Introduction

Helicobacter pylori is a gastric pathogen that colonises human stomach, leading to clinical manifestations of chronic gastritis, peptic ulcer disease, gastric adenocarcinoma and mucosa associated lymphoid tissue (MALT) lymphoma [1]. Current standard H. pylori eradication regime, also known as triple therapy, comprises two antibiotics (amoxicillin and clarithromycin or metronidazole) and a proton pump inhibitor (PPI) [5]. However, due to surging prevalence of clarithromycin and metronidazole resistance in H. pylori, the treatment success rate of either clarithromycin-based or metronidazole-based triple therapy is merely 80% or less in certain global regions [2,3,4,5,6]. To improve H. pylori eradication rate, different treatment strategies have been implemented including concomitant, sequential and bismuth-based quadruple regimens, as well as the culture-based, susceptibility-guided treatment approach [7, 8].

While different antibiotics have been used for the treatment of H. pylori infection, the bacterium has developed several mechanisms to protect itself against the antimicrobial activities of levofloxacin, tetracycline and rifabutin, in addition to metronidazole and clarithromycin [9]. Metronidazole resistance is complex but is primarily due to the inactivation of RdxA, an oxygen-insensitive NADPH nitroreductase [10]. Additionally, it has been demonstrated that the level of metronidazole resistance attributed to RdxA inactivation can be enhanced further by FrxA nitroreductase [11]. For clarithromycin resistance, substitution mutations involving the adenine residues at positions 2142 and 2143 were frequently detected in the 23 rRNA gene [12]. Furthermore, amino acid substitutions at positions 87 and 91 in DNA gyrase subunit A (GyrA) resulted in H. pylori resistance against fluroquinolones, whilst point mutations in pbp1 and 16S rRNA genes have been linked to amoxicillin and tetracycline resistance, respectively [13,14,15].

While high prevalence of metronidazole-resistant H. pylori isolates is common in several places in India like Kolkata, central Gujarat, Chandigarh, Delhi, Lucknow, Hyderabad and Chennai, the rates of H. pylori resistance to other antibiotics such as clarithromycin, ciprofloxacin, amoxicillin and tetracycline varied [16,17,18]. Since the antimicrobial resistance patterns of H. pylori vary between different geographical areas, it is vital to implement local antibiotic resistance surveillance program to provide physicians with more informed drug choices to effectively treat H. pylori-related gastric diseases. In India, regional antimicrobial resistance profile of H. pylori is generally lacking across the nation. Therefore, in this study, we aimed to examine the prevalence of H. pylori antibiotic resistance in Karnataka state of India, specifically against five antibiotics including metronidazole, clarithromycin, levofloxacin, tetracycline and amoxicillin. The findings of our work would provide a better understanding of the prevalence of drug-resistant H. pylori strains in Karnataka to facilitate the design of more rational and effective antibiotic combinations for eradication of H. pylori infection.

Results

From May 2014 to May 2017, 355 Karnataka patients underwent endoscopy at Kastruba Medical College and Tertiary Care Hospital, among which 180 (50.7%) were positive for H. pylori infection based on histopathology examination. A total of 113 H. pylori strains were successfully isolated, 80 (70.8%) from male and 33 (29.2%) from female patients. The average age was 46.2 ± 14 years old. Among 113 culture-positive patients recruited from nine districts in Karnataka state, two-third were from Udupi (31.9%, 36/113), Davangere (19.5%, 22/113) and Shimoga (15.9%, 18/113), and the rest were from Bellary (4.4%, 5/113), Chikmagalur (5.3%, 6/113), Chitradurga (10.6%, 12/113), Hassan (0.9%, 1/113), Haveri (4.4%, 5/113) and Uttar Kannada (7.1%, 8/113). Endoscopic examinations further diagnosed 78 (69%), 30 (26.6%) and 5 (4.4%) patients with gastritis, peptic ulcer disease and intestinal metaplasia, respectively. We next examined the susceptibility of each H. pylori strain against five different antibiotics including metronidazole, clarithromycin, levofloxacin, tetracycline and amoxicillin. The demographic data of patients and the MIC values for all strains are provided in Additional file 1.

Overall, 14.2% (16/113) strains were susceptible to all antibiotics used in this study (Table 1). Notably, 81.4% (92/113) of isolated strains were resistant to metronidazole. We also observed a high levofloxacin resistance rate among our strains (54.9%, 62/113). Resistance rates towards clarithromycin, tetracycline and amoxicillin were 20.4% (23/113), 5.3% (6/113) and 7.1% (8/113), respectively. Of 67 strains (59.3%, 67/113) that were multidrug-resistant, 58 (86.6%, 58/67) were resistant to at least both metronidazole and levofloxacin. While there was no association between patient ages and the metronidazole resistance rate of isolated clinical strains, metronidazole resistance was moderately lower in females than males (P = 0.06) (Table 2). Additionally, resistance to levofloxacin appeared to be the highest in patients aged 60 years and above.

Table 1 Antibiotic resistance rates of 113 H. pylori strains isolated in this study
Full size table
Table 2 Antimicrobial resistance according to gender and age distributions
Full size table

Of 113 strains included in this study, 38 had been sequenced at whole-genome level and reported in our earlier work [19]. Inactivation of RdxA and/or FrxA, nucleotide substitutions in 23S rRNA gene and amino acid substitutions in GyrA protein, conferring metronidazole, clarithromycin and levofloxacin resistance, respectively, were examined in these 38 strains to determine any of these mutations, if present, matches the susceptibility results. Among 36 sequenced metronidazole-resistant strains, 27 were expected to express non-functional or altered RdxA and/or FrxA proteins resulting from nonsense or frameshift mutations in the coding sequences, and partial gene deletions which would have been likely through recombination (Table 3). In the remaining metronidazole-resistant strains, as well as two metronidazole-susceptible strains, both genes were intact.

Table 3 Inactivation mutations in rdxA and frxA genes in sequenced metronidazole-resistant H. pylori strains
Full size table

Mutations in the 23 rRNA gene were detected in all nine clarithromycin-resistant sequenced strains, predominantly A2143G and G2224A single mutations (Table 4). One strain contained double mutations of A2142G and G2224A in its 23 rRNA gene. Out of 31 levofloxacin-resistant strains, 18 (58.1%) had amino acid substitution occurred at D91 in GyrA, eight (25.8%) at N87 and one at both positions (Table 5). We also noticed that two strains exhibited an unprecedented N-terminal extension of GyrA by five amino acid residues (QDNSV), immediately after the start codon, and one showed a mutation at R295.

Table 4 23S rRNA gene mutations in nine sequenced clarithromycin-resistant H. pylori strains
Full size table
Table 5 GyrA mutations in sequenced levofloxacin-resistant H. pylori strains
Full size table

Discussion

To our knowledge, this is the first study to examine the prevalence of antimicrobial resistance of H. pylori in Karnataka state of South India. In India, the prevalence of metronidazole-resistant H. pylori is remarkably high, with resistance rates reaching at least 80% in several areas [16,17,18, 20, 21]. Importantly, such high occurrence of metronidazole resistance was also observed in our isolated clinical strains, indicating that conventional triple therapy consisting of metronidazole should not be used to treat H. pylori infection in Karnataka. While the main mechanisms conferring H. pylori with metronidazole resistance involving RdxA and/or FrxA inactivation mutations were present in 75% of our sequenced metronidazole-resistant strains, in the remaining strains there were considerable number of missense mutations especially in rdxA gene that we cannot possibly rule out their role in metronidazole resistance by inducing protein conformational changes and hence dampening the reduction–activation activity of RdxA towards metronidazole [11, 22, 23].

In Maastricht V/Florence consensus report, it was recommended that clarithromycin-based triple therapy should be avoided if local clarithromycin resistance rate exceeds 15% [24]. As the primary clarithromycin resistance rate in Karnataka was 20.4%, which is above the 15% limit, suggesting that clarithromycin-based triple therapy is also not an appropriate option for the treatment of H. pylori infection in Karnataka. Resistance to clarithromycin is generally associated with point mutations in the 23S rRNA gene particularly at positions 2142 and 2143 [25]. In this study, the predominant mutations were A2143G and G2224A, where the latter has been linked to H. pylori clarithromycin resistance in the northeast part of China [26].

In tandem with a remarkably high prevalence of metronidazole resistance and a moderate occurrence of clarithromycin resistance, more than half of our isolated H. pylori strains were also resistant to levofloxacin, in agreement with prior reports of high levofloxacin resistance rates in several developing nations including Nepal, Vietnam and Bangladesh [27,28,29]. Levofloxacin resistance in H. pylori is acquired primarily by amino acid substitution mutations in GyrA protein involving N87 and D91 [30]. Similarly, 87.1% (27/31) of our sequenced levofloxacin-resistant strains were found to harbour these frequently reported mutations. Of four remaining strains, interestingly, two contained an insertion of QDNSV residues next to the start codon in their GyrA protein. It is highly likely that this five-amino-acid insertion would induce conformational changes in GyrA protein, reducing its binding affinity to fluoroquinolones and thus rendering levofloxacin ineffective. Further, we found one strain that had a R295H mutation in GyrA. Further mutagenesis study would be of interest to confirm the role of these novel GyrA mutations in H. pylori resistance to fluoroquinolones including levofloxacin.

Bearing in mind the high to modest rates of H. pylori resistance to metronidazole, levofloxacin and clarithromycin in this study, the use of triple therapy containing either combination of these drugs will likely result in treatment failure and should therefore be avoided. Our findings also suggest that triple therapy containing amoxicillin, tetracycline and PPI can be used as one of the first-line regimens for treating H. pylori infection in Karnataka. In a recent study conducted in Shanghai, China, where H. pylori metronidazole and clarithromycin resistance rates were 74.2% and 24.2%, respectively, at similar levels to Karnataka, the use of quadruple therapy containing bismuth, high-dose metronidazole to overcome existing resistance, clarithromycin and PPI had high modified intention-to-treat (ITT) and per-protocol (PP) cure rates of 90.3% and 96%, respectively [31]. In another study conducted in Zhejiang, China, where H. pylori clarithromycin resistance is prevalent, bismuth-based quadruple therapy containing amoxicillin and clarithromycin was shown to achieve both ITT and PP eradication rates of 86.1 and 92.3%, respectively [32]. Hence, it is worth examining the use and the effects of the above-mentioned bismuth-based quadruple therapies for the treatment of H. pylori infection in Karnataka population. If proven effective, local medical professionals would be able to continue using both relatively inexpensive metronidazole and clarithromycin drugs as part of H. pylori eradication regimens in Karnataka.

Conclusions

In our study, high resistance rates of metronidazole and levofloxacin resistance, and a modest occurrence of clarithromycin resistance were revealed in H. pylori strains isolated from patients in Karnataka, providing important information that triple therapies based on these drugs are not useful as first-line treatment in this South India region. Both amoxicillin and tetracycline are still useful for eradicating H. pylori infection in Karnataka. We also revealed novel mutations in GyrA protein that possibly contribute to H. pylori resistance in levofloxacin, which merit further investigations.

Methods

Gastric biopsies collection

This study was approved by the Manipal University Human Ethics Committee (Ref. No. IEC301/2014). Gastric biopsies were obtained for both histopathological examination and H. pylori culturing with informed and written consent from dyspeptic patients who underwent endoscopy at Kastruba Medical College and Tertiary Care Hospital, Manipal (KMC) from May 2014 to May 2017. In total, 113 H. pylori clinical strains were successfully isolated from patients who lived in Karnataka state of South India, had not previously received any H2 receptor blocker, PPI or nonsteroidal anti-inflammatory drug (NSAID) medication, and without any history of gastric surgery and H. pylori eradication treatment, were recruited. Four biopsy specimens (two each from antrum and corpus) were collected in Bouin’s fluid for histopathological examination. Two additional specimens (one each from antrum and corpus) were collected in transport medium for bacterial isolation.

Bacterial culturing

Transport medium containing biopsy specimens was vortexed prior to plating 100 µl of the solution on brain heart infusion agar (BHIA) supplemented with 5% (vol/vol) horse serum (Gibco, New Zealand), 0.5% (vol/vol) BBL IsoVitaleX enrichment medium (Becton, Dickinson and Company, USA), trimethoprim (5 µg/ml) and vancomycin (6 µg/ml) (HiMedia Ltd., India). The plates were then incubated in a microaerobic atmosphere (5% O2, 10% CO2 and 85% N2) at 37 °C for a period of 3 to 7 days. Successful H. pylori isolate was identified based on its distinctive colony morphology, appearance of Gram-negative bacillus under the microscope, positive urease, oxidase and catalase reactions, and PCR detection of H. pylori-specific glmM gene.

Minimum inhibitory concentration (MIC) determination using agar dilution method

Plates were prepared containing 5% (vol/vol) horse serum and serial two-fold dilutions of each antibiotic. The range of final concentrations for each antibiotic were 2–64 µg/ml for metronidazole, 0.125–1 µg/ml for clarithromycin, 0.25–1 µg/ml for amoxicillin, and 1–4 µg/ml for both levofloxacin and tetracycline. Bacterial suspension with turbidity of 2.0 McFarland standard was prepared from a 24-h old culture plate in saline. Two µl of bacterial suspension was inoculated onto each antibiotic plate, followed by incubation in a microaerobic atmosphere at 37 °C for 3 days. The MIC was then determined at the lowest antibiotic concentration where bacterial growth was completely inhibited. According to the guidelines of European Committee on Antimicrobial Susceptibility Testing (EUCAST), bacterial isolates are considered to be antibiotic-resistant if the MIC value is > 8 µg/ml for metronidazole, > 0.5 µg/ml for clarithromycin, > 0.12 µg/ml for amoxicillin, and > 1 µg/ml for both tetracycline and levofloxacin.

Mutation analysis

Thirty-eight draft genomes were annotated using Prokka v1.12 [33]. The Prokka-annotated draft genomes are available at the public data repository Figshare (https://figshare.com/), with https://doi.org/10.6084/m9.figshare.8016128. To obtain rdxA, frxA, 23S rRNA and gyrA nucleotide sequences from each strain, blastn search of hp0954, hp0642, hpr01 and hp0701, respectively, from H. pylori 26695 reference genome was performed on CLC Genomic Workbench 11 (Qiagen, Germany). Extracted rdxA, frxA and gyrA genes were translated into amino acid sequences prior to alignment. For extracted 23S rRNA genes, sequence alignment was performed at the nucleotide level. Aligned sequences were then compared between resistant and sensitive strains, with reference sequence included, to examine for reported and novel mutations.

Statistical analysis

For statistical analysis, Fisher’s exact test and Fisher-Freeman-Halton’s test were used, as appropriate. Only p value of < 0.05 was considered statistically significant.