169 infants and children admitted into hospital suffering from gastroenteritis were assigned into two groups (in each group there were 18 patients with bacterial diarrhea). One group was treated with honey whilst the other was treated with the standard oral rehydration therapy (ORT; a 2% solution of glucose and electrolyte). The treatment with honey solution revealed a statistically significant reduction in the duration of bacterial diarrhea (58 hours cf. 93 hours), and gave no increase in the duration of non-bacterial diarrhoea. In another clinical trial, 45 patients with dyspepsia were given no treatment other than 30 ml of honey solution before meals three times daily. After the treatment the number of patients passing blood into feces declined from 37 to 4, the number of patients with dyspepsia from 41 to 8, the number of patients with gastritis or duodenitis from 24 to 15, and the number of patients with duodenal ulcer from 7 to 2.

MANUKA HONEY AND GASTROINTESTINAL PATHOGENS

Comprehensive reviews of the large amount of research carried out on the antimicrobial activity of honey against a large number of microbial species have been published by Molan and Blair. Several in vitro research studies have also shown that honey may reveal antimicrobial efficacy against a wide spectrum of gastrointestinal pathogens.

However, it must be noted that those reported results are usually not comparable with each other. For instance, Mundo tested the sensitivity of gastrointestinal pathogens E. coli O157:H7, Listeria monocytogenes, S. typhimurium and S. aureus with 13 different honeys including manuka honey, and reported that a high concentration of honey (50–100%) was required to inhibit all the bacteria. Relatively high concentrations of manuka honey (25–50%) were required to inhibit the bacteria in that study, whereas Lusby reported that a concentration of 5% of all tested honeys, including manuka honey, were effective against E. coli.

There are several factors that have made it difficult to compare the results from the different reports. Some factors that are usually missed out in the reports include the inoculum density and the state of the tested microorganisms, the floral source and the storage of the honey, the media and the method used in the test, and sometimes whether or not an artificial honey is included in the assay as a control. One of the most important parameters being missed out in most of the studies, however, is that the authors failed to standardize the antibacterial activity of the tested honey using a standard antiseptic, so that it is highly possible that the potency of the antibacterial activity of the honeys used in the studies may have varied many fold. This may have led to the different results when testing the same bacterial species even if the honeys were from the same floral source. For example, Lusby commented that the overall poor activity of honeys against S. aureus was unexpected as previous reports have shown that manuka honey has an excellent activity against this organism. The potency of the antimicrobial activity of honey can in fact vary up to 100-fold. Therefore, it is essential to have the antimicrobial activity in a honey standardized when that honey sample a honey sample is to be assessed against specific microorganisms or to be compared with other types of honey.

Another example of variable results in the published reports is the efficacy of honey against the genus Campylobacter, a widespread causative pathogen of diarrhea. To the author’s knowledge there have been very few reports of testing this genus with honey, and these tests were done on only one isolate of C. jejuni. In the study carried out by Obi, honey was found to have a significant activity against the tested Campylobacter whereas Adebolu reported that natural honeys did not show any antibacterial activity against the tested C. jejuni strain, but did not discuss the reason. In fact, the research study done by Adebolu is highly questionable. Beside the fact that the activity of the tested honeys were not standardized, in this work the agar diffusion method with nutrient media was used, which may not be suitable for testing the sensitivity of slow-growing bacteria like Campylobacter spp. against honey, as the honey may have diffused out into the agar and thus had its concentration decreased to a level below the MIC by the time the microorganism had grown. This dilution by the agar means that the MIC values for any antimicrobial agent will be reported as being higher than the true value when an agar diffusion method is used.

Yet another variable that is commonly missed out in studies that are in the literature is the cell density of the bacteria being tested in the susceptibility assay. It is generally observed that the higher the cell number, the more resistant the cells are to antimicrobial agents. Depending on the species of bacteria or the antibacterial agents being tested, the MIC could rise 4 to 16-fold with as little as 0.5 log10 increase in inoculum density. Wiegand and Burak reported that the MIC of eight tested antibiotics against P. shigelloides dramatically increased from B0.03 mg/l using 105 CFU/ml up to 16 mg/l using 106 CFU/ml. It has also been observed that the influence of inoculum size substantially increased as the inoculum exceeded 9×107 CFU/ml (Edwards et al., 1991). The influence of inoculum density on sensitivity studies is understandable because an increase in the inoculum would reduce the effective concentration or the percell concentration of antibacterial agents. Therefore, without the information on the cell density being given, it would be difficult to determine if the reported sensitivity of the microbes to the antibacterial agents being tested is actually overestimated or underestimated.

In this thesis both the antibacterial potency and the cell density are standardized, and this makes the findings from this thesis of greater value than those from other similar work where this was not done.

MANUKA HONEY AND PROBIOTICS

The gut microflora plays an important role in maintaining gastrointestinal health. It is thought that by maintaining the beneficial microorganisms, humans may decrease the chance of suffering from gastroenteritis.

Several researchers have shown that honey may reveal positive effects on the normal flora, although some of the results may not always agree with others due to several experimental factors such as the floral source of the tested honey and the microbial flora examined. Ezz El-Arab demonstrated that the colon bifidobacteria and lactobacilli counts in male Swiss albino mice were markedly increased in the group receiving food supplemented with a monofloral (cotton) honey even when the mice had administered ochratoxin A (10 ng/kg by weight/day) or aflatoxins (1 μg/kg by weight/day), although the concentration of the honey in the diet was not clearly noted. Kajiwara compared the stimulative effect of 5% honey on the growth of intestinal bifidobacteria with fructooligosaccharide (FOS), galactooligosaccharide (GOS) and inulin, and found them to be equivalently effective. Shamala also noted that Lactobacillus acidophilus and L. plantarum had higher viable counts in a medium with a diluted honey (equivalent to 1% sugar concentration; floral source unknown) than in a medium with sucrose (1%) or a mixture of glucose (0.5%) and lactose (0.5%). An in vivo study conducted by the same authors also showed that viable counts of lactic acid bacteria from both small and large intestines of rats fed with honey were markedly higher than those from rats fed with sucrose.

On the other hand, Varga reported that none of the 1%, 3% and 5% (w/v) acacia honeys added in yogurt had stimulatory or inhibitory effect on Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus in the yogurt. This was partly in agreement with a report by Chick in which L. acidophilus and Bifidobacterium bifidum in addition to the same species (S. thermophilus and L. delbrueckii subsp. bulgaricus) were neither positively nor negatively affected by 5% (w/w) clover honey added in skim milk in comparison with other sweeteners, although the production of lactic acid by bifidobacteria appeared to be significantly enhanced by the honey. However, so far there is no report showing that honey is detrimental to the normal flora, perhaps due in part to the antibacterial activity of the reported honey not being as significant as that of manuka honey.

The antibacterial activity of manuka honey against gastrointestinal organisms has been studied However, all these reports only considered the efficacy of the honey against one or more specific enteropathogenic species but did not take the possible impact of the antibacterial activity of manuka honey on the normal flora into account. Since manuka honey revealed a significant antimicrobial activity against a wide range of microorganisms, it is also possible that the honey may be inhibitive to the probiotic gastrointestinal microflora.

Possible Mechanisms that May be Involved in the Treatment of Gastroenteritis with Manuka Honey

Rehydrating Electrolytes

Oral rehydration therapy, with a solution of glucose and electrolytes, is recommended when acute gastroenteritis occurs because the gastrointestinal tract re-absorbs large amount of fluid daily to maintain the balance of electrolytes in the body, and spontaneous dehydration due to vomiting or diarrhoea can be more critical or fatal than infection itself. In this aspect, honey can be useful because it consists of sugars and also contains ions. Additionally, although an electrolyte solution with glucose added (ORS) is recommended by the World Health Organisation in the clinical trial approached by Haffejee and Moosa an ORS in which glucose was replaced with honey (the glucose concentration and electrolyte content being identical to those in the ORS by WHO) seemed to be more effective than glucose-electrolyte solution in re-hydrating patients. Fordtran suggested that fructose may also promote the uptake of potassium, which is an extra advantage when treating gastrointestinal disorders with honey.

Repairing Damaged Mucosa

The second possible mechanism for honey to alleviate gastroenteritis is through repairing the damaged gastrointestinal mucosa, which may be involved simultaneously with the anti-oxidative and anti-inflammatory activity.

Inflammation and oxygen-derived free radicals have been thought to be involved in gastroenteritis, and honey has been reported to contain a number of anti-inflammatory and antioxidant components. These activities are largely relevant to the topical treatment of wounds with honey and both these activities may also apply to the gastrointestinal ulcers. Mahgoub reported in an animal trial that honey (5 g/kg) provided almost 100% protective effect on Wistar albino rats from acetic acid-induced colitis, whereas a mixture of glucose, fructose, sucrose and maltose did not provide any protection. The same research group also tested several biochemical properties and found that honey prevented depletion of the antioxidant enzymes, reduced glutathione and catalase. Bilsel also investigated the effect of honey on induced colonic inflammation in rats and found that with honey there was a significantly lower percentage in mucosal damage than with prednisolone, a drug commonly used to treat inflammatory bowel disease (P=0.04).

Prebiotic Effect of Manuka Honey

Honey also appears to have a prebiotic effect, which improves the growth of microflora that directly and/or indirectly help retard enteropathogenic infections. Ustunol and Gandhi reported that the growth of Bifidobacterium spp. were significantly improved in 3–5% (w/v) honey solution. Kajiwara demonstrated similar results and also observed that the mean doubling times decreased from 147–690 hours to 9.9–14.3 hours. Whilst Astwood and Sanz demonstrated that some honeys contained oligosaccharides, it is interesting to note that the effects of honey stimulating the growth of microflora presented in Kajiwara and Ustunol and Gandhi were higher than those of fructooligosaccharide, galactooligosaccharide and inulin.

Antimicrobial Actions

The indigenous antimicrobial properties in honey stated above may inhibit the growth of enterobacteria. As demonstrated in the study of Haffejee and Moosa in which the duration of the bacterial diarrhea was halved whereas the viral diarrhea was not, it is possible that the effectiveness of honey on treating diarrhea is partially contributed by the antibacterial activity, although this may also suggest that the antiviral activity in the honey was also partially contributing to this, as otherwise the duration of the viral diarrhea in this trial could have increased as antibacterial agents commonly do. Indeed, in vitro studies have shown that several honeys exhibit significant inhibition on gastrointestinal pathogens including Bacteroides, E. coli, H. pylori or Salmonella. Interestingly, the microbiological action may also be due to the prevention of organisms from adhering to gastrointestinal epithelial cells.

SUSCEPTIBILITY OF ENTEROBACTERIA TO MANUKA HONEY

As the aim of this study was to evaluate if manuka honey is likely to be useful for gut infections, it was necessary to first find how susceptible the gastrointestinal bacteria are to the honey.

In this chapter the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) of manuka honey and artificial honey to common pathogenic gastrointestinal bacteria were tested using the broth microdilution method. The susceptibility of probiotics was also measured to find if they would be likely to survive the treatment of enteric infections with honey.

Manuka Honey and Gastrointestinal Pathogens

Bacterial gastrointestinal disorders are one of the most common clinical diseases. These are largely caused by, but not limited to, Enterobacteriacea (E. coli, Shigella, Enterobacter, Yersinia), Salmonella, Campylobacter and H. pylori. Several “opportunistic” microorganisms such as C. difficile may be associated with gastroenteritis although they normally do not cause disease.

Honey has been reported to show a significant antibacterial activity against a wide range of bacteria, but those reports of efficacy of honey are not always comparable with each other for several reasons. On the one hand, the detail on the honey being tested is not clearly given. For instance, Badawy compared the antibacterial activity of four Egyptian clover honey samples, each of which had been stored for a different period of time (7 months, 12 years, 16 years and 21 years) against E. coli O157:H7 and S. typhimurium and reported that the activity declined with time. However, Badawy et al. (2004) did not state whether honey samples had been stored at room temperature or in a dark refrigerator but it is known that antibacterial activity of honey is sensitive to light and to heat. In some reports, even the floral source of the tested honey was not given but the activity of honey can vary greatly among different floral types. It has also been known that the potency of honeys sharing the same floral source could differ in activity by up to 100-fold. Therefore, it is necessary to have the antibacterial activity of honey standardized.

On the other hand, many published reports failed to give details on the bacteria being tested such as the inoculum density. As the more bacterial cells there are in the testing environment, the more likely for the bacteria to overcome the toxicity of antibacterial agents, it is important to state the number of the cells being involved in the assessment of an antibacterial agent.

The objectives of this study were to investigate the antibacterial activity of a standardized manuka honey against a number of standardized number (inoculum density) of type strains and clinical isolates of gastrointestinal pathogens from patients with diarrhea. Additionally, to understand the potential impact of the antibacterial activity of manuka honey on normal flora, the sensitivity of some probiotic species to manuka honey was also examined in this chapter. The data obtained in the study in this chapter were also used as the basis for the work in the following chapters.

RESULTS

Facultative Anaerobes

All the facultative anaerobes had a lower MIC value with manuka honey M115 than with artificial honey. Generally, manuka honey at a concentration less than 8% could inhibit the growth of the tested gastrointestinal bacteria. Although Enterobacter spp. had higher MICs than other tested microorganisms, the concentrations of manuka honey required to inhibit their growth were still lower than those of sugar syrup (approximately 11% cf >16%). The MBCs of manuka honey were generally higher than the MICs of the same honey by one or two dilution steps. The MIC and MBC of manuka honey M115 against the P. aeruginosa isolate were found to be relatively higher than those against other species of bacteria (approximately 16.6%).

Microaerobes

The results showing the sensitivity of Campylobacter. The susceptibility test revealed that the growth of all 29 species of Campylobacter were strongly inhibited by manuka honey M113. The MIC of manuka honey for Campylobacter ranged from 0.8% to 1.1% whereas that of artificial honey was 3–4 times higher than that of manuka honey (3.1–4.3%), revealing that the MIC of manuka honey for each strain was significantly lower than that of artificial honey (P < 0.05).

The MIC of manuka honey M115 for H. pylori was significantly lower than that for artificial honey (5.64 ± 1.57 cf. 16.6 ± 0; P < 0.05). Given that the potency of manuka honey M113 used for Campylobacter spp. was 1.78-time (29.4/16.5) that of manuka honey M115 which had been used for H. pylori, we may presume that the MIC of manuka honey M115 for Campylobacter spp. would be about 1.42–1.96 %. It may consequently be presumed that the MIC of manuka honey with the same potency would be lower for Campylobacter spp. than for H. pylori.

The subculturing after determining the MIC showed that growth occurred when subculturing from concentrations of honey below the MIC whereas there was no growth from concentrations at and above the MIC. This revealed that the MIC of either manuka honey or artificial honey was also minimum bactericidal concentration (MBC) for H. pylori and for all of the Campylobacter spp. in this study.

Anaerobes

The MIC of M113 manuka honey for C. difficile ranged between 2–3.6%, and the MBC between 4–5.4%. The C. difficile tested were much more resistant to artificial honey than natural honey, as can be seen from the high MIC (> 16.6%).

Probiotics

The results for the sensitivity of the anaerobes to manuka honey M115 are shown in Table 4.5. The results revealed that these species are relatively tolerant to both manuka honey M115 and artificial honey. This is especially obvious in B. animalis subsp. lactis and Lactobacillus spp., for which there is no statistically significantly difference in the MIC values between the two types of honey for each species (P > 0.05).

The MBC of both manuka honey M115 and artificial honey for all tested probiotics are very high. For all species a concentration of higher than 13.3% of manuka honey was required to kill the probiotics. Because higher than 16.6% of manuka honey M115 seems to be required to kill B. animalis subsp. lactis and Lactobacillus spp., it was not possible to compare the MBC values of the manuka honey with those of artificial honey.

DISCUSSION

Because some of the species in this work were done with double-potency manuka honey (M113; with non-peroxide antimicrobial activity equivalent to that of 29.4% phenol) whereas other species were done with manuka honey M115 which is equivalent to 16.5% phenol, the MIC and MBC values obtained in this chapter cannot be compared directly. Despite this, it is possible to convert the mean MIC and MBC values for species tested with manuka honey M113 by multiplying them by 1.78 (i. e. 29.4/16.5) so as to compare them with the MIC/MBC of manuka honey M115.

Overall the average concentration of manuka honey required to inhibit the growth of most organisms tested in this study was very low, which agrees with most other studies that tested or reviewed the efficacy of manuka honey against organisms with manuka honey. The exceptionally high concentrations required to control P. aeruginosa are probably because of the biofilm, as has been widely known to be linked to its resistance against antimicrobial agents.

Most publications only considered the MIC of the honey for microorganisms and rarely did they estimate the antibacterial property further. In this study the MIC and the MBC of the total activity of honey for some common gastrointestinal microorganisms were tested.

Facultative Anaerobes

Most facultative anaerobes tested were inhibited by the manuka honey with antimicrobial potency near the median level (equivalent to approximately 16% phenol) even when the honey was diluted 10-fold or more. On the other hand, the artificial honey which imitates the sugar content of a normal honey failed to inhibit the growth of all tested microorganisms even at the highest concentration used in the test (16.6%). This suggests that it is not the osmolarity but other antibacterial factors in the honey that are responsible for the inhibition of the growth of the bacteria.

In this susceptibility test, S. typhimurium DT104, a multi-antibiotic resistant strain of Salmonella, was included because the prevalence has been increasing for the last few years, and it is revealed that this strain can be inhibited by manuka honey at a concentration less than 10%. Although antibiotic-resistant microorganisms are being increasingly reported and the number of resistant bacteria tested with honey is still limited so far, the efficacy of manuka honey against these seems promising. Prior studies have shown that honey is effective on MRSA and VRE.

Interestingly, bacteria with antibiotic-resistance properties were reported as soon as antibiotics were commonly used among the medical professions, whereas resistance to honey has not been reported regardless of it having been used as a medicine for millennia. The fact that bacteria have failed to develop resistance to honey is perhaps partially because the antibacterial efficacy of honey has not been widely known by medical professionals, and also partially because many of the reported antibacterial activities of honey being used had not been well standardised using a reference antiseptic. The result has been that a wide range of MIC values have been reported in the literature (in some cases these have ranged from less than 20% up to 100% for the same bacterial species), and consequently it is impossible to detect whether or not resistance to honey has developed.

Microorganisms are unlikely to acquire antibacterial resistance if they are treated with compounds targeting multiple loci. Honey is a complex substance and the antibacterial activity is multi-factorial. Rose Cooper at University of Wales Institute has conducted a long-term study to select wound pathogens resistant to manuka honey by continuously exposing bacteria to a sub-lethal concentration of honey, but honey-resistant bacteria have not yet developed successfully. A similar study conducted at Sydney University also failed to develop honey-resistant strains of S. aureus and P. aeruginosa whereas these bacteria had increased their resistancy to other antibacterial agents under similar condition in the study.

Although the number of antibacterial-resistant strains that have been tested with honey is relatively limited and whether or not bacteria would eventually develop resistance to honey was not evaluated in this study, the sensitivity of S. typhimurium DT104 to manuka honey in this study may partially, if not fully, suggest the usefulness of honey on treating multi-resistant strains of bacteria increasingly seen in medical disciplines.

Clostridium difficile

Although the sensitivity of C. difficile to manuka honey was determined with honey with high activity (M113), the MIC and MBC of honey with median activity (M115) for the anaerobes can be estimated to be about twice the MIC or MBC of that shown in Table 4.4. This would give MIC values between 4% and 7.5%, and MBC between 8% and 11%.

The sensitivity of C. difficile to antimicrobials including honey can be somewhat controversial because drugs can be effective against the bacterial cells but not the endospores. Although the C. difficile tested did not reveal resistance to manuka honey, the efficacy of honey against its endospore is unknown. In fact, even a high concentration of manuka honey is unlikely to be effective against bacterial endospores. Honey is stored aseptically in bee hives because of the enzymatic production of hydrogen peroxide during the ripening. When beekeepers collect honey from honey combs, the honey is exposed to the ambient environment and the outside of the honeycomb which may increase the chance it being contaminated by microbes. Microbial contamination is of no concern in an undiluted honey because of the low aw in which most microbes cannot survive. However, spores introduced into honey are not killed because spores can survive in high osmolarity, and although hydrogen peroxide is sporicidal at 0.88 M, it is not produced until honey is diluted because the water activity is too low for glucose oxidase in honey to work. Even if honey is diluted so that the production of hydrogen peroxide is activated, the typical concentration (1 mM) is still far too low to kill spores. Indeed, bacterial spores, particularly the Bacillus genus and Clostridium botulinum, are regularly found in honey, and this is why honey is usually not recommended to be fed to infants. Although bacterial spores in honey are unlikely to germinate, this indirectly indicates that honey may not be useful for eradicating the bacterial endospores in the gut.

Nonetheless, because it is viable bacterial cells but not spores that cause C. difficile-associated diarrhea, the resistance of endospores to honey may not be of practical concern. In modern medicine, metronidazole or vancomycin has been used to treat severe C. difficile-associated diarrhea but none of the drugs target the clostridial spores. Metronidazole is selectively taken up by anaerobe cells, reduced by indigenous enzymes, and the end product disrupts the DNA helical structure. Vancomycin, on the other hand, acts by inhibiting proper biosynthesis of cell wall peptidoglycan in a Gram-positive anaerobe. Clostridial bacteria or their spores could have existed in the gut of a host, but only when the clostridial cells germinate and outnumber other competitors in the gut can the species cause disease. It is possible that the antibacterial activity of manuka honey can be of some help to patients suffering from C. difficile-associated diarrhea by suppressing the bacterial cells. Perhaps the capability of honey to stimulate the growth of probiotics may also indirectly help suppress the clostridia, although this hypothesis needs to be further investigated.

Probiotics

Unlike most facultative anaerobes, microaerobes and C. difficile for which the MIC of the manuka honey with median-activity ranged between 4.5–8%, 2.5–6% and 4–7% respectively, all tested probiotics gave a relatively high MIC value (equivalent to or above 10%). Together with the even higher MBC values (>16%), this indicates that it is unlikely that manuka honey would kill the probiotics in the gut. Although the number of tested species is limited, this suggests that probiotics might be relatively more tolerant to the honey than gastrointestinal pathogens.

The tolerance of anaerobic probiotics to manuka honey is unexpected. Honey, particularly manuka honey, is frequently reported to inhibit pathogens due to several mechanisms, among which hydrogen peroxide may be the major factor that decreases the growth of non-aerobes. Microaerobes and anaerobes do not tolerate the damage caused by reactive oxygen species because they produce no, or small amount of, catalase and superoxide dismutase. B. animalis subsp. lactis and Lactobacilli are considered to be anaerobes and, in theory, they should have revealed equivalent to or higher sensitivity than facultative anaerobes to honey. Although some probiotics have been known to produce hydrogen peroxide as a defence mechanism to compete against other microbiota in the gastrointestinal tract, which may indirectly indicate the relative tolerance of the probiotics of the peroxide activity of the honey, several other possible reasons can also be featured. Firstly, probiotics, especially Bifidobacterium spp., require glucose as their energy source and honey is rich in glucose. This may have compromised the damage the honey caused to the bacterial cells. However, MRS medium used for cultivating probiotics in this project contains 2% glucose that should be sufficient for the organisms, therefore lack of glucose in the medium for supporting their growth is unlikely. Secondly, some components such as galactooligosaccharide and fructooligosaccharide in honey may have functioned as prebiotics so that the growth of the probiotics was stimulated even in the existence of the animicrobial activity of the honey, although further investigation is required to prove the hypothesis. Thirdly, the tested probiotic species may have some mechanisms that overcome, or at least tolerate, the antibacterial factors of manuka honey. As described above, glucose (and hence osmolarity) is not inhibitive to the species. Bifidobacterium and Lactobacillus spp. are also known to be acidotolerant, which was used to design the selective MRS media for these species (it has a pH of 5.7), and this could have made the acidity of honey to be less inhibitive in probiotics. The effect of methylglyoxal in honey to probiotics as well as on other gastrointestinal bacteria is, unfortunately, not well known yet.

The sensitivity results for probiotics in this chapter is in agreement with recent research by Rosendale. Rosendale studied the effect of several traditional medicines including manuka honey on probiotics (L. rhamnosus, Lactobacillus reuteri and B. animalis subsp. lactis) as well as enteropathogens (E. coli, S. typhimurium and S. aureus) by means of observing their growth turbitimetrically rather than determining the MIC. In that report Rosendale denoted the activity of the manuka honey as “UMF20+” and did not state what percentage of phenol the non-peroxide activity was equivalent to. According to the unique manuka factor rating system, a manuka honey labeled UMF 20+ means that honey has a non-peroxide antibacterial activity equivalent to equal to or greater than 20% phenol, which the activity is significantly higher than that of the manuka honey used in this thesis. Despite the high antibacterial activity being used, the group found that all probiotic growth increased while that of enteropathogens decreased. Interestingly, the same authors also noticed that some other traditional medicines (bee pollen, rosehip, blackcurrant oil and propolis) had either synergistic or antagonistic interactions on the effect of the honey on probiotics depending on the combination of the traditional medicines used. Although there has been very limited research on the effect of manuka honey on probiotics in the literature, this small project in this thesis as well as the work by Rosendale suggest the possibility of probiotics being more tolerant to manuka honey than enteropathogens are.

CONCLUSION

In short, most gastrointestinal bacteria are susceptible to the antimicrobial activity of manuka honey but not to artificial honey. The isolated C. jejuni and C. coli were found to be exceptionally sensitive to manuka honey. This may be partly due to their high sensitivity to osmotic action. Most tested organisms can be inhibited by manuka honey even if it is diluted 10–20 fold, and can also be killed with slightly higher concentration of the honey. Probiotics which we do not want to kill are found to be more tolerant to manuka honey than gastrointestinal pathogens are.

THE EF FECT OF ATMOSPHERE

As part of the antibacterial activity of a honey is due to hydrogen peroxide, and because the oxygen is required to get glucose oxidase in a diluted honey solution to transform the glucose content to hydrogen peroxide and gluconic acid (Equation 1.1), it is possible that in the anaerobic gastrointestinal environment this reaction would be suppressed, which consequently could lead to a lower total antibacterial activity in manuka honey.

This chapter describes the study of the effect of gastrointestinal atmosphere on the overall and non-peroxide antimicrobial activities of manuka honey. The MIC and MBC of a normal manuka honey as well as the one treated with the enzyme catalase are compared in aerobiosis and anaerobiosis.

Hydrogen peroxide is one of the major antibacterial factors in honey, and it is produced, slowly, only when honey is diluted (Equation 1.1). When humans ingest honey, ingested water and body secretions dilute the honey and consequently, in theory, hydrogen peroxide is produced in the gastrointestinal tract.

However, oxygen is also required to produce hydrogen peroxide and this may not be available in the gastrointestinal environment because that is generally anaerobic. It is in fact hard to predict whether there would be any effect on the antimicrobial activity of honey that is result from a less oxygenic environment such as in the gastrointestinal tract. Since oxygen is unavailable or limited in the tract, the actual antibacterial activity in honey could be at least impaired due to hydrogen peroxide being unable to be produced.

Environmental atmosphere, on the other hand, can also affect bacterial metabolism and consequently the growth of microorganisms which makes prediction more difficult. Facultative anaerobes can survive in both aerobic and anaerobic conditions by switching their metabolism, and because aerobic respiratory metabolism is more efficient than anaerobic fermentative metabolism, the growth of facultative anaerobes under anaerobic conditions is normally slower than that under aerobic conditions. As the mechanism of action of non-peroxide antibacterial substances in manuka honey remains to be understood, several possible scenarios were foreseen. Faster growth of a microorganism in an aerobic environment may lead it to overcome some levels of the negative effect caused by antibacterial substances in honey; in other words the cells may be more susceptible in anaerobiosis. The aerobic environment could also accelerate the metabolism of bacterial cells, which may therefore speed up the bacterial cells’ uptake (hence make it more sensitive) or efflux and/or metabolism (to detoxify) of extracellular antibacterial components. Yet, it is also possible that the shift in the overall effect is not significant due to the overall uptake and efflux being balanced in the cells.

In a preliminary test, an agar well diffusion trial using some facultative anaerobes was conducted to study the effect of anaerobic incubation on the activity of honey. It was intended to compare the inhibition zones seen on nutrient agar plates in aerobic, microaerobic and anaerobic conditions. However, in this preliminary test, inhibition zones were seen only in aerobic and microaerobic conditions. In contrast, a relatively opaque zone instead of an inhibition zone was seen in the anaerobic condition. As the observed opaque zone around the well was denser than the background, it was thought that the dense zone was an indicative sign of the stimulative effect of honey on the growth of the seeded bacteria. The reason why the antibacterial activity in honey did not take effect in the anaerobic condition in this preliminary test is not clear. Although there would have been no ongoing production of hydrogen peroxide in GasPak canister, non-peroxide antibacterial substances should have taken over the antimicrobial action. In the anaerobic incubation the growth of the tested facultative anaerobe was significantly slower than those in aerobic or microaerobic conditions, which could be observed from the size of the colonies impregnated in the agars. Perhaps the antibacterial substances in the honey had diffused out of the well so that the concentration had been diluted down below the MIC before the substances could inhibit the slow-growing bacteria. Also at the same time, sugar or other components in honey could have helped the bacterial cells to overcome the undesirable environmental stresses (peroxide/nonperoxide activity from honey and low oxygen level in this preliminary test).

Because this informal preliminary test had suggested that the agar well diffusion technique may be not be suitable for evaluating the effect of environmental atmosphere on the antibacterial activity in honey, in the work in this chapter the MIC/MBC values were obtained using the broth microdilution method instead.

Results and Discussion

In this chapter, the effect of atmospheric condition (anaerobic environment) on the antibacterial activity of manuka honey was studied to evaluate the likely efficacy of the honey on the pathogens in the gastrointestinal tract. As far as the author is aware, this is the first study that looks at the alteration of the effectiveness of manuka honey in an oxygen-limited environment. Many clinically important gastrointestinal pathogens have been extensively studied in aerobic conditions, but no consideration on the efficacy against the facultative anaerobes in a low oxygen concentration has been given. In fact, during the study we noticed that in all of the treatments, bacterial growth under anaerobic conditions was much slower than under aerobic conditions as can be observed visually from the size of the pellet of bacteria settled at the bottom of the microplate wells, suggesting the possibility that the anaerobic atmosphere could indeed alter the cells’ morphology and therefore the actual sensitivity to an antimicrobial.

The MIC of honeys in aerobic and anaerobic conditions to be compared. ANOVA reveals that overall both atmosphere and honey type significantly effect the MIC (P < 0.05) for each. The MIC values for manuka honey in anaerobic condition appeared to be equivalent to or lower than those in aerobic condition, and the degree of difference varied depending on the species (varied from less than 1% to more than 4%). This suggests that bacteria are slightly more susceptible to the total activity of manuka honey under anaerobic condtions, with E. faecalis being an exception which had a numerically but not statistically higher MIC in anaerobic condition.

To investigate the effectiveness of the non-peroxide activity in manuka honey, the enzyme catalase was added into the manuka honey to remove any activity due to hydrogen peroxide. The MIC results show that bacteria have a slightly higher sensitivity to the non-peroxide activity under anaerobic conditions, a similar trend that has been previously shown in the total antibacterial activity under anaerobic conditions. The non-peroxide antibacterial activity in manuka honey has now been identified as being due to methylglyoxal, and the MIC may represent the effectiveness of methylglyoxal. However, it is noteworthy that methylglyoxal may combine with the N-terminal segment of proteins that generally exist in a medium and that may result in extra amounts of methylglyoxal being required to interact with bacteria. Therefore, it may be questionable to regard this as the “MIC for methylglyoxal”.

No inhibition on the tested microbes was observed in artificial honey in both atmospheric environments (all MIC values were C16.6%). This indicates that the osmolarity of honey would not practically inhibit bacteria both in intestinal or external conditions. Although there would be some difference in the actual MIC values if the range of honey concentrations tested had been wider, that still would not be of very much practical importance because the concentration of sugar in the gut is unlikely to reach a high level.

Some interesting aspects can be seen when the sensitivity of each bacterial species to the bacteriostatic activity of manuka honey with and without hydrogen peroxide removed are compared. For some species (E. coli, P. aeruginosa and Salmonella) the honey with catalase added revealed equivalent or marginally higher MIC than a normal manuka honey (P>0.05), whereas for some other species a much higher concentration of the manuka honey with catalase was required to inhibit their growth (P<0.05). Unexpectedly, E. faecalis and E. faecium were found to be slightly more sensitive to manuka honey with catalase than without catalase, although this is numerically but not statistically different (P>0.05). Perhaps this numerical increase in the sensitivity result of Enterococcus spp. is simply because of the limited number of the species being examined. Alternatively, it may be that the catalase used is toxic to the bacteria. More species samples would be needed to ascertain this phenomenon.

On the other hand, the bactericidal profile of manuka honey appeared to be largely different from what has been described above for the MIC values, as the atmosphere factor appears to account less for the effect on the MBC (P > 0.1). More species (E. coli, P. aeruginosa, Yersinia, S. enteritidis and E. aerogenes) gave higher MBC values with normal manuka honey in anaerobic than in aerobic condition (P<0.05), whereas equivalent or lower MBC values were seen in S. typhimurium DT104 and E. faecalis (P> 0.05), and significantly lower in E. faecium, E. cloacae and Shigella (P<0.05). This was same for the MBC values of manuka honey with catalase except E. faecalis (significantly higher MBC value in anaerobic condition; P<0.05) and E. cloacae (numerically higher MBC value in anaerobic condition; P>0.05).

The MBC values for manuka honey with catalase had yet higher values than for those without catalase for all atmospheric conditions. Again, E. faecalis in air was the only exception (significantly lower in aerobic condition; P<0.05). Some of the tested microorganisms were marginally more resistant to the bactericidal effect of the non-peroxide activity in the honey (E. coli in air and E. faecium in both atmospheres; P>0.05), but for most species (E. coli and E. cloacae in anaerobic, Yersinia, Salmonella and Shigella) a much higher concentration of honey was required to kill the cells when hydrogen peroxide had been removed (P<0.05).

Although P. aeruginosa is usually considered an aerobe, and in the MIC test this species indeed did not show any growth in anaerobic condition, after subculturing on an agar plate in the air P. aeruginosa was found to be still surviving.

Some species of bacteria showed equivalent sensitivity to manuka honey with or without hydrogen peroxide present, whereas others showed up to 4 times higher MIC values if catalase was added in the honey. This suggests that the non-peroxide antibacterial factors contribute most of the antibacterial activity of manuka honey because otherwise the activity should have been eliminated by enzyme catalase and should have given high MIC values similar to those with artificial honey. Although Enterobacter spp. appeared to be relatively tolerant to the non-peroxide activity, they still were more sensitive than to artificial honey, which suggest that the activity can still be useful to inhibit these species. The MBC test, on the other hand, suggests that the non-peroxide activity of manuka honey has a relatively “mild” effect on bacterial cells. It can be seen in Tables 5.1 and 5.2 that the MBC of the total activity of manuka honey was higher than the MIC and, with some species, catalase appears to destroy the activity much further. If the MBC of manuka honey with and without catalase is compared, it is clear that to some species the concentration of the non-peroxide activity required to kill the cells is more than twice of that of the total activity. The increase in the MIC and MBC values indicates that hydrogen peroxide actually plays a role in manuka honey to inhibit bacteria, and this effect seems to be more significant with some species of bacteria. Manuka honey is usually mistakenly called “non-peroxide honey” in literature for the public. Manuka honey can have both peroxide and non-peroxide activities, although some species appear to be less sensitive to the hydrogen peroxide in the honey.

Although the difference in the MIC/MBC between different atmospheric environments could have been because of enzyme glucose oxidase in honey was unable to convert glucose into hydrogen peroxide in the absence of oxygen, the anaerobic environment might have also caused environmental stress to the facultative anaerobes which may lead to some morphological adaptations in the cells.

To inhabit in virtually every area in the world, prokaryotes have had to develop appropriate mechanisms to survive or thrive under a wide variety of conditions. Not only do they need to be able to inhabit some specific environments (e.g. hyperthermophiles, halophilic/acid-tolerant bacteria), but they also need to be able to adjust to sometimes prolonged extreme conditions. The air can actually be considered an extreme condition because without an appropriate amount of protective enzymes (catalase and superoxide dismutase) cells would be damaged by oxidative stress caused by ROS. Many bacterial stress responses have been found to be genetically regulated. When bacteria transit from an aerobic environment to one that is anaerobic (for instance, when gastrointestinal pathogens enter the gastrointestinal tract), there would be a change in genetic expression which is regulated by the f umarate and nitrate reduction (FNR) regulatory protein. FNR is in a dimer form and binds stably to DNA in an anaerobic environment and, in an aerobic condition (with 1–10 μM oxygen) the structure turns into monomers and decreases in the ability to bind DNA. With the FNR regulation system, organisms in an anaerobic environment can reduce the production of enzymes involved in aerobic metabolism and increase that required for anaerobic respiration.

Atmospheric factors have been reported to significantly affect certain antibiotics. Rosenblatt and Schoenknecht reported that gentamicin, streptomycin, kanamycin and erythromycin were less active in an anaerobic condition whereas tetracycline and chloramphenicol revealed higher activities against E. coli. Goldstein and Sutter noted that active erythromycin deteriorated by more than 50% after 18 hours incubation in CO2.

Environment atmospheric stress may also result in the modification of the morphology of bacterial cells which consequently alters the sensitivity to antimicrobials. P. aeruginosa, either in planktonic or in biofilm forms, appeared to be more resistant to antibiotics if oxygen was limited, whereas anaerobically grown E. coli K-12 was reported to fail in biofilm formation. Likewise, it is possible that the FNR or other gene regulations regarding environmental stress response may affect the sensitivity of bacterial cells to antibiotics or antibacterials such as hydrogen peroxide or methylglyoxal in honey. This, together with the response to other stresses such as pH and enzymes.

THE PHARMACODYNAMICS OF MANUKA HONEY

In this chapter the pharmacodynamics of the antibacterial activity of manuka honey is described. The hypotheses are: that the pharmacodynamics of manuka honey are like those of conventional antimicrobial agents; that the number of viable organisms continuously exposed to a constant concentration of manuka honey declines as exposure time passes; and that after bacterial cells have been exposed to manuka honey for a short period the antimicrobial effect of manuka honey persists for a while even after the removal of the honey.

In the first part, the time-to-kill test that evaluates the time required to eliminate bacterial cultures persistently exposed in a constant level of manuka honey was investigated. The second part of this study investigated the postantibiotic effect of manuka honey in which bacterial cells were exposed to the honey temporarily and the survival of the treated cells after that was monitored.

Introduction

Antimicrobial agents are capable of inhibiting or killing organisms if their concentrations are above the MIC or MBC, and this has widely been used for evaluating the efficacy of drugs. The principle of the MIC test is to expose the targeted micro-organisms to an antibacterial agent of interest at a series of concentrations of antibacterial agents for a pre-defined time (usually overnight) and monitor the growth of the organisms. The minimum concentration which inhibits the growth of the organisms is then recorded as the MIC. Similarly, the minimum concentration of the antimicrobial agents that is required to kill the tested organisms is termed the MBC. This is usually approached by subculturing the broth media from the MIC test to a fresh solid medium on which microbes may still be alive.

In the MIC test the concentrations of antibacterial agent remain at a high or constant level through the assay. However, this is clearly not feasible in vivo for several reasons.

On some occasion in vivo it is desirable for an antimicrobial agent to take effect on the bacterial cells in a short time because the actual exposure time of the cells to the active ingredient may actually be much shorter than is the case when the MIC test is conducted in vitro. A significant example is the usage of antibiotics in ophthalmology. As eyedrops are usually washed off within minutes after being applied on the infected eyes, the bacteriostatic and bactericidal information obtained from the MIC/MBC tests alone are not useful for evaluating the efficacy of antimicrobial agents in treating such cases. Several efficacy assays of the antibiotics used for H. pylori eradication have been criticised to be inadequate because they fail to control the viability of the organisms within 3 hours. Similarly, MIC/MBC values usually do not account for the significantly fluctuating drug concentration within a body as time passes due to various factors in the host. The active amount of antimicrobial agents may decline, as can be seen on the degradation of penicillins, cephalosporins, cephamycins, ertapenems and carbapenems by ESBL (extended spectrum -lactamase) secreted by Gram-negative bacteria such as E. coli, Klebsiella pneumoniae, Proteus mirabilis, P. aeruginosa and some Enterobacteriaceae. The active ingredient may also be substantially diluted by large amount of body fluid or exudate, and therefore the total concentration eventually goes below the MBC or even MIC. Alternatively the active ingredient may be excreted out of the body through the urinary system before it has time to be of effect on bacteria. Therefore the efficacy of antibacterial agents in treating infections may not be as good as that indicated from in vitro susceptibility tests.

The situation described above is not limited to chemical antibiotics but is applicable to honey too. The efficacy of honey is generally considered to be lower than that of antibiotics, which also means the concentration required to suppress the growth of bacteria is higher than that of antibiotics (approx. 100 mg/ml cf. μg/ml). In wound treatment the active component in honey would diffuse into the wound from the dressing, and eventually the active concentration would decline to that below the MIC to the pathogens on the wound. Likewise, honey intaken orally would be greatly affected by dilution by large amounts of body fluid e. g. saliva, gastric juice and intestinal fluid and water from food and drink. A short transit time due to rapid peristalsis in the gastrointestinal tract may also result in a short period of contact with bacterial cells. Although it is unknown how much of the antibacterial activity of honey exactly would be able to reach the stomach or the lower intestine so as to inhibit gastrointestinal pathogens, it is very clear that the achievable concentration would be much lower than the initial concentration taken in. To evaluate the antibacterial efficacy of honey on bacteria in such conditions, several pharmaceutical parameters other than the conventional sensitivity test are required.

In pharmacology several parameters, namely pharmacokinetics and pharmacodynamics (PK/PD), are used to evaluate dynamically the efficacy of a drug. Pharmacokinetics in antibiotic research monitors the fluctuation of the concentration of the antibiotic after delivery into a living host. As the drug concentration during the treatment can be affected greatly by biological metabolism e. g. preantibiotic, the pharmacokinetics evaluation is usually approached by quantitatively analysing the parent drug and its metabolites using chromatographic systems and producing serum concentration versus time curves. Pharmacokinetics, however, is not always possible because this parameter involves monitoring the level of the known substances, and evaluating the pharmacokinetics of a complex substance of unknown composition such as manuka honey is not feasible.

Pharmacodynamics, on the other hand, is used in antibiotic research for predicting the eradication of bacteria by an antibacterial agent. It describes the exact effect of an antimicrobial substance on bacterial cells and thus is increasingly used for evaluating the mode of action of antibacterial agents and for predicting bacterial eradication. Several suggested pharmacodynamics parameters include, but are not limited to, the area under the plasma concentration-time curve (AUC), the drug’s peak concentration (Cmax), and the time the drug concentration exceeds the MIC (T > MIC). The effectiveness of a number of antibiotics has been well predicted by these models.

The postantibiotic effect (PAE) is one of the pharmacodynamics parameters. It describes the persistent suppression of bacterial growth after a short exposure of bacteria to antibacterial agents. This was first systematically observed by Parker and Marsh who noticed that after staphylococci had been exposed to penicillin, the organisms did not resume multiplying immediately after the removal of the drug but instead remained constant in number for a while, before beginning to multiply again at their normal rate. The duration of the lag seems to depend on individual bacterial species or even strains, the concentration of the drug and the time for which bacteria have been exposed to the drug. Eagle and Musselman tested several cocci including - and -haemolytic streptococci, staphylococci, pneumococci and enterococci with penicillin, and noticed that the persistent effect of the drug reached a maximum level at the concentration that was most rapidly bactericidal for the organism. In this study the authors also found that beyond the maximum effect, the persistence of effect did not increase further even with a 10 000-fold increase in the level of the drug. Similarly, the persistence reached its maximum within one to two hours of exposure time, and even 24 hours of exposure did not significantly prolong the period of persistence. Interestingly, for some strains the duration of persistence increased directly as the proceeding exposure time to penicillin increased. Eagle noticed that once a bacterium had been treated with a large enough amount of antibiotic even for a short time, the cells were ‘bombarded’ so that the microorganism required extra time to recover from the damage, and meanwhile the cells would not replicate until they had completely recovered from the damage, regardless of the absence of antibiotics. In other words, antimicrobial agents may not need to be present continuously during treatment of infections. Indeed, this is one of the main principles when estimating the interval required between dosages of a drug. A thorough review of the persistent effect of antimicrobial agents has been published by Craig and Gudmundsson.

The principle of assaying the postantibiotic effect is to compare the time required for bacteria treated within an antibacterial agent and for untreated control bacteria to reach the same growth speed as each other. This is generally done by exposing bacteria to a medium containing an antimicrobial agent for a short term (normally 1–2 hours), followed by removing the antibacterial agent by means of centrifugation, dilution, inactivation or membrane filtration. The bacterial cells are then incubated in a fresh medium without the antibacterial agent and the growth curve is observed. It must be noted that because the PK/PD parameters may vary greatly among bacterial species, and also that because they may not be directly relevant to the MIC/MBC, it is unlikely that PK/PD parameters can be predicted with the usual susceptibility tests.

Although honey has been proven to have substantial antimicrobial activity even against the organisms that modern antibiotics fail to eradicate, this has mostly been demonstrated with conventional susceptibility tests or with observations during clinical trials. And despite the large number of published report showing the significant antimicrobial activity and efficacy of the honey, it is not known at all whether or not the honey would reveal any of the PK/PD parameters against microbes. This may be partially because honey is a complex mixture of carbohydrates, and it is unlikely to approach a pharmacokinetics study without knowing the exact antimicrobial substance. Also, there has been comparatively little interest in the PK/PD properties even of modern antibiotics until recently. Therefore it is important to study the pharmacodynamics of the antimicrobial activity of honey, so as to evaluate the utility of manuka honey for internal use. This information would also be of help for the medical professionals to understand the usefulness of honey if the pharmaceutical properties of honey could be studied in further details.

Therefore in this study it was intended to first investigate how long it would take manuka honey to eliminate micro-organisms, and to investigate if honey would have the postantibiotic effect that other common drugs do, so that a ‘honey-exposed’ microorganism would still be inhibited even if the honey gets diluted to a low level afterwards.

Materials

It was clear that the concentration of artificial honey required to inhibit organisms was much higher than that of manuka honey. Therefore only manuka honey (M115) was used in this pharmacodynamics assay.

Results

This section reveals the results of the time-to-kill assay which estimates the time required for manuka honey to kill E. coli, Salmonella spp., P. aeruginosa, Y. enterocolitica, Shigella spp., Enterobacter spp. and Enterococcus spp. when the bacterial cells are constantly exposed to honey solutions.

For most of the enteropathogens tested in this study, honey solutions with concentrations higher than the MBC or even MIC were capable of killing the majority (90%) of the organisms within a few hours. E. coli, Yersinia, Pseudomonas and S. typhimurium DT104 lost their viability within 2–4 hours in honey solutions whereas it took 4–6 h to kill S. enteritidis, Shigella, Enterobacter.

Some species, on the other hand, were found to be resistant to the bactericidal activity of manuka honey. Enterococcus spp., unlike the other gastrointestinal bacteria, appeared to be tolerant to the honey for more than 12 hours. E. faecium seems to be slightly more susceptible to the honey than E. faecalis but still the rate of decline in the cell density was much lower than with the other other gastrointestinal bacteria.

Discussion

The susceptibility to manuka honey (MIC) with several gastrointestinal organisms was evaluated using the microdilution method which measures the growth curve of an organism in honey solutions at a series of constant concentration through 18 hours. The method is widely used in many laboratories due to its convenience and simplicity, but this also draws some concerns from practical viewpoint. The growth curve only reflects the multiplication of bacterial cells in the media whereas it does not indicate if the cells are being killed or merely temporarily inhibited by the antibacterial agent. To evaluate whether or not the bacterial cells are killed when exposed to the honey for long time, the MBC test was also conducted. In that test, most enteropathogens were found to have MBC values at around 6–7% manuka honey whereas for some species the MBC value was rather higher (Pseudomonas, Enterobacter and Enterococcus spp. at approx. 13%; B. animalis subsp. lactis and Lactobacillus at over 16%), suggesting that some species may be relatively more tolerent to the antibacterial activity of manuka honey than are the other organisms.

This time-to-kill study, however, seems to have uncovered some hidden information which had not been found in the MIC and MBC studies. In theory, the time required for honey to kill bacteria should be longer if the bacteria reveal higher MBC of the honey because that would mean the bacteria are more tolerant to it. It might take some time for the honey to diffuse into the bacterial cells and have its bioactivity have full effect on the metabolisms in the cells. Comparing with the MBCs in aerobic condition, however, it can be seen that species with high MBC are not necessarily able to survive in the honey for long time. Among those species that gave high MBC result (approximately 16%), P. aeruginosa completely lost its viability within as short time as Salmonella typhimurium as long as the honey concentration is higher than the MBC. This suggests that P. aeruginosa, and perhaps other species that are not included in this research project, may be actually not as hardy as the impression the high MIC/MBC gives. As long as the honey concentration is high enough, it is possible to eradicate this species as easily as most of other tested facultative anaerobes like E. coli. In other words, the time-to-kill profile may not always be in agreement with the MBC for bacteria.

In contrast to other species, hardly any decline in viable cells could be seen with Enterococcus spp. after 11 hours incubation regardless of being constantly incubated even in a high concentration (20%) of honey solution. The number of surviving Enterococcus cells might eventually decline down to zero if the incubation is elongated. This would account for the finding that the MBC for this species was ranged from 13.3% to 16% honey, the incubation time in that experiment being 18 hours. The relatively high tolerence of this species in this study is in agreement with Lamont that Enterococcus is a rather resistant pathogenic genus in comparison with other microbes. According to the results obtained in this study, the honey would need to be kept in contact with the bacterial cells to eradicate this species. This is possible if the honey is used to eradicate enterococcal infection on topical wounds because it is very easy to treat the bacteria by constant cover of the infected site with honey wound dressings or honey gels; however this would not be feasible in the gastrointestinal tract because honey can be easily flushed, diluted or absorbed in the tract. On the other hand, since enterococci play a role in maintaining the microenvironmental ecology in the gut, perhaps this also suggests that some probiotics like gastrointestinal enterococci would generally not be affected by intake of manuka honey.

It must be noted that although manuka honey is now increasingly recognized as an exceptionally effective antibacterial agent, the overall time-to-kill results show that one should not expect manuka honey to eradicate or inhibit microbes within a matter of seconds or minutes. As shown in the time-to-kill test, manuka honey takes at least one hour to have a microbicidal effect on bacterial cells when at a concentration of 20% and a longer time if its concentration is lower. Practically, this also means that the MIC/MBC for gastrointestinal pathogens is not meaningful because the actual concentration of the honey is unlikely to be maintained above the MIC level for that length of time. In the next section, therefore, what would happen to bacterial growth if honey contacts the cells only for a short time is explored.

The Post-Antibiotic Effect of Manuka Honey on Organisms

Because of the complexity of the postantibiotic test, a brief outline of the whole process is firstly given here, and the exact details are described chronologically in the next subsection.

Firstly, a honey solution was made by preparing 10/9 time that of the desired concentration and transferred in a 15 ml centrifuge tube. MHB with the same volume as the honey solution was also prepared in another centrifuge tube as a control. For each species, a tube with honey and another tube for control were handled simultaneously (i. e. two tubes for one species). A broth culture with approximately 3×106 cfu/ml was prepared. The broth culture with 0.1× volume of that of the honey solution was then added so that the desired honey concentrations to be tested were obtained, and at the same time the final cell density would be approximately 3×105 cfu/ml, the cell density that had been used in other chapters. All tubes were incubated for 1 h to induce the postantibiotic effect, with viable cell counting being conducted right after the incubation. The bacteria were washed by centrifuging the tubes and replacing the supernatant with pre-warmed MHB. After this, all tubes were incubated with viable cell counting being conducted at hourly intervals until a significant increase in the turbidity was seen visually in the honey-treated tube.

To check the sterility, an additional centrifuge tube with MHB was also included. This MHB tube was handled in the same way as the testing group described above through the whole process, except that the bacterial broth culture was replaced with MHB.

Most the viable cell counting was conducted using the track dilution method as described in Section 2.6. Before the postantibiotic test being conducted, 0.9 ml of MHB in bijou bottles was prepared as the dilutent for serial dilution. The dilution range of the series was 10−1×, 10−2×, 10−3×, 10−4×, 10−5×, 10−6×, 10−7× and 10−8×. Also, the conventional streak plate technique was used to count the cell number in case the cell density was predicted to be less than the detectable range of the track dilution method (i. e. < 103 CFU/ml). This was done by spreading 0.1 ml and 1 ml of broth culture on two MHA plates. The detectable range of the cell number in the works in the postantibiotic effect test, therefore, would be from 10 CFU/ml to 1012 CFU/ml.

To control the growth speed of bacteria, all media and solutions used in this study were pre-warmed at 37°C before used.

Methods

The following describes the method used for the postantibiotic assay for each species in 20% manuka honey M115.

Results and Discussion

This section describes the result of postantibiotic effect (PAE) on enterobacteria caused by 20% (v/v) manuka honey. It must be emphasised that the postantibiotic effect test is to examine whether the exposure to honey has affected the metabolism of the bacterial cells that remain alive, so that their subsequent growth is slowed. In other words, the study is to investigate for how long the rate of growth after 1 hour exposure to and 1 hour washing process of honey (i.e. gradient of the log plots after t2) remains less than the control before they becoming parallel.

Traditionally, the postantibiotic effect assay is done by monitoring the recovery of antibiotic-treated bacteria in terms of CFU/ml. In carrying out this work here, several serious drawbacks with this assay came to light. It takes a long time to prepare, execute and finish the test, which results in taking up as long as almost one week to run one single test. The lengthy process during the test could largely increase the chance of contamination and mistake, both of which are critical because otherwise it is almost impossible to obtain a good quality colony counting results at the end. It was also found to be impossible to handle more than 2 species i.e. 5 tubes (two tubes for each species, plus one tube for sterility control) in a run because it would take too much time to handle the samples, and the time needs to be strictly controlled (particularly when washing honey off by centrifuging, and also when serially diluting and inoculating bacteria on agar plates hourly). The requirement in time control also had made it impossible to conduct any replication in each single test because this would greatly increase the number of apparatus and media, and therefore the handling time. The lack of replication could have led to the difference in the bacterial cell number at the starting point between the tube with honey treated and the tube for control in some species. Repetitive fluctuation in broth temperature may effect the growth of “recovering” bacteria. Not knowing the range of CFU during the incubation in fresh broth after washing means a wide range of serial dilution was needed (hence large amount of agar plates, diluent and handling time is required). Also, not knowing how long it would take bacteria to re-grow after honey is removed means a single test could take as long as or more than 12 hours. With all the issues described above, the estimation of the postantibiotic effect values therefore cannot be precise. Nevertheless, these data do allow the author to see that there are big differences between species in the postantibiotic effect of manuka honey.

Similar to the findings with the time-to-kill test, the results from the postantibiotic tests showed that the MIC/MBC may not fully describe the susceptibility of bacteria to manuka honey. The susceptibility test result showed that for most species the MIC was 5–10% honey and the MBC was 8.5–16% honey, whereas postantibiotic effects of manuka honey on these species seems to be not always proportional to the MIC or the MBC. Generally most organisms such as Salmonella spp. and E. coli have the MIC of 7–8% and the postantibiotic effect of around 2–2.5 h. For Enterobacter spp., for which the MIC and MBC values are relatively higher, a moderate postantibiotic effect of around 2–2.5 h was also obtained. For Y. enterocolitica the MIC and MBC were very low. With this species there was a longer postantibiotic effect than with most other organisms (nearly 4.5 hours). With P. aeruginosa it was expected that there would be a minor postantibiotic effect because of the high MIC for this species, whereas the results revealed that the postantibiotic effect was surprisingly long (more than 3.5 hours). On the other hand, hardly any postantibiotic effect was observed with E. faecalis and E. faecium (less than 0.5 hour). A similar result was seen with the probiotics B. animalis subsp. lactis, L. plantarum and L. rhamnosus.

The length of the prevention of bacterial regrowth by manuka honey may be due more to its bactericidal effect on bacterial cells than to the MIC. For comparison, the results of the MIC test as well as those of pharmacodynamics are placed side by side. It is also clear that the species with higher time-to-kill results require a very short time for recovery. It is possible that the honey concentration used in the time-to-kill and the PAE tests (20%) being just above the MIC for B. animalis subsp. lactis, Lactobacillus and Enterococcus may have been responsible for the long time-to-kill result and short PAE. A shorter time-to-kill and longer PAE would be expected if higher concentration of honey were used to treat the species. However, this hypothesis is not supported by the observations with P. aeruginosa because the results with P. aeruginosa should have had a similar trend as there was with these species. Also, for those species with short time-to-kill results, the honey concentration at just above MIC did not elongate the result in the time-to-kill studies as much as more than 10 hours. An alternative possibility is that perhaps manuka honey has different modes of bactericidal action on different species. Several antibiotics, for instance daptomycin, are known to have rapid bactericidal effects at or just above the MIC, and the situation observed in P. aeruginosa is somewhat similar to this, although a surprising phenomenon seen in daptomycin, the delay in the growth of microbes at concentration just below the MIC or even as low as 0.2 times the MIC for more than 12 hours is not observed in our study. It is possible that the bactericidal mechanisms in manuka honey may be especially effective with some species (P. aeruginosa and Y. enterocolitica in this study), and be moderately effective to most other species (E. coli, Salmonella, Enterobacter and Shigella) and be of relatively low effectiveness to the probiotics used in this study.

The postantibiotic effect and time-to-kill are increasingly important pharmacodynamic parameters in clinical areas because these provide information on optimising the therapeutic regime in response to the increasing prevalence of bacterial species resistant to antibiotics. Especially the former may help reduce the dose or increase the interval between doses by investigating how an antimicrobial agent affects bacterial cells after being removed.

THE EFFECT OF GASTROINTESTINAL ACIDITY AND ENZYMES ON THE ANTIBACTERIAL ACTIVITY OF MANUKA HONEY

In this chapter the effect of acidity and digestive enzymes on the antibacterial activity of manuka honey are described.

Introduction

Any ingested food would get into the colon through the stomach. At the same time, a large number of microorganisms can also be ingested: in addition to topical wounds, oral ingestion is thought to be the major “entrance” through which microorganisms can invade in a host. It is estimated that over 1010 bacteria each day may enter the host through the mouth, and therefore a defence system against these microbes is essential to maintain the host’s health.

In healthy individuals, before the food reaches the colon it is digested by the digestive system while passing through the tract. As it passes, the gastrointestinal tract secretes fluid with a range of functionality that helps degrade food ingredients into small molecules. The enzymes digest proteins, carbohydrates and lipids in the food. The gastric protease, pepsin, hydrolyses the amino-terminal side of the aromatic amino acids phenylalanine, tryptophan and tyrosine. The acidity, on the other hand, provides an optimum environment for the enzymes to react and also assists food degradation. Gastric digestion is then followed by intestinal digestion carried out by various enzymes such as trypsin, chymotrypsin, amylase and lipase, which work in association with bile salts. The pH in the chyme is spontaneously neutralized to approximately 7–8 by pancreatic fluid, so acidity no longer plays a role in digesting food in the intestine.

The gastrointestinal environment also plays an essential role as a defence line in the body, of which the gastric acidity may contribute the major part of the antiseptic action. The acidity in the stomach (pH 2) prevents bacterial colonisation on gastro-epithelial cells and also eliminates the majority, if not all, of the invasive organisms. It is estimated that the number of viable cells can decline from incoming 1010 bacteria to less than 103 CFU/ml (Wilson, 2008), although this number can also rise up to 105–106 depending on individuals and the time in a day, especially the time after a meal. The significant rise in pH in the small intestine, however, decreases the potential to inhibit the growth of bacteria so that the viable cells could increase up to approximately 1013–1014 in the large intestine. In a simulated digestion study it was reported that E. coli O157:H7 and S. flexneri are inactivated significantly in simulated gastric fluid, with the inactivation rate decreased as the dose of antacid increased.

Gastric enzymes may also affect the bacterial cells. However, the short transit time in the stomach may make the antiseptic effect of pepsin less significant than that of acidity. Although the small intestine is believed to be the main part of the digestion in a host because of the long stay in the tract, the antimicrobial effect of intestinal enzymes is also not obvious as can be seen from a study done by Gorbach. They reported that the estimated colonisation in the gastrointestinal tract in 18 healthy individuals increased from less than 50 CFU/ml in the stomach to 50–100 CFU/ml in the duodenum, to 100–1 000 CFU/ml between the upper jejunum and the upper ileum, and this then abruptly increased to more than 500 000 CFU/ml in the distal ileum. This can be because the digestive enzymes together with bile acids are absorbed by the tract, but it can also be because the indigenous microbiota have adapted to the intestinal environment.

The enzymic activity and the acidity of digestive fluid not only affects bacterial cells but may also affect the property of antimicrobial agents. In fact, several antibiotics have been reported to lose their stability or efficacy in the gastrointestinal environment, which is largely due to the acidity. An example is that amoxycillin and metronidazole appear to be stable at a normal gastric pH (1.0–2.0) whereas clarithromycin can be degraded at this pH in about 4 hours, and some proton pump inhibitors are therefore administered to ensure the acid-susceptible drugs are not inactivated by the acid during H. pylori eradication therapy.

In honey, hydrogen peroxide is one of the major antibacterial factors and its production (Equation 1.1) can be affected by many factors including heat, light and acidity. It is reported that the enzyme glucose oxidase that catalyses the reaction in which hydrogen peroxide is formed works well at pH 7 whereas the activity, hence the accumulation of hydrogen peroxide, is negligible at pH 3. Thus the gastric environment is obviously too acidic for the enzyme to produce hydrogen peroxide. Also, pepsin as well as other digestive enzymes and salts in the tract may destroy the enzyme and affect the non-peroxide antibacterial activity of manuka honey. Therefore it was necessary to investigate how well the activity of the honey would work when subjected to acidity and the enzymic factors. In this chapter, the influence of acidity and then that of gastrointestinal enzymes on the antibacterial activity of manuka honey is investigated.

The Effect of Acidity on the Antibacterial Activity of Manuka Honey

Before the study was carried out, it was considered to use buffered media to minimise the possible pH fluctuation due to the bacterial metabolism during the incubation. However, different buffer solutions have different optimum buffering capacity, and to cover the wide range of pH that would be used in this study (pH 2–8) it would be required to use media made up with different buffer solution for different pH range, and the different buffering acids and salts could also further influence the bacterial metabolisms and the activity of manuka honey. This would add more undesirable complexity in addition to the effect of acidity being studied. It is also reported that the pH of both inoculated and non-inoculated media do not fluctuate significantly during incubation unless it is anaerobically incubated for more than 24 hours. Therefore it was decided to make broth and honey solutions with various pH values by adjusting the pH with 1 M HCl or 1 M NaOH to minimise the possible variables involved.

It was also realised in a preliminary test that the range of the final concentration of honey in microplate wells, which is achievable by serially diluting a single initial concentration of honey with the dilution method used above, was too narrow to cover the possible MIC in this study. For example, the detectable range of MIC is 2.23–16.6% if 50% honey solution is used, and it was found that all bacteria could grow in 16.6% alkaline honey whereas they could not survive in 2.23% acidified honey. Because a range of pH values (pH 2–8) was to be examined simultaneously, it was required to expand the range of the final concentration in microplate wells. Therefore a binary dilution method starting with honey of a high concentration (40%) was used in this study. The range of concentrations of honey prepared in this study was 0.156%, 0.312%, 0.625%, 1.25%, 2.5%, 5%, 10%, 20%, and 40%.

The manuka honey used was the medium activity M115. Artificial honey was also used to distinguish the antibacterial efficacy of honey at each pH from antibacterial activity caused by osmolarity.

For each pH tested, MHB and honey solutions of the same pH were prepared. Aliquots of single strength MHB with different pH values were prepared by adjusting the pH with HCl or NaOH. Aliquots of honey with different pH values were prepared by firstly making a double strength honey solution and adding an equivalent volume of double strength MHB, and then adjusting the pH with HCl or NaOH. All solutions were sterilised by filtering through a 0.22 μm membrane.

To measure the MIC at each pH, the microdilution method was conducted, in triplicate, as described previously. After the MIC test, the MBC for each condition was also examined as described previously.

Results and Discussion

The results of the susceptibility measurements with the various species at various pH values. Despite the wide interval between each concentration level, it is still obvious that generally there was a declining MIC and MBC as the environmental acidity increased. Most microorganisms failed to survive at a pH below 3 or 4 in the controls (no honey present). The presence of honey made no difference to survival at these pH values. The decline in MIC and MBC can be seen both in manuka honey and artificial honey, although the MIC and MBC values of artificial honey are significantly higher than those of manuka honey. Like what has been shown above, this result shows that the osmolarity again is not practical.

It is possible that the antimicrobial activity provided by the unknown non-peroxide antimicrobial component itself is to some degree destroyed honey at low pH should have been as high as that with artificial honey, but it was not. This indicates that, before the pH decreases to the level at which bacteria can not survive at all, the antimicrobial activity of manuka honey remains stable in acidic condition.

In contrast to the acidic conditions, the alkalic environment (pH above 7) seems to be slightly destructive to the antimicrobial activity in manuka honey, as seen in the remarkedly higher MIC and MBC values at pH 8 with manuka honey for all species of bacteria. The alteration in the antimicrobial activity in acidic and alkalic condition may possibly be partly caused by the decrease in activity of glucose oxidase that is responsible for the generation of hydrogen peroxide, as the activity of the glucose oxidase of honey is maximized at around pH 6.1 and declines as the pH moves toward the acidic and alkaline sides. In the study, the activity of glucose oxidase declined from 100% at pH 6 to approximately 40–50% at pH 8–9 or 20–30% at pH 2–3. While acid may compensate the loss in the activity of the enzyme, pH 7–8 is in fact preferable to many microorganisms and therefore the loss in the activity of glucose oxidase in such moderate alkaline environment could have also reflected on the MIC and MBC results. This suggests that the acidity itself can influence the antibacterial efficacy of honey. On the other hand, it is possible that the alkaline environment also has a negative effect on the nonperoxide components in manuka honey. Snow and Manley-Harris found using an agar diffusion assays that exposure to a range of alkaline pH caused the non-peroxide activity of manuka honey to be lost in less than 30 min (within 10 min, the honey lost 50%, 100% and 100% activity at pH 9, 10, and 11 respectively). The finding in the work in this chapter is in agreement with the work by Adams in which treatment of manuka honey with NaOH and then titration back to the original pH resulted in irreversible loss of the non-peroxide antibacterial component in their HPLC assay. The lability of dicarbonyls, hence of methylglyoxal, in an alkaline environment has also been reported by Bowden and Fabian. The increase in the MIC and MBC at pH 8, however, shows only a partial loss of activity because the efficacy is still greater than that of the artificial honey in the identical condition, suggesting that at pH 8 manuka honey still can inhibit microorganisms by means of peroxide or non-peroxide action in addition to osmolarity.

Acidity, especially that in the stomach, in comparison with other defence mechanisms in the gastrointestinal tract (e.g. enzymes, mucus, immunoglobulin), is the primary barrier in a host to defend against incoming bacteria, and the failure in the growth of bacteria in the extreme pH as shown in this study may make humans feel that the importance of the antibacterial property of manuka honey is less obvious. One should not forget, however, that bacteria may break through the acid barrier because they may have adapted to the extreme pH before being ingested in the body. Many species of bacteria including gastrointestinally important pathogens such as E. coli, Shigella, V. cholerae, S. typhimurium and H. pylori have been reported to be able to tolerate extreme pH using many mechanisms, including altering membrane lipid composition or proteins, or through numerous pH-influenced gene regulations. Such acid or alkaline adaptation can be induced by exposing the cells in mildly acidic or alkaline conditions for one to two cell-doubling generations. As such mild environmental stress can easily be found in food or food processing, bacteria causing food-borne gastroenteritis in fact may have been able to tolerate acid/alkaline stress, and consequently the acidity in the stomach or intestinal tract may be less effective against these adapted bacteria. In this respect, the antibacterial activity of manuka honey may be of additional help to the gastrointestinal defense system.

Whereas bacteria adapted to acidic conditions would be tolerant to a yet more extreme environment, at the same time they can also be more sensitive to other types of environmental stress. E. faecalis that is pre-incubated at pH 10.8 can subsequently tolerate pH 11.9 but not pH 3.2. In the same study, the same species exposed to pH 4.8 was found to survive at pH 3.2 but not at pH 11.9. A similar phenomenon can also be observed with E. coli. As acidic/alkaline (and, in fact, many other environments) stress responses involves various mechanisms, one adaptation could lead the cells to be adversely more sensitive to other stress. Similarly, cells of the same species with and without adaptation to acidic/alkaline environment may differ in the sensitivity to manuka honey. As in this study the bacterial cultures were prepared in neutral Mueller-Hinton broth (pH 7.4) before the sensitivity test was carried out, it would not be surprising if the efficacy of manuka honey in acidic/alkaline environment in vivo happen to vary from this in vitro study, and of course it would also be of interest to investigate the sensitivity of the acid/alkaline adapted strains to manuka honey in the future.

The Effect of Gastrointestinal Enzymes on the Activity of Manuka Honey

Following study of the effect of pH on the total antimicrobial activity of manuka honey, the effect of gastrointestinal enzymes on the honey was investigated. It has been reported that glucose oxidase from Aspergillus niger and Penicillium notatum is stable against pepsin and an acidic environment, however the activity was lost completely at pH 1. However, it is not known whether or not the glucose oxidase of honey has the same stability as that of fungal glucose oxidase. If there were any difference in the result of the sensitivity test to be carried out in this section, that could be either because the bacterial cells are digested or damaged by the proteinases, or because the enzymes responsible for total and non-peroxide antibacterial activities in manuka honey are affected by the proteinases through an unknown mechanism.

To simulate the gastric and intestinal environments, simulated gastric fluid (SGF) and simulated intestinal fluid (SIF) were used in this section.

Whereas bile salt is not an enzyme, nor is it included in the simulated intestinal fluid formula in the United States Pharmacopeia, bile was also used in the study to evaluate whether or not bile would affect the overall antibacterial activity of the honey. This decision was based on two reasons. One was that bile is secreted into the intestine. The other was it has been reported that bile appears to have antibacterial activity against some microorganisms and therefore bile, together with the other proteinases, could assist the antibacterial action of the honey.

Results and Discussions

In the test associated with simulated gastric fluid (SGF) which includes pepsin and extremely high acidity, all organisms failed to survive even in the positive control without honey and pepsin. This is in fact predictable as SGF had a pH of 1.2 whereas most microorganisms are known to lose their viability at pH below 4 or 5 as shown in Section 7.2. To understand the effect of the gastric enzyme on the efficacy of the honey, it would be of benefit to use microorganisms such as H. pylori that are capable of tolerating an acidic condition. Unfortunately it was not possible to test H. pylori in this study due to the difficulty of culturing that species. Also to be noted is that H. pylori survives the acidic conditions in the stomach by secreting the enzyme urease to hydrolyse urea in the gastric mucosa and create a less acidic micro-environment which may be at a pH at which manuka honey is known to work. Because of the urease activity of this species, it would be somewhat misleading to denote this testing model as to determine the effect of acidic conditions on the antibacterial activity in manuka honey. Alternatively, possible solutions to investigate the effect of pepsin would be to reduce the time for which bacteria are exposed to the simulated gastric fluid. However, it would be difficult to predict how long the antibacterial component, especially non-peroxide component, would be staying in the stomach, because any fluid like honey solution would promptly pass through the gut whereas it is not known whether or not the non-peroxide components would persist in the gastric mucosa or epithelial tissue. Another challenge that can be foreseen is that, if an organism is to be treated with gastric fluid followed by being transferred to a solid medium, the sudden fluctuation in the environmental pH would corrupt the cells which would consequently influence the results.

Perhaps the best way to assess the effect of pepsin would be to investigate from the viewpoint of the mechanism of antimicrobial action of manuka honey. Hydrogen peroxide is unlikely to be relevant to pepsin because although glucose oxidase would not be destroyed by pepsin, the pH required for pepsin to react is too low for glucose oxidase to generate peroxide action. On the other hand, the mechanism of the non-peroxide activity in the honey is not thoroughly understood at this stage.

The MIC and MBC values of manuka honey in the various digestive environments with each species of bacteria. The MBC values are generally slightly higher than the MIC for all groups (i.e. 1–2 dilution steps different from the MIC). Kruskal-Wallis analysis showed that the enzymatic environments statistically affect both MIC and MBC values (P < 0.05). The adjusted pair-wise analyses showed that the results of environments 3 and 5 are statistically higher than that of environment 1 (P < 0.05). Overall, however, 10–15% manuka honey under intestinal digestive conditions was still capable of inhibiting the growth of most gastrointestinal bacteria. Comparing the MIC of manuka honey with and without pancreatin or bile being added (environments 1 and 2 cf. environments 3–5), it is seen that the addition of the intestinal components can negatively affect the efficacy of manuka honey.

Interestingly, although there was variation in the degree of fluctuation, most tested species revealed a similar trend in that the addition of either pancreatin or bile raised the MIC readings (environments 3 and 4). The increase in MIC is more significant when pancreatin and bile were added simultaneously (environment 5) (P < 0.05). More interestingly, for most tested species, the degree of increase in the MIC when both pancreatin and bile were present (i.e. the difference between the MIC of environments 5 and 2) is almost identical to the sum of the degree of the increase due to the individual components (i.e. the sum of “the difference between environments 3 and 2” and “that between 4 and 2”), suggesting that the influences of pancreatin and bile are probably additive.

Heated pancreatin (environments 6 and 7) did not dicrease the activity of manuka honey as great an effect as unheated pancreatin (P > 0.05), suggesting that denatured pancreatin can less likely affect the efficacy of the honey negatively. This also indicates that pancreatin can influence the antimicrobial activity of manuka honey enzymatically. Perhaps this is approached by destroying proteins that are responsible for the antibacterial activity of the honey. This hypothesis, however, will need to be proved in future by investigating the stability of proteins in honey against enzymes.

While the antimicrobial activity of manuka honey still works in the simulated intestinal environment, it is still somewhat surprising to note that intestinal enzymes slightly reduced the activity of the honey, and especially that bile seem to have an additive effect on the reduction. The reduction in the activity may partially be due to the environmental pH which is slightly more alkaline than the regular testing environment, but since pancreatin is a mixture of catalytic enzymes including protease (trypsin) produced by the exocrine cells of the pancreas, the proteolytic enzyme could also have some membrane damaging effects on the bacterial cells, which in theory should have resulted in a decrease in the MIC values. The role of pancreatin in pathogenesis is not well known and the research is limited. Paramithiotis assessed the survival of Staphylococcus spp. in the gastrointestinal tract and found the isolates to be resistant to pancreatin. On the other hand Chaignon tested the susceptibility of staphylococcal biofilms to gastrointestinal enzymatic treatment and found that the eradication depends on the constituents and the clinical isolates. As the biofilm form of growth protects organisms from the hosts’ immune systems and also from antimicrobial therapy, the disintegration of the biofilm matrix from the hosts’ epithelial cell surface into the environment may help antibiotics to eliminate the bacteria. It is thought that it is this detaching function, rather than causing damage to the bacterial cells, is the way pancreatin and probably pepsin too, helps the host to defeat the bacterial infection.

Unlike the limited reports on pancreatin, a number of studies regarding the interaction between bile and bacteria have been done and antimicrobial activity of bile has been reported.

The major function of bile in vivo is to act as a biological detergent which emulsifies fats in the intestine. This suggests that bile has the potential for antimicrobial activity by disrupting cell membranes which may be more effective to Gram-negative than to Gram-positive bacteria due to the difference in the structure of their cell membranes. Other antimicrobial actions of bile that have been suggested are: disturbing macromolecule stability, inducing secondary structure formation in RNA, inducing DNA damage, altering the conformation of proteins resulting in misfolding or denaturation, causing oxidative stress by generating oxygen free radicals and chelating calcium and iron.

The type of the bile used may affect the inhibitory effect of bile to bacterial cells. Grill reported that bovine bile that contains trihydroxyconjugated bile salts, was less inhibitory to bacterial cells than porcine bile, which contains dihydroxyconjugated bile salts. However, Begley suggested that although bovine bile is commonly chosen to assess the tolerance of bacteria to bile, porcine bile would be more appropriate to use as it has more similar salt/cholesterol, phospholipid/cholesterol and glycine/taurine ratios to human bile. It is unknown if the constituents of bile would interact with the major antimicrobial ingredient of manuka honey. It is expected that the overall antimicrobial activity of manuka honey with porcine or human bile would be higher than that obtained in this study as human bile is more likely to provide inhibition of the bacterial strains. However, this hypothesis needs to be explored with human bile in future.

Despite of the number of report showing the deleterious effect of bile on cells, many microorganisms have been found to be resistant to bile. It is believed that Gram-positive bacteria are generally more sensitive to bile than Gram-negative bacteria, and therefore bile salt is usually used as a selective agent in media such as MacConkey medium, Salmonella-Shigella medium and esculin bile agar. The MIC and MBC values of bile reported for Gram-negative bacteria are generally substantially higher than those for Gram-positive bacteria, but some Gram-positive bacteria seem to be resistant to bile too. Brook, Carpenter, Flores and Saito reported that Enterococcus spp. may be isolated from bile or biliary drain devices. Since bacteria may survive on biliary drain devices where the concentration of bile is very high, the fact that the species of bacteria in the study in this thesis are not killed by the addition of 0.3% bile would not be unexpected.

A possible explanation for the observation that the organisms were more resistant to the honey when bile (and perhaps pancreatin too) were present was the induction of adaptation. This is a phenomenon commonly seen among bacteria that a short-term pre-exposure of bacterial cells to sublethal levels of stress may induce the cells to adapt to the subsequent normal level of the stress, and as many stresses may have similar effects on the cells, an adaptation may also mean they adapt to other stresses at the same time. This is also termed “cross-adaptation”. Begley noted that low levels of bile may rapidly intercalate with membrane lipids after solubilizing them, which results in an increase in the resistance to further stresses. It is possible that this cross-protection caused manuka honey to be less effective in the simulated intestinal environment.

It must be emphasized that the bacterial growth, the effect of gastrointestinal enzymes and the antimicrobial action of manuka honey cannot be fully modeled from in vitro studies, because of the complex nature of the gastrointestinal environment that cannot be fully simulated in a laboratory. Microorganisms can alter their morphologies and genetic expressions in response to environmental stress, and the regulation can differ greatly even from strain to strain. It has also been shown in an E. coli macroarrays assay that manuka honey can up-regulate genes that involve in stress responses, and down-regulate those involved in protein synthesis. What has made the puzzle even more complex is that, unlike most antibiotics for which the chemical composition and the mechanisms of action are understood, there are still numerous of aspects in manuka honey that remain to be cleared. These difficulties also apply to other gastrointestinal environment factors described in previous chapters.

Conclusion

Despite acidity and intestinal enzymes being able to negatively affect the antibacterial activity of manuka honey, the total antimicrobial action of the honey under these conditions is still significantly higher than that of sugar syrup. Environmental stress such as gastrointestinal acidity and enzymes, perhaps as well as the atmospheres as discussed above, may actually not only affect manuka honey but also cause alteration in bacterial morphology. To thoroughly understand the characteristics and the mechanism of non-peroxide activity in manuka honey it is therefore necessary to further investigate the relationship between the honey, the microorganisms and their stress responses.

The Antibacterial Efficacy of Standardized Manuka Honey

In this thesis manuka honey with standardized antibacterial activity has been used to evaluate its potential to treat gastrointestinal bacterial infections. Honey has been utilised historically to treat infections worldwide, but unfortunately many of the reports estimating the antibacterial activity in honey are not comparable because those reports did not standardize the honey of interest using a reference antiseptic. Furthermore, in those reports some details that in fact may influence the result of the sensitivity test are absent.

Manuka honey is now widely known by the public to have exceptional antibacterial activity, but it is widely believed that all manuka honey has the same activity. An extensive survey done by Allen has shown clearly that not all manuka honey is the same, and also that some manuka honey samples in fact do not have any detectable antibacterial activity if the enzyme catalase is added to remove hydrogen peroxide. Allen found that the antibacterial activity of 50 manuka honey samples tested for both total activity and non-peroxide activity varied greatly in potency, from not detectable to equivalent to that of over 30% phenol solution. Indeed, at the beginning of this PhD project, a manuka honey with a very high level of activity (M113) was used because it was not realized that the activity was much higher than that stated on the label. After it was realized, a honey with a near-median level of activity was used (M115).

Some species of bacteria are inhibited by quite low levels of osmolarity, so inhibition by honey that is observed may be due to the sugar content rather than to hydrogen peroxide or non-peroxide factors. This is why it is important to include an artificial honey as a reference, so that the antibacterial efficacy of factors other than the osmolarity can be distinguished.

Another variable that is commonly missed out in studies that are in the literature is the cell density of the bacteria being tested in the susceptibility assay. It is generally observed that the higher the cell number, the more resistant the cells are to antimicrobial agents. Depending on the species of bacteria or the antibacterial agents being tested, the MIC could rise 4 to 16-fold with as little as 0.5 log 10 increase in inoculum density. Wiegand and Burak reported that the MIC of eight tested antibiotics against P. shigelloides dramatically increased from B0.03 mg/l with 105 CFU/ml up to 16 mg/l with 106 CFU/ml. It has also been observed that the influence of inoculum size substantially increased if the inoculum exceeded 9×107 CFU/ml, if the antibacterial agents were tested under anaerobic conditions, or if the antibacterial agents were not rapidly bactericidal. The influence of inoculum density on sensitivity studies is understandable because an increase in the inoculum would reduce the effective concentration or the per-cell concentration of antibacterial agents. Therefore, without the information on the cell density being given, it would be difficult to determine if the reported sensitivity of the microbes to the antibacterial agents being tested is actually over-estimated or under-estimated. In this thesis both the antibacterial potency and the cell density are standardized, and this makes the findings from this thesis of greater value than those from other similar work where this was not done.

Although the materials used in the work in this thesis have been controlled, it must not be forgotten that one should not expect the results obtained in this in vitro project to be reflecting the actual efficacy of manuka honey when it is used to treat bacterial gastroenteritis in vivo, because the gastrointestinal tract is an extremely complex micro-ecosystem, and it is impossible in a laboratory to mimic all factors that are involved in the tract.

In this PhD project, the sensitivity of a number of gastrointestinal microbes to manuka honey was first determined in the simplest condition (i.e. an in vitro environment that is commonly used in an antimicrobial agent sensitivity test), followed by considering the anaerobic conditions that could result in there being no hydrogen peroxide produced in manuka honey. The possible recovery in the growth of bacterial cells simulating if honey is removed from the bacteria by absorption from the intestine or by flow past immobilized bacteria in the gut was studied, then lastly the possible influence of the gastrointestinal pH as well as enzymes on the potency of manuka honey was studied. Note that in these chapters, each factor was considered separately. For instance, the effects of acidity and enzymes on the MIC and MBC values was studied, the possible influence caused by the anaerobic conditions is not taken into count. Similarly, in the pharmacodynamic assays neither the presence / absence of oxygen nor the gastrointestinal acidity / enzymes were considered. In vivo, these factors certainly would be involved simultaneously, but that would make the investigation much more complex.

Despite the fact that the work in this thesis is over-simplified from the viewpoint of the real gastrointestinal environment, the results of this thesis do demonstrate that the antibacterial properties of manuka honey may aid in the treatment of bacterial infection in the gut. Additionally, the unexpected tolerance of probiotics to the antibacterial activity of manuka honey may also help improve the gastroenteritis symptoms, as honey could be used therapeutically without disturbing their growth. This may be helped by the prebiotic action of honey on probiotics. However, the major concern to evaluate whether or not honey is likely to help clear infection in the gastrointestinal environment would be whether a sufficient concentration of honey could be achieved in the tract, because the concentration of manuka honey would be diluted in the gut by secretions and by water in food or drink. ORS (oral rehydration solution) has been widely used to treat gastroenteritis, and perhaps one may replace the carbohydrate component in ORS with manuka honey so that the rehydration solution may provide an antibacterial function in addition to merely rehydrating the patient. To achieve this, the amount of the honey in the rehydration solution would need to be re-evaluated because the concentration of the carbohydrate in ORS is 75 mmol/l (World Health Organisation), or 0.87% v/v glucose. Considering that honey contains approximately 40% glucose, the concentration of glucose in ORS would be equivalent to that of 2.175% honey solution. This concentration of honey is clearly too low to have a practical antibacterial effect on bacteria in the gut, and therefore the recipe of the honey rehydration solution needs to be modified rather than just replacing the glucose in ORS with the equivalent strength of manuka honey. In the clinical trial done by Haffejee and Moosa, an oral rehydration soltuion that contained 5% (v/v) of pure honey (floral source not given) was used and achieved suscessful results and faster clearance of bacterial infection. Another possible solution that may be of some help to ensure the achievable antibacterial concentration of manuka honey in the gut, and also the time the concentration of honey remains above the MIC/MBC, would be to ingest the honey when the tract is empty so that the active component in the honey would not be largely diluted in vivo. It is also possible that there would be an increase in concentration of antibacterial activity when water is absorbed in the gut. However it is not known whether or not the antibacterial factors in manuka honey would also be absorbed in the gut, and further research is required to elucidate this.

The effect of acidic conditions and that of pepsin on the antibacterial activity of manuka honey was unfortunately not determined successfully because the acidity inhibited the bacteria. In theory, the antibacterial activity of the honey due to hydrogen peroxide would be destroyed in harsh acid conditions. On the other hand, it is not known if the non-peroxide antibacterial component in manuka honey (now identified as a high concentration of methylglyoxal; see later) is likely to be destroyed in the gastric environment. Methylglyoxal has been reported to react with arginine and lysine residues and is more reactive (i.e. less stable) in alkaline than in acidic conditions. However, methylglyoxal accounts for only half of the non-peroxide activity because an unidentified synergist is now known to be involved in the action of nonperoxide antibacterial activity of the honey. Therefore the stability of this activity in the presence of pepsin and acid still needs to be experimentally proved. If the gastric factor is of much concern, perhaps this issue can again be minimized by ingesting manuka honey when the stomach is empty because in the absence of chyme containing large quantities of undigested proteins there would be less stimulation of secretion of gastrin, and therefore acids and enzymes.

Although measurements of MIC and MBC are commonly used to evaluate the efficacy of antibacterial agents, these parameters would not be meaningful in the case of evaluating the treatment of enteropathogens using honey. It is presumable that the fluidity of honey and the possibility of the absorption of the antibacterial components in the honey would result in a very short contact time of honey with bacteria in the gastrointestinal tract. The time-to-kill and the post-antibiotic assays have shown that the MIC/MBC parameters are not sufficient to evaluate the efficacy of honey in vivo because the conditions for the MIC/MBC test do not resemble that in the tract in terms of the contact time. Also, although the results from the postantibiotic effect assay conducted in this thesis have shown that the antibacterial effect of manuka honey persists for a few hours after one hour of partial inhibition, it is not known whether or not this effect would still exist at lower honey concentration than that has been used in this thesis (20%) and/or with shorter contact time of the honey with bacteria.

Note also that honey has also been reported to detach bacterial cells from the host cells. Attachment to the gastrointestinal mucosa usually is the first step of infection by enteropathogens, and therefore even if the antibacterial activity of honey is not strong enough to inhibit bacteria, it may still be useful to treat the pathogens if the honey is effective in detaching bacteria. Like the postantibiotic effect assays, it would be of interest to test the effect of different concentrations of the honey as well as that of different contact times on the detaching effect in the future. Future work should also examine a range of honeys to find the best type for detaching bacteria. Breton and Pineau investigated Robinia pseudoacacia honey, lime-tree (Tilia spp.) honey, lavender honey, chestnut honey, honeydew from conifers, citrus honey, alpine flowers honey and “Gatinais” honey. Of these, R. pseudoacacia honey was found the best. This suggests that the ability to detach varies between honeys.

Furthermore, honey has been shown to have anti-inflammatory and antioxidant activities. Since inflammation and reactive oxygen species (ROS) play a role in gastroenteritis, the possible usefulness of the anti-inflammatory and anti-oxidative activities of manuka honey at different conditions may also warrant further studies.

Methylglyoxal in Manuka Honey

The component(s) responsible for the non-peroxide antibacterial activity in manuka honey have remained unidentified for decades since the discovery of this unusual activity. This has been because honey is complex in its composition. It was not until much of the work in this thesis had been completed that methylglyoxal (MGO) was identified as the substance responsible for the non-peroxide antibacterial activity in manuka honey. However, an asyet unidentified synergist is also present which doubles the antibacterial potency of MGO.

Methylglyoxal, also called 2-oxo-propanal, is formed in organisms as a side-product of several metabolic pathways, in which glycolysis is the most important source. Methylglyoxal is highly cytotoxic, although generally cells may detoxify it through several mechanisms such as the glyoxalase systems.

The effects of methylglyoxal on living organisms is controversial and remains to be clarified. While there have been a number of reports regarding the toxicity of methylglyoxal to living organisms, some other research has shown that methylglyoxal may not be toxic when used with ascorbic acid and creatine, and that it may be safely used as an anticancer drug. The high concentration of methylglyoxal in manuka honey should not be of concern from a practical point of view. According to Adams and Mavric, the amount of methylglyoxal in manuka honey varies from 38 to 828 mg/kg. Most reports concerning the cytotoxicity demonstrated the results by treating plants, mammals, yeasts (Saccharomyces cerevisiae) and prokaryotes, either in vivo or in vitro, with approximately 10 mmol/l (720 mg/l) or higher concentration of methylglyoxal long-term. If we apply this to human ingestion of manuka honey and also consider that the human body contains about 50 litres of water, this would mean one needs to take a dose of approximately 50 kg of manuka honey with a high level of methylglyoxal to reach that cytotoxic amount of methylglyoxal from the honey. Even then, there is a possibility that not all of the methylglyoxal will be absorbed from the gut, and that some of what does get absorbed will be broken down by metabolism. A number of trials have been conducted in mammals, but the exceptionally high doses the animals were subjected to in those trials may render questionable the interpretations. For instance, Kalapos reported that 800 mg/kg body weight of methylglyoxal was lethal in mice within 4 hours after intraperitoneal injection. It is questionable if this is of meaning when applied to humans consuming honey, as it equates to consuming about 50 kg of honey in a single dose, even if all of the methylglyoxal is absorbed from the gut. Similarly, Furihata administered methylglyoxal ranging 300–600 mg/kg body weight by gastric tube to demonstrate an unscheduled DNA synthesis in male F344 rats. Interestingly, despite the high level of methylglyoxal being reported by Adams and Mavric, toxicity or related disease caused by ingesting manuka honey has never been reported. Also to be noted is that in published reports on animal trials, the animals were observed over days and but it is not clear how rapidly nor how much the methylglyoxal in food would be absorbed in the gut. The interacting effect of other ingredient (e.g. antioxidants) in food/honey on the absorption of methylglyoxal and the action of the synergist in manuka honey also remains to be elucidated.

Conclusion

Manuka honey has been shown in published literature and in the work in this thesis to be effective against a wide range of gastrointestinal pathogens, although it is too early to conclude from the demonstrated data in this thesis that the antibacterial activity of manuka honey would still be effective in the intestine. In the work in this thesis it has been found that manuka honey with a near median level of antibacterial activity tested against most of the common enteropathogens inhibits at < 10%, kills at < 16.6%, and stunts the growth for 2–3 hours at 20% if used in vitro. It was also found that the honey loses no more than 50% of its antibacterial efficacy when tested in vitro under most conditions simulating those in the intestine (anaerobiosis and in the presence of intestinal components). Because acidity itself inhibits bacteria, it is not known whether or not the honey would work in the stomach. However, published findings have shown that the activity of honey is not harmed by exposure to low pH then neutralisation, therefore it should still be effective once it passes through the stomach to the intestines.

Gastroenteritis has generally been treated with oral rehydration therapy (ORT) that consists of carbohydrates and electrolytes. In addition to its antibacterial action, manuka honey could also be used to provide carbohydrates in an electrolyte solution to rehydrate patients with gastroenteritis. Furthermore, because honey also provides additional bioactivities that the conventional ORS does not have (such as antioxidant, anti-inflammatory and prebiotic activity), it is likely that a rehydration solution containing manuka honey and electrolytes would be superior to ORS. With an appropriate recipe of honey rehydration solution and further investigation to find the most appropriate dosage and frequency of doses, a clinical trial may be warranted.