Molecular Basis of Resistance Displayed by Highly Ciprofloxacin-Resistant Salmonella enterica Serovar Typhi in Bangladesh

The emergence of multidrug-resistant typhoid strains has led to the use of ciprofloxacin as the drug of choice for routine outpatient treatment (11, 13). This has resulted in the rapid emergence of Salmonella enterica serovar Typhi strains with reduced susceptibility to ciprofloxacin due to a point mutation in the gyrA gene, at either Ser-83 or Asp-87 (1, 7, 12, 14, 17). These strains are associated with failure of or a delayed response to ciprofloxacin therapy (16). These strains are still considered conventionally susceptible to ciprofloxacin based on their Clinical Laboratory Standards Institute (CLSI) cutoff (6) (MICs of ≥2.0 μg/ml), but they are resistant to nalidixic acid and thus are termed nalidixic acid-resistant S. enterica serovar Typhi (NARST). However, screening with nalidixic acid discs, as recently recommended by the CLSI, is not yet practiced routinely in clinical laboratories. Until now, only one other typhoid strain with complete resistance (MIC, 16 μg/ml) to ciprofloxacin has been reported (10). Another strain with complete resistance was isolated from a case of osteomyelitis (18).

In this report, we describe the characteristics of three highly ciprofloxacin-resistant (MIC, 512 μg/ml) strains of S. enterica serovar Typhi isolated from the blood of patients diagnosed with typhoid fever, and we report the mechanism of resistance of these strains.

In the first week of November 2005, ciprofloxacin-resistant S. enterica Typhi strains were isolated from the blood of three patients (an 11-year-old female, a 21-year-old female, and a 21-year-old male) from different parts of Dhaka, Bangladesh, all of whom had a piped water supply. Blood was collected from two of the patients on days 4 and 6 after the initiation of ciprofloxacin therapy (20 to 25 mg/kg of body weight/day). The third patient presented on day 1 of illness without a history of prior antibiotic use.

Strains were identified by standard biochemical and serological procedures (4) and subjected to testing by Etest to determine the MICs for various antibiotics. Susceptibility to azithromycin was determined by disc diffusion only, as an Estrip for this drug was not available. The MIC of ciprofloxacin was further determined by microbroth dilution to find the end point. The results were interpreted following CLSI guidelines (6). Isolates were highly resistant to ampicillin (>256 μg/ml), cotrimoxazole (>32 μg/ml), chloramphenicol (>256 μg/ml), ciprofloxacin (512 μg/ml), and nalidixic acid (>256 μg/ml) and were susceptible to ceftriaxone (0.094 μg/ml) and azithromycin (zone, 20 to 22 mm). Based on the culture results, all three patients were treated with ceftriaxone and were cured without complications.

Biochemical profiles were determined by Api 20E testing (bioMerieux SA, Marcy l'Etoile, France), and scores were recorded according to the manufacturer's instructions.

Molecular typing was done by multiplex PCR for variable-number tandem repeats (VNTRs), using primers flanking three VNTR loci (TR1, TR2, and TR3) (Table 1), as described by Liu et al. (9). In brief, each 25-μl reaction mixture contained 1.5 μl of the bacterial lysate suspension and 10 to 12 pmol each of the corresponding forward and reverse primers for the TR1, TR2, and TR3 loci in 15 μl of Taq PCR master mix (QIAGEN GmbH, Hilden, Germany). After initial denaturation at 94°C for 5 min, the PCR was run for 35 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min, followed by a final extension at 72°C for 7 min. The PCR products, along with a 100-bp DNA marker, were subjected to electrophoresis on a 1.3% agarose gel (Invitrogen, Carlsbad, CA).

DNA sequencing was carried out by the dideoxynucleotide chain termination method, using an automated DNA sequencer (ABI Prism 3100 genetic analyzer; Perkin-Elmer Applied Biosystems, Foster City, CA). Sequences were edited using DNASTAR's Lasergene sequence analysis software (2). Multiple sequence alignment was performed using CLUSTAL X (15) to check for a mutation(s) in the amino acid sequence at positions 54 to 171 of GyrA, which contains the quinolone resistance-determining region.

All three isolates were found to be identical by their Api 20E profile (4404552) and VNTR type, but their VNTR type differed from the VNTR patterns found for ciprofloxacin-sensitive strains (Fig. 1). Sequencing of the PCR products showed the presence of a double mutation. There were nucleotide changes at codons 83 and 87, from TCC to TTC and GAC to GGC, respectively, resulting in corresponding changes from serine to phenylalanine and aspartate to glycine, which were identical for all three isolates. In contrast, previous studies on the mutation of NARST strains identified a single mutation at either position Ser-83 (TCC) or Asp-87 (GAC) (14). The isolation of indistinguishable strains at the same time and from unrelated cases suggests that they are from a common source (possibly a piped water supply) and probably evolved recently.

Our study has the limitation that patient information was collected retrospectively, only after confirmation that the strains were S. enterica serovar Typhi and resistant to ciprofloxacin. Furthermore, we could not investigate for possible mutations at other known loci for fluoroquinolone resistance, including gyrB, parC, and parE (8). Thus, the high levels of resistance of these S. enterica serovar Typhi strains could be the cumulative impact of simultaneous mutations at other sites and/or activation of an efflux pump (3).

The observation of the emergence of strains of S. enterica serovar Typhi with complete resistance to ciprofloxacin is of major public health concern. Oral ciprofloxacin is the drug of choice for typhoid fever and is relatively affordable and widely used. In recent years, we have observed many cases of ciprofloxacin treatment failure, although the MICs of the strains ranged from 0.25 to 0.75 μg/ml (unpublished data). Nonetheless, ciprofloxacin is used extensively in the community because laboratories routinely report these strains as sensitive, based on the current CLSI guideline (cutoff, zone of ≥20 mm or MIC of ≤2 μg/ml), and screening with nalidixic acid, as recently recommended by CLSI, is not generally practiced.

The emergence of this highly resistant strain in Bangladesh may be due to the widespread use of ciprofloxacin in a population with a high prevalence of NARST. Under such selective pressure, the antibiotic-multiresistant clone described here is likely to become dominant. Consequently, the only readily available treatment will be either azithromycin (5) or expanded-spectrum cephalosporins, which are unaffordable to the majority of the Bangladeshi population. Furthermore, the efficacy of these more expensive antibiotic classes is not known for truly ciprofloxacin-resistant S. enterica serovar Typhi strains. An investigation of the extent of this resistance problem in the Indian subcontinent is urgently needed. If such strains are widespread, as expected, a randomized, controlled trial comparing oral azithromycin to expanded-spectrum cephalosporin treatment of typhoid fever due to these resistant strains is needed to better inform the choice of optimal treatment.