In Vitro Susceptibilities of Four Bartonella bacilliformis Strains to 30 Antibiotic Compounds

Abstract

We have evaluated for the first time in vitro antibiotic susceptibilities of four human strains of Bartonella bacilliformis, the agent of Carrion’s disease. Our results show that B. bacilliformis, like other Bartonella species, is highly susceptible to antibiotics, including most beta-lactams, aminoglycosides, chloramphenicol, rifampin, macrolides, tetracyclines, cotrimoxazole, and fluoroquinolones.

Until 1992, Bartonella bacilliformis, the agent of Carrion’s disease, was the only member of the genus Bartonella (15). Following several taxonomic reappraisals, 13 validated species now exist, including five species which have been implicated in human diseases (5, 6, 10, 13). Unlike other members of the genus, which are widely considered emerging or reemerging pathogens, B. bacilliformis has been continually implicated in morbidity and mortality since its recognition in 1905 (15, 23). Carrion’s disease is endemic in the mountainous regions of Peru, Ecuador, and Colombia (1, 2, 12, 19), where the sand fly Lutzomyia verrucarum is considered the principal vector of B. bacilliformis. No reservoir other than humans has been demonstrated. Asymptomatic infection has been reported in several studies (2, 14). B. bacilliformis infection may be responsible for a life-threatening septicemia with acute hemolysis known as Oroya fever, especially in nonendemic people such as tourists and transient workers (12, 19, 22, 23), usually successfully treated with chloramphenicol (8, 19, 26). Among the endemic population, Carrion’s disease also presents as a chronic illness called verruga peruana which is characterized by benign cutaneous vascular lesions (3, 19). In this disease, caused by the same bacterium, chloramphenicol is ineffective, and streptomycin is considered the most reliable antibiotic therapy (19). However, these antibiotic regimens have been determined empirically, and in vitro antibiotic susceptibilities of B. bacilliformis have never been characterized. We have determined the MICs of 30 antibiotic compounds for four B. bacilliformis human strains by using an antibiotic agar dilution technique adapted to the fastidious growth of this group of bacteria.

The type strains of B. bacilliformis KC583 (ATCC 35685) and B. bacilliformis KC584 (ATCC 35686), which correspond to human isolates, were obtained from the American Type Culture Collection (Manassas, Va.). B. bacilliformis Acochaca 812 and B. bacilliformis Monzon 269 were isolated from the blood of Oroya fever patients in February 1998 and were kindly provided by H. Vizcarra (Instituto de Medicina Tropical Daniel A. Carrión, Universidad Nacional Mayor de San Marcos, Lima, Peru). Because of the fastidious nature of Bartonella spp., MIC determination assays as recommended by the National Committee for Clinical Laboratory Standards are not suitable. Thus, MICs were determined by an agar dilution technique previously described for other Bartonella species (20). The antibiotic test medium was Columbia agar (bioMérieux, Marcy l’Etoile, France) supplemented with 10% horse blood (bioMérieux). Antibiotics were diluted in Columbia agar to obtain final concentrations ranging from 0.0015 to 128 μg/ml. The B. bacilliformis strains were grown on horse blood agar, harvested after 6 days of incubation at 30°C, and suspended in phosphate buffer (pH 7.4) to obtain bacterial suspensions at a concentration equivalent to that of a 0.5 McFarland standard. Ten microliters of a 100-fold dilution of each inoculum (corresponding approximately to 10⁶ CFU/ml, i.e., 10⁴ CFU/spot, as determined by the CFU technique) was plated onto blood-supplemented agar. The plates were incubated at 30°C in a 5% CO₂ atmosphere for 6 days, an incubation period determined to be optimal for the evaluation of bacterial growth in antibiotic-containing agar and in drug-free controls. MICs required for complete inhibition of bacterial growth were recorded. Escherichia coli CIP 53126 and Staphylococcus aureus CIP 103811 were obtained from the Pasteur Institute (Institut Pasteur, Marnes La Coquette, France) and used as antibiotic test controls. MICs were determined for control strains either with Mueller-Hinton agar (bioMérieux) incubated at 37°C for 24 h or with 10% horse blood-supplemented Columbia agar, incubated at 37°C in a 5% CO₂ atmosphere, plates being read each day during a 6-day incubation period.

The four strains of B. bacilliformis displayed a homogeneous pattern of antibiotic susceptibilities (Table 1), and the MICs for the strains were also comparable to those previously determined for Bartonella quintana, Bartonella henselae, Bartonella elizabethae, and Bartonella vinsonii (16, 20). B. bacilliformis strains were highly susceptible to most beta-lactams (except oxacillin, cephalothin, and cefotetan), aminoglycosides, macrolides, doxycycline, and rifampin. High susceptibility to penicillin G and relative resistance to cephalothin have previously been described for other Bartonella species (20), as well as for Eikenella corrodens (27) and Capnocytophaga sp. (25). Clindamycin was also poorly effective in preventing bacterial growth. Among the fluoroquinolones, sparfloxacin and ciprofloxacin were the most effective. Vancomycin, whose action is limited almost entirely to gram-positive microorganisms, displayed surprisingly low MICs for the gram-negative B. bacilliformis strains. In contrast, MICs of colistin were higher than those usually described for gram-negative bacteria. These results are in accordance with previous experiments using other Bartonella species (20). They suggest specific properties of the Bartonella sp. cell wall. MICs determined for the control strains, with Mueller-Hinton agar and a 24-h incubation period, were very similar to those reported by the Pasteur Institute. The use of blood-supplemented Columbia agar, a CO₂-enriched atmosphere, and a prolonged incubation time did not significantly change MICs for the control strains (i.e., twofold or less increase in MICs). These results indicated that antibiotics tested may not have been significantly altered by specific incubation conditions used for B. bacilliformis MIC determination.

TABLE 1.

MICs for B. bacilliformis strains, as determined by the antibiotic agar dilution technique with Columbia agar supplemented with 10% horse blood

Antibiotic	MIC (μg/ml) for B. bacilliformis strain:
	ATCC 35685	ATCC 35686	Acochaca 812	Monzon 269
Penicillin G	0.015	0.015	0.03	0.03
Oxacillin	0.5	0.25	0.5	0.5
Amoxicillin	0.03	0.03	0.06	0.06
Amoxicillin-clavulanate	0.03-0.003	0.03-0.003	0.06-0.006	0.06-0.006
Ticarcillin	0.06	0.06	0.12	0.12
Cephalothin	8	4	8	8
Cefotetan	2	2	2	2
Cefotaxime	0.06	0.03	0.12	0.12
Ceftriaxone	0.003	0.003	0.006	0.003
Ceftazidime	0.25	0.12	0.25	0.25
Imipenem	1	0.5	1	1
Streptomycin	4	4	4	4
Gentamicin	1	1	2	2
Tobramycin	2	2	4	4
Amikacin	4	2	8	4
Erythromycin	0.06	0.06	0.06	0.06
Azithromycin	0.015	0.015	0.015	0.015
Clarithromycin	0.03	0.015	0.03	0.015
Roxithromycin	0.03	0.03	0.03	0.03
Clindamycin	64	32	64	64
Doxycycline	0.03	0.03	0.03	0.06
Pefloxacin	2	1	2	1
Ciprofloxacin	0.25	0.25	0.5	0.5
Sparfloxacin	0.25	0.25	0.25	0.25
Thiamphenicol	0.25	0.25	0.25	0.25
Trimethoprim-sulfamethoxazole	0.8-4	0.4-2	0.8-4	0.4-2
Rifampin	0.003	0.003	0.003	0.003
Fosfomycin	16	8	16	16
Colistin	16	16	16	16
Vancomycin	8	4	8	8

Before the antibiotic era, the only available treatment for the acute anemia associated with Oroya fever was blood transfusion, but this treatment had poor effectiveness and a high mortality rate of about 80% of cases (23, 26). Penicillin G, chloramphenicol, tetracycline compounds, and erythromycin have since been successfully introduced for the treatment of Oroya fever. However, as many patients suffer secondary infections, of which salmonellosis is the most common, chloramphenicol has become the recommended antibiotic therapy (8, 26). Nonetheless, therapeutic failures and persistent bacteremia have been reported when this drug is used. Cotrimoxazole has been used occasionally with success in some patients (19). More recently, anecdotal reports have indicated that newer macrolides (roxithromycin) and fluoroquinolones (norfloxacin and ciprofloxacin) may be effective in bacteremic patients (19). Penicillin G and chloramphenicol are not effective in the treatment of verruga peruana (19). Moreover, the treatment of Oroya fever with chloramphenicol does not prevent the appearance of verruga peruana (19). Streptomycin (2 mg/kg of body weight/day, 10 days) is considered the drug of choice in this chronic stage of Carrion’s disease. Intramuscular injections may, however, be difficult with children, and rifampin has been proposed as an alternative. In some series, the efficacy of rifampin has been evaluated as comparable to that of streptomycin (19), with the disappearance of cutaneous lesions within a month of therapy. However, failures have also been reported when this antibiotic is used, with subsequent improvement when the antibiotic therapy is changed to streptomycin (19). More recently, ciprofloxacin and the macrolide compounds erythromycin and roxithromycin have been used successfully in a limited number of patients (19).

We report for the first time an extensive study of the antibiotic susceptibilities of B. bacilliformis. These data are necessary not only to characterize newer antimicrobial agents potentially active in patients with Carrion’s disease but also to help define the optimum antibiotic therapy for other Bartonella-related diseases. The pathophysiology of Carrion’s disease and that of bacillary angiomatosis warrant comparison. Both include an acute stage, corresponding to a life-threatening bacteremia, and a chronic stage with cutaneous angiomatous lesions which are histologically indistinguishable (3, 7). Both B. henselae and B. bacilliformis may infect erythrocytes (4, 18) and endothelial cells (9, 11), in vitro and in vivo. Our results show that B. bacilliformis like other Bartonella species displays high in vitro susceptibility to most antibiotics. However, clinical data available for bacillary angiomatosis have previously shown that MICs may not be predictive of in vivo efficacy of the antibiotic therapy (17). In vitro, only the aminoglycosides are bactericidal against B. henselae, the agent of bacillary angiomatosis, both in axenic medium and in cell systems (21). Although many antibiotics are effective in the treatment of B. henselae bacteremia (24), cutaneous lesions of bacillary angiomatosis often relapse on antibiotic withdrawal (17). Likewise, the clinical experience of Peruvian medical doctors is that only streptomycin is a reliable antibiotic therapy for verruga peruana (8, 19, 26). We hypothesize that only antibiotics with intracellular penetration and bactericidal activity, such as aminoglycosides, will allow eradication of bacteria in patients chronically infected with Bartonella species. Investigations on the bactericidal activity of antibiotics against B. bacilliformis grown in axenic medium and in cells are under way in our laboratory.