In vitro and in vivo anti-Helicobacter/Campylobacter activity of the aqueous extract of Enantia chlorantha

Abstract

The aqueous extract of Enantia chlorantha Oliver (Annonaceae) stem bark, a plant widely used in Cameroon for the traditional treatment of gastritis and stomach problems, was assessed for in vitro and in vivo anti-Helicobacter/Campylobacter properties using the well diffusion assay, agar dilution assay, and killing rate determination. The in vitro activity was dose-dependent, and the same antimicrobial parameters (MAQ = 0.63 mg; MIC = 0.39 mg/mL; MBC = 1.56 mg/mL; ET₁₀₀ = 8 h) were obtained for both H. pylori and C. jejuni/coli. When the plasma active principle concentration equivalence was determined in vitro using plasma from rats exposed to a single dose (3000 mg/kg) of the extract, the peak absorption of E. chlorantha active principle against H. pylori occurred at 2 h. Plasma activity was nil 8 h after extract administration. The in vivo H. pylori eradication potency of the extract was assessed using mice infected with H. pylori. Antral mucus sample cultures from mice treated with E. chlorantha extract (500 and 1000 mg/kg for 3 days) did not yield any growth. The results suggest that in addition to its in vitro activity, E. chlorantha water extract also possesses in vivo antibiotic effects against H. pylori.

Keywords:

• Enantia chlorantha
• Helicobacter pylori
• in vivo
• in vitro activity

Introduction

The eradication of Helicobacter pylori (H. pylori) from patients with chronic gastritis, peptic ulcer disease, and gastric cancer as well as mucosa-associated lymphoid tissue (MALT) lymphoma has been widely recommended as key to the effective treatment of these disease conditions. This recommendation is based on the major role of H. pylori in the etiology of these diseases (Soll, 1996; Malfertheiner & EHPSG, 1997), since H. pylori eradication results in their successful management (Skirrow, 1992; Petschow et al., 1996; Schillio, 1996; Korwin, 1999; Sizemore et al., 2002; Gisbert & Pajares, 2003). H. pylori strains resistant to amoxicillin, clarithromycin, and metronidazole (the mainstay therapy) quickly emerged, and treatment failure following treatment using these drugs has become frequent (Megraud, 1998; Okamoto et al., 2002). In an attempt to overcome treatment failures due to resistance, the Maastricht consensus (1996, 2000) recommended several 7-day tri-therapy regimens which are usually composed of a proton pump inhibitor associated with two antibiotics (nitro-imidazole combined with either amoxicillin or clarithromycin) (Malfertheiner et al., 2002). Other multidrug anti-H. pylori regimens have been proposed, but none has yielded complete satisfaction, and microbial resistance as well as drug side effects remain the main causes of therapeutic inefficiency (Kato et al., 2000; Beales, 2001; Courillon-Mallet, 2003). For these reasons a new area of H. pylori research consists in the assessment of the in vitro and in vivo susceptibility to compounds and extracts of plant origin. For an anti-H. pylori agent to be efficient, it is necessary for it to remain active after trans-intestinal passage. It is known that clearance of H. pylori from the stomachs of infected patients can be due to the direct topical activity of the ingested drugs at the level of the gastric mucosal epithelium. In addition, a secondary systemic therapeutic activity can result from the back secretion and re-entry of the absorbed active principle from the basal to the apical side of the gastric epithelium (Lamouliatte et al., 1991; Adamek et al., 1993; Goddard et al., 1996; Matysiak-Budnik, 2002). In this process, the rate of absorption, bioavailability, and half life of the active principle are important factors.

Enantia chlorantha Oliver (Annonaceae), also known as the African yellow wood, is a dense forest tree found in Cameroon, Nigeria, and Gabon. In the southern forest zone of Cameroon, it is used for the traditional treatment of stomach problems, jaundice, urinary tract infections, malaria, tuberculosis, hepatitis, and some forms of ulcer (Adjanohoun et al., 1996). A protoberberine-type alkaloid identified as 7,8-dihydro-8-hydroxypalmatine (PAL) prepared from the stem bark of this plant was previously shown to possess gastric cytoprotective action against lesions induced using various irritants as well as healing effects on chronic acetic acid-induced gastric ulcers (Tan et al., 2000, 2002). In this study, the in vitro anti-Helicobacter/Campylobacter properties as well as the possible in vivo H. pylori eradication activity of the aqueous stem bark extract of E. chlorantha was assessed. The in vitro anti-Helicobacter activity of plasma obtained from animals exposed to a single dose of the plant extract was also evaluated. This bioassay was used to determine the plasma concentration equivalence of the active principle.

Materials and methods

Plant material

The stem bark of E. chlorantha was harvested in April 2005 at Ambam (South Province of Cameroon), and identified by Barthelemie Tiengue by comparison with the existing Cameroon National Herbarium voucher specimen N° 25918/SRFCAM. The bark was sun dried untill constant weight, and ground to a powder. A 10% (w/v) mixture of the bark powder and distilled water was boiled by simmering for 15–20 min. After cooling to room temperature, the preparation was filtered through four layers of cotton fabric gauze. The filtrate was allowed to stand for 90–120 min, after which the supernatant was filtered through Whatman filter paper N° 1. The decoction obtained was evaporated at 40°C till total dryness using a convection air oven (Jencons-PLS, Bedfordshire, England), and a dry solid material was obtained with a yield of 4.6%, w/w. The extract was used immediately or stored at 4°C.

Culture media, reference antibiotics, and culture media supplements

Brain heart infusion (CM0225), Columbia agar (CM0331), horse serum (SR0035), lacked horse blood (SR0048), Vitox supplement (SR0090), Helicobacter pylori DENT supplement (SR0147), and CampyGen gas packs (CN0024A) were purchased from Oxoid, Basingstoke, England. The reference antibiotics [amoxicillin (Amoxyl 500 mg/5 mL; GlaxoSmithKline Australia Pty Ltd.)] and metronidazole [(Metronidazole Injection, USP RTU® 500 mg/100 mL; Baxter Healthcare Corp., Deerfield, IL)] were purchased from a local pharmacy. Sterile normal saline was used in the preparation of dilute solutions for the in vitro and in vivo tests. For the in vitro well diffusion test, the following concentrations of metronidazole were used: 25, 50, 100, 150, and 200 μg/mL.

In vitro analyses

Test organisms and culture conditions

The locally isolated human fecal Campylobacter jejuni/coli sample (CPC-022004) was provided by the Centre Pasteur du Cameroun. Helicobacter pylori CCUG 39500 was obtained in lyophilized form from the Culture Collection University of Göteborg (CCUG), Sweden. The strain was revived following CCUG recommendations and its identity was confirmed using the rapid urease and the catalase/oxidase tests. The strains were maintained at 37°C on Columbia agar supplemented with 5% (v/v) lacked horse blood and 1% (v/v) Vitox (CA-Vitox), under microaerophilic conditions generated using CampyGen in an airtight 2.5 L anaerobic jar (Oxoid). The test organisms were then sub-cultured into fresh medium every 72–96 h. The in vitro antimicrobial diffusion test was performed on 5–6 mm thick CA-Vitox. The test organism inocula used for the diffusion and dilution tests were prepared by suspending 72 h colonies in 2 mL of sterile distilled water to make a turbidity standard of McFairland N° 2 and N° 1, respectively, for H. pylori and C. jejuni/coli (NCCLS, 1999; Cheesbrough, 2000). These turbidity standards containing about 10⁹ and 10⁸ CFU/mL, respectively, for H. pylori and C. jejuni/coli produced semi-confluent or confluent growth.

Antimicrobial well diffusion test

This test was performed as previously described (Tan et al., 2006). In brief, two-fold serially decreasing concentrations of the extract (200–3.125 mg/mL) were prepared using sterile distilled water. Each solution (50 μL) was dropped into a 6 mm diameter well drilled on a CA-Vitox plate previously inoculated with 0.1 mL of the standard inoculum. Metronidazole (0, 1.25, 2.5, 5, 7.5, and 10 μg) and sterile distilled water were used as the reference antibiotic and the negative control, respectively. The plates were then incubated under microaerophilic conditions. After 72 h, extract concentrations that showed a diameter of inhibition (DI) of at least 7 mm were considered to be active. Each test was done in triplicate and the mean DI was calculated. The minimum active quantity (MAQ) was determined as the lowest quantity of extract or reference antibiotic that produced the lowest DI.

Antimicrobial agar dilution test

CA-Vitox plates (3.5 cm diameter) incorporated with two-fold decreasing concentrations of extract (60–0.23 mg/mL) or metronidazole (25–0.19  μg/mL) as well as extract-free plates (controls) were prepared in duplicate. For each concentration of extract or reference antibiotic, one of the plates served as the negative control while the other was used for the test. The test plates were inoculated with 0.1 mL of the standard inoculum of test organisms and all the plates were incubated as described above for 72 h. After incubation, the lowest concentration of extract that prevented visible growth was considered as the minimum inhibitory concentration (MIC) (Mahady et al., 2003). The surfaces of all the plates that showed no visible growth were each washed with 100 μl of sterile distilled water and the resulting solutions were plated on extract-free plates of CA-Vitox and then reincubated appropriately. After incubation the lowest concentration corresponding to the plate that yielded no growth was considered as the minimum bactericidal concentration (MBC).

Killing rate determination

Three milliliters of brain–heart infusion (BHI)–serum were incorporated with the test extract in capped tubes to obtain a 25 mg/mL solution, and each tube was inoculated with 0.3 mL of the standard bacterial cell suspension. The inoculated broth was incubated under microaerophilic conditions at 37°C, and 0.1 mL portions were removed at different time points (0, 2, 4, 6, 8, 10, 12, and 24 h) for viability counts. The portions removed were each serially diluted 100-fold in BHI, and the diluted solutions were each plated (0.1 mL) in triplicate on CA-Vitox agar plates. The resulting bacterial colonies were counted after 72 h of incubation to determine the mean viable count (Hassan et al., 1999). From the mean viable count (CFU/mL), the killing rate curve was plotted as the log₁₀ CFU/mL as a function of the length of time of contact of the microbes with the extract. The ET₁₀₀ (exposure time within which the viable count drops to the lowest detectable limit) was determined graphically.

In vivo analyses

Animals

Healthy inbred albino Wistar rats weighing between 130 and 210 g were used for the in vivo assessment of active principle absorption and determination of the plasma concentration equivalence. The animals were maintained on a 12 h light/dark cycle. They were given fresh water ad libitum, and fed with laboratory baked chow composed of maize (50%), soybeans (25%), and wheat flour (25%) supplemented with table salt, palm oil, and fish and bone powder. For the in vivo anti-Helicobacter study, 6–8-week-old inbred healthy Swiss mice weighing between 17 and 23 g were used. The housing conditions were as previously described (Boda et al., 2006). Prior authorization for the use of laboratory animals was obtained from the Cameroon National Ethics Committee (Reg. No. FWA-IRB00001954).

Determination of the plasma inhibitory concentration equivalence

A 240 mg/mL stock solution was prepared in distilled water using 7.2 g of extract. The animals were divided into two groups of 18 rats each and a single dose of extract (3000 mg/kg b.w.) or metronidazole (25 mg/kg b.w.) was intragastrically administered using a gavage needle mounted on a 5 mL syringe. At different time points (0.5, 1, 2, 4, 6, 8, 10, 12, and 24 h), two rats from each of the groups were ether-anesthetized and blood was collected by cardiac puncture using a 2 mL syringe. Two rats that served as controls were administered 2 mL of distilled water and were sacrificed at 24 h. The plasma obtained from the harvested blood was used immediately or stored frozen. Two-fold serial dilutions of the plasma (1:1–1:64) were aseptically prepared using sterile normal saline. The antimicrobial activity of each diluted solution was assessed by well diffusion (dilutions 1:1–1:4) and agar dilution (dilutions 1:2–1:64) methods as described above. Each diffusion test was performed in triplicate and the mean DI was calculated. The highest dilution which yielded a visible zone of inhibition (well diffusion diameter ≥7 mm) or visible growth in the Petri dishes (agar dilution test) was considered as the maximum inhibitory dilution (MID). The determination of the active principle concentration equivalence in the plasma was done using two methods: (1) the plot of the mean diameter of inhibition as a function of the quantity of the antimicrobial was used to establish the linear regression equation; the active principle concentration equivalence was then estimated from the linear regression equation as the concentration corresponding to the diameter of inhibition obtained with the MID; (2) assuming that the quantity of the antimicrobial at the highest plasma dilution was equal to the MIC, the active principle concentration equivalence in the plasma was estimated by multiplying the MID by the MIC (Kotecka & Rieckmann, 1993).

In vivo assessment of anti-H. pylori activity in infected mice

This assay was done as previously described (Boda et al., 2006). Briefly, the mice were pre-treated orally for 7 consecutive days with amoxicillin and metronidazole (25 mg/kg b.w.) to make sure that they were free from any Helicobacter-like organisms that could have been acquired through natural infection. A resting period of 7 days following the pre-treatment was allowed, at the end of which the animals were infected using 200 μL of a bacterial culture containing 10⁸ CFU/mL of H. pylori CCUG-39500. The inoculation was done four times a week with a 1-day interval between inoculations. On the second day following the last inoculation, the infected animals were grouped according to the following doses: 100, 500, and 1000 mg/kg b.w. or metronidazole 25 mg/kg b.w., and were treated for three consecutive days (one administration per day). On the day following the last administration of the extracts the animals were fasted for 10–12 h, after which they were sacrificed. The stomachs were excised and gastric biopsy samples were taken and used to assess H. pylori clearance by Gram staining, rapid urease test, and culture (Fauchere, 1999; Boda et al., 2006).

Results

In vitro analyses

In order to assess the antibacterial property of E. chlorantha extract on the two test organisms and to verify dose-dependent effects, the antimicrobial well diffusion test was first carried out. The results obtained show that the anti-Helicobacter/Campylobacter activity of the aqueous extract of E. chlorantha is dose-dependent, since inhibition diameters increased with increasing quantity of extract in the wells (Figures 1–3). The MAQ value on both pathogens was found to be 0.63 mg (Table 1). The activity of the reference antibiotic (metronidazole) when tested on H. pylori was also dose-dependent, with a MAQ value of 2.5 μg (Table 1).

Figure 1.  Anti-H. pylori activity of E. chlorantha aqueous extract: diameter of inhibition (DI) as a function of the quantity of extract in the well (mg). y, linear regression equation; R², regression coefficient. Values in brackets represent the well and the corresponding DI.

Figure 2.  Anti-H. pylori activity of metronidazole: diameter of inhibition (DI) as a function of the quantity of antibiotic (Qty) in the well (μg). y, linear regression equation; R², regression coefficient. Values in brackets represent the quantity of antibiotic in the well and the corresponding DI.

Figure 3.  Anti-C. jejuni/coli activity of E. chlorantha extract: diameter of inhibition (DI) as a function of the quantity of extract (Qty) in the well (mg). Values in brackets represent the quantity of extract and the corresponding DI.

Table 1.  In vitro antimicrobial activities of E. chlorantha and metronidazole on H. pylori and C. jejuni/coli.

	MAQ	MIC	MBC
Activity on H. pylori
E. chlorantha	0.63 mg	0.39 mg/mL	1.56 mg/mL
Metronidazole	2.50 μg	1.56 μg/mL	3.12 μg/mL
Activity on C. jejuni/coli
E. chlorantha	0.63 mg	0.39 mg/mL	1.56 mg/mL
Metronidazole	/	3.12 μg/mL	3.12 μg/mL

MIC, minimum inhibitory concentration; MBC, minimum bactericidal concentration; MAQ, minimum active quantity; /, not determined.

Agar dilution test

The growth of the test organisms was assessed on CA-Vitox plates incorporated with decreasing concentrations of the extract or the antibiotic in order to determine the MIC and MBC values. The results obtained are also presented in Table 1. The MIC and MBC values were 0.39 and 1.56 mg/mL, respectively, with E. chlorantha extract on both pathogens. Corresponding values for metronidazole were 1.56 and 3.12 μg/mL on H. pylori and were 3.12 and 3.12 μg/mL on C. jejuni/coli. The time-kill kinetic studies were done at extract concentrations equal to 16 times the MBC value. Since the ET₁₀₀ was considered as an index of bactericidal activity, this parameter was determined from the time-kill curves (Figure 4). It was observed that the original bacterial load dropped to the lowest detectable limit within 8 h for both pathogens (ET₁₀₀ = 8 h).

Figure 4.  Time-kill curves of E. chlorantha extract on H. pylori CCUG 39500 (16 × MBC) and C. jejuni/coli CPC-022004 (16 × MBC) in batch cultures at pH 7.3 and temperature 37°C.

In vivo analyses

Active concentration equivalence

Test animals were exposed to a single oral dose of the extract or metronidazole, and the anti-Helicobacter activity of their plasma samples was assessed at different time points using the diffusion and dilution methods. The linear regression equation whose best line of fit was used to estimate the active principle concentration equivalence was y = 2.9439x + 5.6413; with R² = 0.8843 (Figure 1). It was found that the plasma concentration of metronidazole at any of the time-points after the single oral dose (25 mg/kg b.w.) was too low to produce an antimicrobial effect detectable by the diffusion test. However, with the dilution test, antimicrobial activity of metronidazole was obtained at 30 min (6.24 μg/mL), 1 h (3.12 μg/mL), and 2–6 h (1.56 μg/mL), and, above 6 h, the plasma activity dropped to zero level (Figure 5). Using the diffusion test, the activity of E. chlorantha in plasma peaked (15.72 mg/mL) at 2 h, and no activity was detected above 6 h (Figure 6). Following the dilution test, an active concentration equivalence of 3.12 mg/mL was obtained 30 min after administration of 3000 mg/kg b.w. of extract (Figure 5). The plasma concentration of E. chlorantha active principle equivalence rose to a peak of 12.48 mg/mL during 2–4 h and then dropped to undetectable levels 8 h after administration (Figure 5).

Figure 5.  Active principle concentration equivalence of E. chlorantha extract and metronidazole in plasma as a function of time after a single oral dose (concentration determined as the product of the MID and MIC). Metronidazole, 25 mg/kg b.w.; active principle equivalence of E. chlorantha extract at 3000 mg/kg b.w. Values in brackets represent the active concentration equivalence in plasma.

Figure 6.  Active principle concentration equivalence of E. chlorantha extract in plasma as a function of time after a single oral dose of 3000 mg/kg b.w. (concentration determined using the linear regression equation of the diffusion test). Values in brackets represent the active concentration equivalence in plasma.

In vivo anti-H. pylori activity in infected mice

In order to assess the possible in vivo anti-Helicobacter activity of E. chlorantha extract, H. pylori-infected mice were treated with 100, 500, and 1000 mg/kg b.w. of E. chlorantha extract and compared with two control groups treated, respectively, with 25 mg/kg b.w. of metronidazole and normal saline. The results obtained (Table 2) indicate that the controls remained infected after 3 days of treatment with normal saline. In contrast, the cultures of the gastric tissue samples from the infected mice treated with E. chlorantha extract (500 and 1000 mg/kg b.w.) as well as metronidazole showed a negative rapid urease reaction and did not yield any growth after 3 days of treatment. Treatment with 100 mg/kg b.w. of extract produced 50% negative cases with gram staining, 62.5% with the rapid urease reaction, and 87.5% with culture of the gastric biopsy samples. In the group treated with metronidazole, the Gram stain was negative in 75% of the cases.

Table 2.  H. pylori clearance from mouse stomach following 3 days of treatment with E. chlorantha extract and metronidazole.

Animal no.	ECSB									Metronidazole, 25 mg/kg b.w.			Control, normal saline
	100 mg/kg b.w.			500 mg/kg b.w.			1000 mg/kg b.w.
	G	U	C	G	U	C	G	U	C	G	U	C	G	U	C
1	−	−	−	−	−	−	−	−	−	−	−	−	+	+	+
2	+	+	+	−	−	−	−	−	−	−	−	−	+	+	+
3	+	+	+	−	−	−	−	−	−	−	−	−	+	+	+
4	−	−	−	−	−	−	−	−	−	−	−	−	+	+	+
5	+	+	−	−	−	−	−	−	−	−	−	−	+	+	+
6	+	−	−	−	−	−	−	−	−	+	−	−	+	+	+
7	−	−	−	−	−	−	−	−	−	−	−	−	+	+	+
8	−	−	−	+	−	−	−	−	−	+	−	−	+	+	+
% Negative	50	62.5	87.5	87.5	100	100	100	100	100	75	100	100	0	0	0

ECSB, E. chlorantha stem bark extract; G, Gram stain; U, rapid urease test; C, culture; +, positive test or reaction; −, negative test or reaction. Animals: nos. 1–4 male; nos. 5–8 female.

Discussion

In the southern forest area of Cameroon, the decoction of E. chlorantha stem bark is frequently recommended for the traditional treatment of malaria and stomach problems (Christopher Mezui, personal communication). Previous studies showed that a protoberberine-type alkaloid (7,8-dihydro-8-hydroxypalmatine (PAL)) (Wafo et al., 1999) prepared from the stem bark of E. chlorantha possessed prophylactic antiulcer effects and healing effects on chronic acetic acid-induced gastric ulcers (Tan et al., 2000). The same compound was recently found to possess in vitro and in vivo anti-Helicobacter properties (Boda et al., 2006; Tan et al., 2006). In the present study, the crude water extract of E. chlorantha has also shown anti-Helicobacter/Campylobacter properties. This finding lends more credence to the use of the decoction from this plant for the traditional treatment of stomach complaints symptomatic of peptic ulcer disease. Unlike in the previous study where the ET₁₀₀ of the time-kill kinetics for PAL was found to be 6 h (at 3 × MIC), an ET₁₀₀ value of 8 h (at 64 × MIC) was obtained for the crude water extract of E. chlorantha in this study. The apparently slower bactericidal activity of E. chlorantha aqueous extract as compared to that of PAL (Tan et al., 2006) may result from a low presence of the active principle in the water fraction of the bark extract. It is, however, not excluded that the observed anti-Helicobacter/Campylobacter activity may be due to the presence of different antimicrobial principles.

H. pylori and C. jejuni/coli are classically known as Campylobacter-like organisms (Skirrow, 1992; Schillio, 1996; Dunn et al., 1997). They belong to the same family, the Campylobacteriaceae (Vandamme & De Ley, 1991), and are known to express the same sensitivity profiles and the same response mechanisms to antibiotics. The results of the in vitro analyses obtained in the present study confirm this notion, since similar MAQ, MIC, MBC, and ET₁₀₀ values were obtained for both pathogens (Table 1, Figure 4).

Since the in vitro anti-Helicobacter activity is not a reliable predictor of in vivo efficacy, the in vivo anti-Helicobacter property of E. chlorantha extract was also assessed using the previously described rapid screening mouse model (Boda et al., 2006). At 500 mg/kg b.w., E. chlorantha extract completely cleared H. pylori from the stomach of infected mice, suggesting the possible in vivo efficacy if used for eradication.

In order to determine the plasma concentration of chloroquine and to study the pharmacokinetic profile of this antimalarial drug, Kotecka and Rieckmann (1993) used the dilution technique to assess the in vitro antimicrobial activity of the plasma obtained from individuals treated with chloroquine. After comparison with high performance liquid chromatography values, they recommended the bioassay technique as a useful tool for the determination of drug concentration equivalence in the blood. In this study, the antimicrobial activity of the plasma from animals exposed to a single dose (3000 mg/kg b.w.) of the plant extract was used to evaluate the quantity of antimicrobial in the blood. The dose was chosen based on pilot experiments that were carried out in order to obtain a measurable in vitro reaction using the collected plasma. This dose of the extract was not toxic during acute toxicity tests (Tan et al., 2007). Antimicrobial activity of the plasma was detected from 30 min after exposure up to 8 h (half life = 4 h) using the dilution technique (Figure 5) and from 1–6 h using the diffusion technique (half life = 3 h) (Figure 6). Though both methods revealed that the peak of E. chlorantha active principle concentration in the plasma was situated at around 2 h, it was, however, noticed that the dilution technique was more sensitive than the diffusion technique.

A number of studies have shown that H. pylori eradication by certain drugs (namely, metronidazole, amoxicillin, and clarithromycin) is achieved by both systemic and topical action (Lamouliatte et al., 1991; Adamek et al., 1993; Goddard et al., 1996). After absorption into the bloodstream, the drug is secreted from the basal side back to the apical surface of the gastric epithelium where H. pylori is located (Matysiak-Budnik et al., 2002). Systemic antimicrobial activity cannot be observed if the drug is metabolized into an inactive form in the liver or during its trans-intestinal passage. The anti-Helicobacter activity of the plasma was observed within 30 min and up to 8 h after exposure to a single oral dose of 3000 mg/kg b.w. of E. chlorantha extract. This suggests that the in vivo activity of the extract may not result solely from the topical action but may also take place by a systemic component.

Multidrug therapy is currently recommended as one way of avoiding the rapid development of resistance by microorganisms that is partly responsible for treatment failure. This is being applied for the treatment of tuberculosis (Sterling et al., 1999) and malaria (Bloland et al., 2000), and for H. pylori eradication (Kato et al., 2000; Beales, 2001). Crude extracts are known to contain several active principles, and as such may be considered as combinations of active principles with potentiating or synergistic effects. This may explain why the development of resistance to crude extracts has hardly been reported. Sub-acute toxicity studies have revealed that the water extract of E. chlorantha is safe up to 500 mg/kg, and higher doses may provoke pathological incidents involving the heart, lungs, and liver (Tan et al., 2007). More detailed information on the in vivo H. pylori eradication potency using the mouse-adapted human H. pylori (the SS1 strain) is needed so that the water extract of E. chlorantha can be considered for the production of a cheap phytomedicine as an alternative to the expensive multi-antibiotic peptic ulcer regimens.

Declaration of interest

This research was supported by the International Foundation for Science, Stockholm, Sweden, and the United Nations University (UNU), Tokyo, Japan, through grant F/2882-3F awarded to one of the authors (P.V.T.).