Emergence of amoxicillin resistance and identification of novel mutations of the pbp1A gene in Helicobacter pylori in Vietnam

Background Amoxicillin-resistant Helicobacter pylori (H. pylori) strains seem to have increased over time in Vietnam. This threatens the effectiveness of H. pylori eradication therapies with this antibiotic. This study aimed to investigate the prevalence of primary resistance of H. pylori to amoxicillin and to assess its association with pbp1A point mutations in Vietnamese patients. Materials and methods Naive patients who presented with dyspepsia undergoing upper gastrointestinal endoscopy were recruited. Rapid urease tests and PCR assays were used to diagnose H. pylori infection. Amoxicillin susceptibility was examined by E-tests. Molecular detection of the mutant pbp1A gene conferring amoxicillin resistance was carried out by real-time PCR followed by direct sequencing of the PCR products. Phylogenetic analyses were performed using the Tamura-Nei genetic distance model and the neighbor-joining tree building method. Results There were 308 patients (46.1% men and 53.9% women, p = 0.190) with H. pylori infection. The mean age of the patients was 40.5 ± 11.4 years, ranging from 18 to 74 years old. The E-test was used to determine the susceptibility to amoxicillin (minimum inhibitory concentration (MIC) ≤ 0.125 μg/ml) in 101 isolates, among which the rate of primarily resistant strains to amoxicillin was 25.7%. Then, 270 sequences of pbp1A gene fragments were analysed. There were 77 amino acid substitution positions investigated, spanning amino acids 310–596, with the proportion varying from 0.4 to 100%. Seven amino acid changes were significantly different between amoxicillin-sensitive (AmoxS) and amoxicillin-resistant (AmoxR) samples, including Phe366 to Leu (p <  0.001), Ser414 to Arg (p <  0.001), Glu/Asn464–465 (p = 0.009), Val469 to Met (p = 0.021), Phe473 to Val (p <  0.001), Asp479 to Glu (p = 0.044), and Ser/Ala/Gly595–596 (p = 0.001). Phylogenetic analyses suggested that other molecular mechanisms might contribute to amoxicillin resistance in H. pylori in addition to the alterations in PBP1A. Conclusions We reported the emergence of amoxicillin-resistant Helicobacter pylori strains in Vietnam and new mutations statistically associated with this antimicrobial resistance. Additional studies are necessary to identify the mechanisms contributing to this resistance in Vietnam.


Introduction
Helicobacter pylori (H. pylori) is a Gram-negative microaerophilic pathogenic bacterium that colonises the human gastric mucosa. It has been categorised as a Group I carcinogen for gastric cancer [1]. It infects approximately 50% of individuals worldwide. In Vietnam, the prevalence of H. pylori is very high (75%) [2]. The rates of H. pylori infection in Vietnamese patients with chronic gastritis, gastric ulcers, duodenal ulcers, and gastroduodenal ulcers range from 59.9-69.9, 77.8, 85-95%, and 85.3-93.6%, respectively [3]. The incidence of gastric cancer in Vietnam has been reported to be the highest compared to other southeast Asian countries [4]. Successful H. pylori eradication has been demonstrated to not only heal peptic ulcers but also prevent their recurrence and reduce the incidence of gastric cancer [3,5].
Amoxicillin is one of the key antibiotics often used to eradicate H. pylori in standard triple therapy (amoxicillin, clarithromycin, and proton pump inhibitor) as the first-line treatment [6]. However, the spread of amoxicillin-and/or clarithromycin-resistant H. pylori has become an important cause of eradication failure. A high rate of clarithromycin resistance of H. pylori (72.6%) has been well documented in Vietnam [7]. Additionally, recent data in Vietnam have shown that the amoxicillin resistance rate of H. pylori differs across regions and it seems to have increased over time, from 1.1% in 2015 to 10.4% in 2016 and 15% in 2018 [3,8,9]. This threatens the effectiveness of H. pylori eradication therapies with these antibiotics.
Point mutations in the genes coding for penicillinbinding proteins (PBPs) lead to a decreased affinity for the drug and result in amoxicillin resistance [10,11]. Among these genes, the pbp1A gene seems to be one of the keys [12][13][14][15]. There have been several studies on the molecular antibiotic resistance of H. pylori in Vietnam, but most of them focused on clarithromycin and levofloxacin resistance [7,9,16]. This is the first study conducted to investigate the molecular mechanism of amoxicillin resistance in H. pylori in Vietnam.
The present study aimed to evaluate the prevalence of amoxicillin-resistant H. pylori-positive gastric biopsy specimens and to investigate point mutations in the pbp1A gene in H. pylori-sensitive and primary-resistant samples. This study contributes to uncovering novel point mutations in the pbp1A gene underlying this resistance effect.

Patient population and clinical specimens
This study was conducted between January 2019 and April 2021 at the University Medical Center, Ho Chi Minh City, Vietnam. A total of 308 Vietnamese patients with positive H. pylori infection by rapid urease test, presenting with naive dyspepsia and aged ≥18 years were recruited. The exclusion criteria included patients with gastric cancer and/or those who took any antibiotics within the last 4 weeks.
The sample size was calculated by RaoSoft ® software (http:// www. raoso ft. com/ sampl esize. html) at a 95% confidence level and alpha set at 0.05. Based on the primary amoxicillin resistance rate of 15% [3] in H. pylori in Vietnam, the sample size was estimated at 195 for amoxicillin resistance study. We also set up a larger sample size at the same time for studying other molecular antimicrobial resistance in H. pylori as well. Therefore, the sample size of 308 was finally used for the analysis of amoxicillin resistance in H. pylori as the result of the available data. During upper gastrointestinal endoscopy, four biopsies were obtained from each of the 308 patients recruited to the study to make sure the highest rate of H. pylori recovered from the sampling and to limit the influence of H. pylori distribution in the gastric mucosa. The first pair of specimens (one from the antrum and one from the corpus) was tested using both the rapid urease test (NK-Pylori test, Nam Khoa Biotek Co., Ltd.) and a polymerase chain reaction-based test (PCR) (AccuPid H. pylori Genotyping Kit, ref. No. Q01HPY03.1A, Khoa Thuong Biotech Co., Ltd.). The second pair was stored in transport medium [17] and kept at 2-8 °C until processed for culture within 3 h after sampling. Samples from a total of 101 patients were collected for culture. The samples can also be stored at − 80 °C in BHI (BD) supplemented with 25% glycerol until culture. The Pbp1A gene fragment from all samples was subjected to direct sequencing. The patient characteristics are presented in Table 1. samples were then ground in 100 μL brain heart infusion (BHI, BD) broth supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich) and cultured on Columbia agar (BD) plates supplemented with 10% lysed sheep blood (Nam Khoa Biotek Co., Ltd.), 1% isoVitaleX (BD), a skin antibiotic mixture (Sigma-Aldrich), and 2.5 μg/ mL amphotericin B (Sigma-Aldrich). Plates were incubated for 4-5 days at 37 °C in a microaerobic atmosphere. A single colony of a 4-5-day-old culture was identified through morphological observation, urease tests and PCR specific to the cagA and vacA genes. A colony with specific morphology, a positive urease test and identified cagA and vacA was subcultured on Columbia agar plates as described above for 4 days. Amplification and detection of the cagA and vacA genes was performed based on a previously published protocol [18].

Determination of the minimal inhibitory concentration
The minimal inhibitory concentration (MIC) of amoxicillin was determined by the E-test method (BioMerieux, ref.# 412243). H. pylori strains stored at − 80 °C were thawed on Columbia agar plates supplemented with 10% lysed sheep blood and 1% isoVitaleX and incubated at 37 °C for 3-4 days in a microaerobic atmosphere. Colonies were subcultured on another plate for 3 more days. Each plate was swabbed with a sterile cotton-tipped applicator, and the cells were suspended in BHI to obtain a turbidity equivalent to a 3.0 McFarland standard. A 100 μl bacterial suspension was inoculated on a Columbia plate containing 10% lysed sheep blood and 1% isoVi-taleX, and another sterile cotton-tipped swab was used to streak in three directions across the plate. The plates were left to dry for 2-3 min, and then the E-test strips containing a continuous exponential gradient of amoxicillin were placed on the agar surface. The plates were incubated at 37 °C for 3-4 days in a microaerobic atmosphere. At the end of the incubation period, the MICs were determined by the intercept of the zones of growth inhibition with the graded E-test strip. The reference H. pylori strain ATCC ® 43504 ™ susceptible to amoxicillin (Amx S ) was always used as a quality control strain, and MIC determination was accepted only if the MIC of amoxicillin for this strain was less than or equal to 0.125 μg/ml. The susceptibility of the H. pylori isolates to amoxicillin was classified as resistant if the MIC was greater than 0.125 μg/ ml, according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints (https:// eucast. org/ clini cal_ break points/).

Direct sequencing of the pbp1A gene fragment
Total genomic DNA was extracted from the gastric biopsy specimens using a Qiacube automated purification system and kit (QIAGEN) according to the manufacturer's standard instructions. Optimisation of the real-time PCR components and conditions was performed to determine the optimal conditions for the amplification of a specific fragment of the pbp1A gene of the bacteria in an automated thermal cycler (Roto-Gene Q, Qiagen). Sequences of the pbp1A primers were designed in this study as follows: pbp1A-F: 5′-CGA TAG ATT TGG ATT ACC AACGC-3′; pbp1A-R: 5′-ACG ATT TCT TTA CGC AAG CC-3′. The size of the expected amplicon was 1035 bp. The pbp1A-F/R The PCR products were purified with ethanol. The purified products were sequenced using a BigDye ™ Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA). The sequencing PCR products were purified with the BigDye XTerminator ™ and read in an ABI 3130 Genetic Analyser. Nucleotide sequences of both chains obtained were aligned and transformed into amino acid sequences using Geneious Prime software version 2021.1.1 (Auckland, New Zealand) and compared to the sequences of strains 26,695 (AE000511) and Hargenberg (AF479617) deposited in GenBank (http:// www. ncbi. nlm. nih. gov/ Genba nk/) for detection of mutations in the pbp1A gene.

Statistical analysis
Statistical analysis was performed using the Statistical Package for Social Science (SPSS) version 20.0. Descriptive statistical analysis was used to describe the characteristics of the patients' gender, age, residency, gastric disease status and susceptibility to amoxicillin of the strains isolated from the clinical samples. A one-sample binomial test was used to determine whether the proportion of cases (sex, amoxicillin resistance, coinfection) was equal to the previously documented corresponding proportion. The chisquare test was used to correlate the presence of mutations in the pbp1A gene and the susceptibility to amoxicillin. Fisher's exact test was used alternatively when more than 20% of the expected counts were less than 5. Monte Carlo estimates of the exact significance were used when the data did not meet the assumption of the asymptotic method. A p-value less than 0.05 was considered significant.

Demographic characteristics
There were 308 qualified patients recruited in this study. Details on the demographic characteristics of these patients are presented in Table 1.
Amoxicillin MIC was determined randomly in 101 samples (32.8% of the entire study group). Among these samples, 74.3% (75/101) of strains were sensitive, and 25.7% (26/101) of strains were primarily resistant to amoxicillin. The isolates with phenotypic resistance to amoxicillin exhibited a MIC range of 0.190-1.5 mg/l ( Fig. 1).

Virulence factors of H. pylori
The presence of the cagA or vacA gene was investigated in all 308 clinical specimens positive for the rapid urease test. Based on the vacA genotypes, coinfection was discovered: 38 (12.3%) specimens were colonised by at least two H. pylori strains (vacA m1 and vacA m2). All of the coinfection specimens were excluded from the analysis to assess the association between bacterial genotypes and clinical outcome as well as antibiotic resistance.
The cagA gene was present in 225 (83.3%) of the 270 H. pylori specimens. There was no association between gastritis and peptic ulcers with the presence or absence of the cagA gene (p = 1.000).
Regarding the vacA genotypes, most of the specimens had the vacA s1 allele (98.4%). No association was detected between the vacA signal sequences and the clinical outcome (p = 0.265). However, there was a relationship between the vacA middle sequences and the clinical outcome (p = 0.021). The vacA m1 allele was found to be significantly higher in 12 (75.0%) specimens from gastric ulcer patients, and the vacA m2 allele was significantly higher in 141 (55.5%) of the specimens from gastritis patients. The combination of the vacA signal and middle sequences with the clinical outcome also presented a similar significant relationship (p = 0.008). VacA s1m1 was higher in the ulcer group (75.0%), while vacA s1m2 was higher in the gastritis group (53.9%). The combination of cagA status and vacA genotypes was also investigated. There was an association between the combined genotypes and the clinical outcome (p = 0.032). In particular, no association was observed between the genotypes and the resistance to amoxicillin (p > 0.05) ( Table 2).

Mutational changes in or adjacent to the acyl transpeptidase conserved sequence of the pbp1A gene
Amoxicillin resistance in Amx R H. pylori isolates is mediated by mutations in the pbp1A gene [13]. To investigate the genetic diversity of pbp1A genes in this study by gene sequencing, the pbp1A gene sequences of the H. pylori reference strain 26,695 (O25319, Amx S ) and Hargenberg (AF479617, Amx R ) were used as controls to compare and number the isolates according to the corresponding deduced amino acid sequences. A total of 270 sequences of pbp1A gene fragments spanning amino acids 310-596 were analysed. The results showed that the proportion  of amino acid substitutions varied from 0.4 to 100% depending on the amino acid position (Fig. 2). Several amino acid positions in the acyl transpeptidase domain of PBP1A had a high proportion of substitutions, such as Asp 535 to Asn (100%), Ser 589 to Gly (96.7%), Asp 479 to Glu (86.6%), and Asn 504 to Asp (74.1%). Amino acid positions with substitution proportions equal to or lower than 1% are not shown on the graph (Fig. 2).

Association between amino acid substitution and amoxicillin susceptibility
To investigate which pbp1A gene mutations were involved in amoxicillin resistance in the Amox R isolates in this study, we evaluated all 77 amino acid substitutions in the acyl transpeptidase domain of PBP-1A, from amino acids 310 to 596. Seven amino acid changes were found to be significantly different between the Amox S and Amox R samples (  [11,13,15]. We noticed that the previously described amino acid changes, such as Glu 406 to Ala (x 2 = 0.017; p = 1.000), Ser 417 to Thr (x 2 = 0.291; p = 1.000), Thr 556 to Ser (x 2 = 0.895; p = 1.000), and Asn 562 to Tyr (x 2 = 4.338; p = 0.072), associated with a high level of beta-lactam resistance in acquired multidrug-resisting H. pylori [19] were insignificant in this population (Table 3). Additionally, other novel amino acid substitutions mentioned previously [13] in selected Amx R transformants comprising Thr 540 to Ile, Ser 542 to Arg, Thr 555 to Ser, and Asn 561 to Tyr were unchanged in this study. The results showed that several mutated positions had statistical significance between amoxicillin-sensitive and amoxicillin-resistant samples and were completely different from amino acid positions reported to be related to amoxicillin resistance published previously [15]. These mutations are adjacent to the second and third PBP motifs, which were observed only in Amx R H. pylori colonies after transformation [13].

Phylogenetic analyses of the pbp1A sequences
Phylogenetic analyses were performed on all pbp1A gene fragment sequences to determine whether the sequences with identified mutations were in the same group of sequences related to amoxicillin resistance. Trees were constructed from 85 sequences based on the pbp1A amplification fragments trimmed to 879 bp by the Tamura-Nei genetic distance model and the neighbourjoining tree building method (Geneious Prime ® 2021.1.1).

Fig. 2
Amino acid substitution proportion of deduced PBP-1A discovered in the population compared to the Amx S H. pylori strain 26,695. Red represents amino acid substitution differences with statistical significance from Amx S and Amx R samples by the E-test. A proportion lower than 1% was not included in the picture This primary phylogenetic analysis suggested that the pbp1A sequences obtained in the present study might belong to different sequence groups. The sequences from which H. pylori isolates showed resistance to amoxicillin could be combined into two groups with close relatedness to those from which H. pylori isolates were sensitive to amoxicillin, except for samples A003, A075, A138 and A163. These two groups (a, b) also shared the most recent common ancestor of the pbp1A gene (Fig. 3). Group (c) including sensitive strains was also created as a control group, along with 26,695 and Hagenberg strains. Table 4 shows that groups (a) and (b) had very similar amino acid substitution and insertion profiles and were quite different from the reference sequences at the seven observed positions. There were four novel strains from our samples presenting single amino acid substitutions at Leu 366 and Val 473 , as well as amino acid insertions at Glu/ Asn 464-465 and Ser/Ala/Gly 595-596 . The other two amino acid replacements (Met 469 and Glu 479 ) also appeared uniquely in our data when compared to the references. However, the presence of specific substitutions and insertions was present not only in the resistant samples but also in the sensitive samples. For example, Met 469 appeared in just two sensitive strains (MIC values were equal to 0.016 and 0.023 μg/mL). Leu 366 and Val 473 also appeared in sensitive strains, but not the sensitive ones with MIC ≤0.016 μg/mL.

Discussion
Amoxicillin resistance in H. pylori is one of the greatest concerns of clinicians because anti-H. pylori regimens often consist of amoxicillin in addition to other antibiotics and proton pump inhibitors [20,21]. Recent studies have suggested that amoxicillin resistance in H. pylori results from alterations in PBP1A [12,13,15,22]. One of the main purposes of this study was to investigate the molecular mechanism of amoxicillin resistance in H. pylori strains collected from endoscopic biopsies in Vietnam.
The amoxicillin-resistant H. pylori proportion discovered in this study (25. [28,29]. Compared to previous studies in Vietnam, the amoxicillin resistance proportion has been trending up significantly, for instance, from 0% (p < 0.001) in 2013 [30], 1.1% (p < 0.001) in 2015 [8], 10.4% (p < 0.001) in 2016 [9], 15% (p = 0.002) in 2019 [3] and to 25.7% in 2020 for this study. Although the previous local studies were different from each other regarding geographical areas, sample sizes, study period and antimicrobial testing methods, an increasing proportion of amoxicillin resistance in H. pylori in Vietnam has been generally demonstrated, suggesting that this is a serious emerging threat to the success of amoxicillinbased regimens. The high rate of amoxicillin-resisting H. pylori in our study might be explained by the fact that the combination of amoxicillin and clavulanate potassium is often empirically prescribed for various infectious diseases [31]. The incorrect use of antimicrobials can accelerate the selection of drug-resistant strains [23].
Multiple strains of H. pylori can coinfect the same patient [32]. The coinfection detected by vacA genotype in our study presented a rate of 12.3%. A similar proportion in terms of coinfection determination was obtained when confirmed by random amplified polymorphic DNA (RAPD) fingerprinting (12.5%; p = 0.500) [32] or through vacA and iceA genotyping (11.0%; p = 0.255) [33]. Therefore, coinfection needs to be excluded before evaluating the relationship between H. pylori genotypes and other factors, such as disease status and clinical symptoms, to maintain accuracy. Moreover, coinfection could undermine the success of eradication therapy and should be considered when interpreting the results of antimicrobial susceptibility tests [32].
Based on the H. pylori genotype analysis, our data showed an association between the vacA m1 genotype and gastric ulcers. This result suggests that individuals colonised with vacA m1 -positive H. pylori strains are at an increased risk of developing gastric ulcers. These results confirmed the vacA m1 genotype is associated with an increased risk of peptic ulcers, which has been reported by Nguyen et al. [34] and Trang et al. [35] in Vietnam. Conversely, Milad et al. revealed that the vacA m2 genotype was significantly higher in patients with peptic ulcer disease than in patients with gastritis in Iran [36]. On the other hand, several previous studies published by Godoy et al. in Brazil and Loivukene et al. in Estonia did not find any association between virulence factors such as vacA m1 genotype and clinical outcomes or bacterial resistance to metronidazole, although the coinfection by multiple strains has been well considered in these studies [33,37]. The difference in cagA and vacA genotype proportions as well as the association of these genotypes with the clinical outcomes might result from the fact that various populations have been evaluated with different ethnic groups and that a high genetic variability of strains in different countries exists [38,39]. However, there is universal agreement regarding the role of the vacA m1 genotype among studies carried out in Vietnam at different time points. In addition, we did not find any association between the virulence factors and resistance to amoxicillin.
Amino acid substitutions in the acyl transpeptidase domain of PBP1A are required for resistance to amoxicillin [12,13], especially F473L alteration recently discovered to be the important genetic determinant of resistance to amoxicillin of H. pylori in Cambodia [40]. However, other mechanisms could be involved in increasing the MIC value and contributing to the levels of high Amx R strains, such as decreasing membrane permeability due to altered porin proteins (HopC, HopH), increasing the activity of efflux pumps to eject antibiotics from the periplasm, or even reducing the binding of antibiotics to other PBPs, especially PBP2 [19,41].
In this study, we discovered a high rate of resistance to amoxicillin. To identify the mutations of pbp1A responsible for amoxicillin resistance, sequences of the pbp1A gene were analysed. We found seven amino acid changes possibly linked to amoxicillin resistance in clinical samples. Among them, the Ser 414 to Arg substitution has been proven to be the main factor in amoxicillin resistance of the Hardenberg strain by site-directed mutagenesis [12] and it was also common in clinical Amx R strains by natural transformation [11,13]. The Phe 366 to Leu alteration was reported to be present in the clinical Amx R strain SZ79 in combination with the Ser 414 to Arg substitution [13]. Phe 473 was recognised in strains sensitive to but less susceptible to amoxicillin but also in Amx R transformants in the absence of Ser 414 to Arg substitutions [15], while Val 473 was common in our Amx R samples in the context of Ser 414 to Arg changes. Although it has become clear that amino acid variations conferring resistance vary by the geographical origins of the strains [15], these data have confirmed the combination of amino acid substitutions or mutations in multiple loci to amoxicillin resistance [13].
To investigate the relatedness, as well as the combination of different mutations of pbp1A sequences in the resistance to amoxicillin, we created a phylogenetic tree of 85 obtained pbp1A gene fragments. The grouping of the pbp1A gene sequences suggested that there were other mechanisms in addition to the mutations in the pbp1A gene contributing to amoxicillin resistance in H. pylori in Vietnam. Other studies have also shown that distinct mechanisms of antimicrobial resistance also play important roles in the resistance to amoxicillin in H. pylori [13].
Our study had several limitations. First, resistant mutations were not identified directly from isolates but only from gastric biopsy specimens. We excluded coinfection cases from the data analysis, and this approach might not be ideal for identifying the molecular mechanisms related to amoxicillin resistance. Second, the role of other genes that could have synergistic effects in amoxicillin resistance could not be excluded. We did not evaluate coinfection by multiple H. pylori strains by fingerprinting methods such as random amplified polymorphic DNA (RAPD) or multilocus sequence typing (MLST). Last but not least, resistance has not been confirmed to have a direct correlation with eradication efficacy in real-life practice.
However, the strength of this study is that it was conducted at one of the largest hospitals in southern Vietnam, which usually admits patients from Ho Chi Minh City and many nearby areas. This is the first study that reported mutations related to Amx R H. pylori in Vietnamese patients and it has identified some novel mutations, especially insertion mutations Glu/Asn 464-465 and Ser/Ala/Gly 595-596 in the pbp1A gene and other nearby mutations, which are likely specific to H. pylori strains in Vietnamese. Further studies are required to validate the role of these novel mutations in conferring amoxicillin resistance. In addition to the emerging prevalence of amoxicillin-resistant H. pylori strains, direct detection of pbp1A gene mutations from H. pylori-positive biopsy specimens may lead to novel diagnostic strategies for amoxicillin resistance determination and would be useful in clinical practice. More importantly, other mechanisms, such as the acquisition or expression of β-lactamase or changes in other proteins involved in cell wall synthesis, such as PBP2, PBP3, and PBP4, should also be evaluated.

Conclusion
Our study has identified new mutations that are statistically associated with amoxicillin resistance and demonstrated the importance of the detection of amoxicillin-resistant Helicobacter pylori in clinical practice because of the emergence of these strains in Vietnam. However, further studies should be carried out to identify additional mechanisms contributing to amoxicillin resistance in H. pylori.