Population structure of Helicobacter pylori and antibiotic resistance-associated variants in a high-risk area of gastric cancer

Abstract

The increasing antibiotic resistance of Helicobacter pylori has had a serious impact on gastric cancer prevention. Our study aimed to profile the genomic characteristics and explore variants associated with resistance in H. pylori strains from a high-risk area of gastric cancer in China. We isolated 153 strains from a community-based cohort and assessed their susceptibility to six antibiotics by MIC Test Strip and genomic characteristics by whole-genome sequencing. Phylogenetic analysis identified the strains as an independent cluster within H. pylori East Asian population (hpEastAsia). HefA, an efflux pump gene, showed the highest differentiation in the Linqu strains compared with the other Chinese strains. Bacterial genome-wide association study (GWAS) identified 86 resistance variants covering 44 genes. Novel resistance variants were found in lon and babA for metronidazole, HP1168 for clarithromycin, hcpC for levofloxacin, and sabA for rifamycin. Two newly identified hefA mutations (R229K and A283V) showed significant associations with metronidazole (P = 0.012) and tetracycline (P = 0.044) resistance, respectively. HefA mutations and GWAS variants were integrated with the significant literature-reported mutations to optimize the prediction models for metronidazole, levofloxacin, clarithromycin, and tetracycline resistance with area under the receiver operating characteristic curves of 0.82-0.93. Double-antibiotic resistance models were established for clinical applicability. Furthermore, hefA expression may play a potential mediating role in the associations between mutations and resistance. This study identified genetic independence in the representative H. pylori strains from a high-risk area of gastric cancer. Optimized resistance prediction panels, including novel hefA mutations and GWAS variants, may provide preliminary guidance for localized precise treatment and helpful experiences for the similar high-risk populations.IMPORTANCEHelicobacter pylori is a remarkable pathogen due to its virulence in gastric cancer and high genetic plasticity. Linqu County in China, a high-risk area of gastric cancer, faces serious antibiotic resistance issues and necessitates genomic profiling of local H. pylori strains. Phylogenetic analysis revealed the Linqu strains as a relatively independent cluster within the hpEastAsia population. Novel antibiotic resistance-associated hefA mutations and variants from our bacterial genome-wide association study in the Linqu strains were optimized to improve the prediction performances for single antibiotic and double-drug combination resistance compared with traditional literature-reported mutations. This study identified relative genetic independence and high differentiation in the representative H. pylori strains from a population with high risk of gastric cancer and high prevalence of antibiotic resistance. The optimized panels with novel variants improve antibiotic resistance prediction models compared with literature-reported mutations, providing guidance for localized precise treatment and suggesting prevention strategies for similar high-risk populations.

Keywords: Helicobacter pylori; antibiotic resistance; gastric cancer; whole-genome sequencing.

Fig 1

Antibiotic resistance profiling of the Linqu strains. (A) The resistance rates in the Linqu strains to AC, CH, LE, MZ, TC, and RI. (B) The antibiotic resistance patterns in the Linqu strains, describing the numbers of resistant strains and antibiotic combinations in the single- and multiple-drug resistance groups. AC, amoxicillin; CH, clarithromycin; LE, levofloxacin; MZ, metronidazole; RI, rifamycin; TC, tetracycline; R, resistant; S, susceptible.

Fig 2

Population structure of Linqu strains and global published strains co-ancestry matrix was calculated using fineSTRUCTURE, including 153 Linqu strains and 1541 published strains with 824 from other regions of China, 494 from Japan, 71 from Korea, and 152 from other continents. The columns on the top and left side show the different geographic regions and populations of the H. pylori strains with distinct colors.

Fig 3

Manhattan plots of the genome-wide association study for antibiotic resistance. Significant antibiotic resistance-associated genetic variations were selected by GWAS between the sensitive and resistant strains for (A) AC, (B) CH, (C) LE, (D) TC, (E) MZ, and (F) RI. Log₁₀ (P) for each hit is recorded on the vertical axis. The gray dashed line indicates P = 10⁻⁵. GWAS, genome-wide association study; AC, amoxicillin; CH, clarithromycin; LE, levofloxacin; TC, tetracycline; MZ, metronidazole; RI, rifamycin.

Fig 4

The resistance prediction model construction and validation in the Linqu strains. Prediction models were constructed and validated for single and double-drug resistance, including (A) CH, (B) LE, (C) MZ, (D) TC, (E) CH + LE, CH + MZ, MZ + TC. ROC curve analyses were conducted to evaluate the prediction models including model 1 (literature-reported mutation panel) and model 2 (optimized panel combining literature-reported mutations, hefA mutations, and novel GWAS variants). The model uses genetic data as input to predict susceptible or resistant phenotypes as output. GWAS, genome-wide association study; CH, clarithromycin; LE, levofloxacin; MZ, metronidazole; TC, tetracycline; ROC, receiver operating characteristic.

Fig 5

Associations among hefA mutations, expression, and antibiotic resistance. (A) Association between A283V mutation and the MIC values for TC. (B) Association between R229K mutation and the MIC values for MZ. (C) Association between hefA expression and antibiotic resistance patterns. (D) Mediation analysis for the involvement of hefA expression in the association between A283V mutation and TC resistance. Mediated effect: percentage of the total effect of hefA variants on antibiotic resistance mediated by hefA expression; IE: indirect effect; DE: direct effect. (E) Mediation analysis for the involvement of hefA expression in the association between R229K mutation and MZ resistance. MIC, minimum inhibitory concentration; TC, tetracycline; MZ, metronidazole.