In Vitro Antimicrobial Susceptibility Testing of Helicobacter felis, H. bizzozeronii, and H. salomonis

Abstract

The susceptibilities of Helicobacter felis (15 strains), H. bizzozeronii (7 strains), and H. salomonis (3 strains) to 10 antimicrobial agents were investigated by determination of the MIC using the agar dilution method. No consistent differences were noticed between the different Helicobacter species, which were all highly susceptible to ampicillin, clarithromycin, tetracycline, tylosin, enrofloxacin, gentamicin, and neomycin, as demonstrated by low MICs. Higher MICs were obtained for lincomycin (up to 8 μg/ml) and spectinomycin (up to 4 μg/ml). Two H. felis strains showed a MIC of 16 μg/ml for metronidazole, suggesting acquired resistance to this antimicrobial agent.

“Helicobacter heilmannii,” the provisional name of tightly coiled gram-negative bacteria occurring in 0.2 to 2.4% of human gastric biopsy specimens, is associated with various types of gastric pathology (1, 16). Although chronic active gastritis is the predominating pathological profile (6, 16), it appears to a milder degree than gastritis induced by H. pylori, a leading cause of peptic ulcer disease (7, 29). The gastritis may worsen into gastric erosions (6, 30), gastric carcinomas, or mucosa-associated lymphoid tissue lymphomas (9, 23, 28, 29), requiring an effective treatment protocol (21). Commonly a multiple drug regimen is used for H. pylori infections, since single-drug treatment proved to be inefficient (11). Recommended antibiotic agents are clarithromycin and amoxicillin as first-line therapy, and metronidazole together with tetracycline as second-line treatment, both schemes in combination with a gastric-acid-secretion inhibitor (21). Nevertheless, the success rate of these regimens is relatively low, due to the emerging antimicrobial resistance of H. pylori (8, 10). Similar treatment regimens have been used empirically in cases of “H. heilmannii” infections (23), since these bacteria are not culturable in vitro (23, 31), preventing the exploration of their susceptibility pattern. Only once has the in vitro culture of a tightly coiled bacterium from a human stomach, subsequently identified as H. bizzozeronii, been described (2, 18).

Bacteria strongly affiliating with “H. heilmannii,” both on a morphological and on a genetic basis, occur in the stomachs of dogs and cats. Hitherto, three different species, H. felis, H. bizzozeronii, and H. salomonis, phenotypically and phylogenetically highly related to one another, have been identified as natural inhabitants of the gastric mucosae of these pet animals (19, 26). Their pathological significance for dogs and cats is still a matter of debate. Antimicrobial treatment of these animals is considered only when helical bacteria are abundant and accompanied by mucosal inflammation and clinical symptoms (26). Veterinarians have to rely on antimicrobials used to eradicate H. pylori (14), as the fastidious nature of H. felis, H. bizzozeronii, and H. salomonis has hitherto impeded the exploration of their antibiotic susceptibility patterns. Recently, these Helicobacter species have been identified in human gastric biopsy specimens histologically found positive for “H. heilmannii,” indicating that the latter name comprises an amalgam of different Helicobacter species, at least to some extent originating from dogs and cats (31).

It was the purpose of the present study to investigate the susceptibilities of H. felis, H. bizzozeronii, and H. salomonis to antimicrobial agents commonly used for the treatment of gastritis, in order to improve the management of “H. heilmannii” infections both in humans and in animals.

A total of 15 strains of H. felis, 7 strains of H. bizzozeronii, and 3 strains of H. salomonis, isolated from the gastric mucosae of different cats and dogs and identified by a recently developed multiplex PCR (3), were included in this study (Table 1). The organisms were cultured on brain heart infusion agar (Oxoid, Basingstoke, England) as described previously (3). All incubations occurred at 37°C under microaerobic conditions with 5% oxygen in a closed circuit, created by evacuating 80% of the ambient atmosphere and introducing a gas mixture of 8% CO₂, 8% H₂, and 84% N₂. The isolates were passaged twice to ensure reliable growth and purity.

TABLE 1.

Canine and feline gastric Helicobacter strains used in this study

Species and strain no.	Source	Origin
H. felis
CS1 (CCUG 28539T)	Cat	Australia
DS1	Dog	Australia
CS5	Cat	Australia
CS6	Cat	Australia
CCUG 37471	Dog	Finland
M32	Dog	Slovenia
M31	Dog	Slovenia
JKM1	Cat	Belgium
JKM2	Dog	Belgium
JKM3	Dog	Belgium
M26	Dog	Slovenia
M29	Dog	Slovenia
M35	Dog	Slovenia
M38	Dog	Slovenia
M39	Dog	Slovenia
H. bizzozeronii
Storkis (CCUG 35545T)	Dog	Finland
M20	Dog	Slovenia
M5	Dog	Slovenia
M7	Dog	Slovenia
M10	Dog	Slovenia
M12	Dog	Slovenia
M25	Dog	Slovenia
H. salomonis
Inkinen (CCUG 37845T)	Dog	Finland
M45	Dog	Slovenia
M50	Dog	Slovenia

Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 29213, and Campylobacter jejuni ATCC 33560, grown on Columbia agar (Oxoid) supplemented with 5% sheep blood, were included as reference strains.

Susceptibility to metronidazole, ampicillin, clarithromycin, lincomycin, tetracycline, tylosin, enrofloxacin, gentamicin, spectinomycin, and neomycin was investigated by the agar dilution method. All antibiotics were supplied by Sigma (St. Louis, MO) as standard powders with known potencies, except for enrofloxacin, purchased from Bayer (Brussels, Belgium). The compounds were dissolved and diluted according to the guidelines of the National Committee for Clinical Laboratory Standards (NCCLS) (25). MIC tests were carried out on Mueller-Hinton II agar (Becton Dickinson, Cockeysville, Md.), supplemented with 10% horse blood. Agar plates contained serial twofold dilutions of the above-stated antibiotics, with final concentrations ranging from 0.03 to 512 μg/ml. Agar plates free of the tested antibiotics were included as controls.

Helicobacter and Campylobacter bacteria grown for 72 h were harvested and suspended in sterile saline to a density of 3 on the McFarland turbidity scale. Suspensions with a density of 0.5 McFarland standard were prepared from overnight-grown reference strains of E. coli and S. aureus. Antibiotic-containing and antibiotic-free plates were seeded by a Steers inoculum replicator (MAST, London, United Kingdom). Plates were read after 1 day for the reference strains and after 4 and 7 days of incubation for both the Helicobacter isolates and the reference strains. The MIC was determined as the lowest concentration of an antimicrobial drug inhibiting visible growth, disregarding a faint haze, of the tested Helicobacter isolates after 7 days of incubation in a microaerophilic atmosphere.

Results for the Helicobacter species are summarized in Table 2. Not all MICs of the antimicrobials tested for the C. jejuni reference strain are included in the NCCLS guidelines. MICs ranged from 0.5 to 1 μg/ml for metronidazole, from 0.5 to 2 μg/ml for clarithromycin, from 1 to 4 μg/ml for spectinomycin, from 4 to 8 μg/ml for ampicillin, and from 4 to 32 μg/ml for lincomycin. The MICs determined for tylosin and neomycin were 16 and 1 μg/ml, respectively.

TABLE 2.

Distribution of MICs for 15 H. felis, 7 H. bizzozeronii, and 3 H. salomonis strains of animal origin

Antimicrobial agent	Species	No. of strains with a MIC (μg/ml) of:
		≤0.03	0.06	0.125	0.25	0.5	1	2	4	8	16	32	64	128	256	512
Ampicillina	H. felis	1	7	5	2
	H. bizzozeronii		4	2
	H. salomonis	1	1	1
Clarithromycin	H. felis	10	3	1	1
	H. bizzozeronii	6		1
	H. salomonis	3
Enrofloxacin	H. felis	8	7
	H. bizzozeronii	3	2	2
	H. salomonis	1	1	1
Gentamicin	H. felis		5	8	2
	H. bizzozeronii		3	3	1
	H. salomonis			3
Lincomycinab	H. felis				1	2	5	4	1	2
	H. bizzozeronii					1		3		1
	H. salomonis				2		1
Metronidazolea	H. felis					6	4	3			2
	H. bizzozeronii						1	3	1	1
	H. salomonis					3
Neomycin	H. felis			3	9	3
	H. bizzozeronii				5	2
	H. salomonis				2	1
Spectinomycinac	H. felis					2	2	9	1
	H. bizzozeronii					3		3
	H. salomonis							3
Tetracycline	H. felis			2	12		1
	H. bizzozeronii				4	1	2
	H. salomonis			1	1	1
Tylosina	H. felis	6	3	5	1
	H. bizzozeronii			1	3	2
	H. salomonis	1		2

^aThe MIC of H. bizzozeronii M25 could not be determined, as no growth was obtained on control plates.

^bThe MIC of H. bizzozeronii M20 could not be determined, as no growth was obtained on control plates.

^cThe MIC of H. felis CCUG 37471 could not be determined, as growth on control plates was inferior.

The fastidious and slow-growing nature of these bacteria led us to incubate the plates longer than 72 h, which is the time recommended by the NCCLS for susceptibility testing of H. pylori (25). In a first stage, the incubation period was extended to 4 days. Growth was definable only for a minority of the isolates after this time, and a 7-day incubation period revealed no different outcome of these MICs. This prolongation markedly facilitated determination of the MIC due to a more pronounced growth of the Helicobacter isolates. After incubation for 7 days, the MICs of the reference strains for all antimicrobial agents tested were 1 to 2 log₂ dilutions higher than the MICs recorded after 1 day of incubation, but still in the range premised by the NCCLS (24).

In general, the MICs did not show consistent differences between the different Helicobacter species. All isolates seemed to be highly susceptible to ampicillin, clarithromycin, enrofloxacin, gentamicin, tetracycline, tylosin, and neomycin, as low MIC levels were noted. The MICs of clarithromycin, amoxicillin, and tetracycline are completely in line with those indicating susceptibility in H. pylori (10). MICs of lincomycin, metronidazole, and spectinomycin were markedly higher. For metronidazole, a broad MIC range was noticed.

Selection of antimicrobial agents was based on their common application for H. pylori eradication in humans and/or for treatment of gastrointestinal disease in companion animal medicine. The latter justifies the inclusion of lincomycin, tylosin, spectinomycin, gentamicin, and neomycin in the study protocol next to ampicillin, clarithromycin, metronidazole, and tetracycline. Enrofloxacin is fairly frequently used in small-animal veterinary medicine for the treatment of infections with principally gram-negative bacteria. Ciprofloxacin, the primary active metabolite of enrofloxacin, expressing an antimicrobial activity similar to that of enrofloxacin (32), is considered an important alternative for H. pylori eradication in cases of metronidazole and/or clarithromycin resistance (4, 27).

Currently, no standard methods are described for MIC determination for non-H. pylori Helicobacter species. The NCCLS proposes the agar dilution method using Mueller-Hinton agar supplemented with sheep blood as the method of choice for Campylobacter jejuni and related species (24) and for H. pylori (25). In the present study, however, horse blood was chosen, because it is known to support excellent growth of the organisms of interest and its supplementation is advised by the NCCLS for testing fastidious organisms. A typical feature of H. felis, H. bizzozeronii, and H. salomonis is the need for fresh, moist agar plates, on which viable bacterial cells grow as a spreading film (13, 17). This growth characteristic was thought to impede agar dilution testing, leading to the assumption that animal models would be the only way to assess antimicrobial drug susceptibility, a method which has already been applied to test H. felis (5). In the present study, all isolates except H. bizzozeronii strains M20 and M25 consistently grew well on dry plates as single colonies. The explanation for the poor growth of these particular strains may be found in the lack of specificity of the McFarland standard, which is an indication of the amount of cellular material rather than of the quantity of viable bacteria (15). The same phenomenon has already been noted for H. pylori (15).

Up to now, the only antimicrobial agent for which the NCCLS guidelines have established a breakpoint in the case of H. pylori susceptibility testing is clarithromycin. H. pylori is considered resistant when the MIC is ≥1 μg/ml (25). Other breakpoints to define the resistance level of H. pylori have only been suggested in the scientific literature for metronidazole (≥8 μg/ml), tetracycline (≥2 μg/ml), and ampicillin (≥8 μg/ml) (20, 22). Upon extrapolation of these breakpoints to H. felis, H. bizzozeronii, and H. salomonis, acquired resistance to metronidazole was observed for one H. bizzozeronii isolate (MIC, 8 μg/ml) and two H. felis isolates (MIC, 16 μg/ml). All three H. salomonis isolates in the present study were susceptible to metronidazole, although resistance to this agent has already been reported in the past (17). These observations urge new investigations on the mechanism responsible for metronidazole resistance in non-H. pylori Helicobacter species. The mechanism in H. pylori is based on nonsense mutations in the rdxA gene, a gene coding for an oxygen-insensitive NADPH nitroreductase, which enables the reduction of metronidazole into its active component hydroxylamine (12). The genetic alterations generate a premature stop in the translated protein, thereby inactivating the nitroreductase and the antimicrobial activity of metronidazole.

Whether our findings suggest the possibility of efficiently applying these antimicrobials in monotherapy for eradication of non-H. pylori Helicobacter spp. is questionable. Experimental studies with mice intragastrically infected with H. felis proved monotherapy of metronidazole, erythromycin, tetracycline, or amoxicillin all ineffective for eradication of these spirally shaped organisms (5). With respect to the possible discrepancy between in vitro and in vivo studies, one may refer to the behavior of H. pylori, which is highly susceptible to antimicrobials in vitro while its eradication in patients requires aggressive triple or quadruple therapy (8, 11).

The present study is the first to investigate the normal in vitro susceptibilities of H. felis, H. bizzozeronii, and H. salomonis to several antimicrobials, notwithstanding their fastidious nature and the lack of standardized guidelines on MIC testing for these Helicobacter species.