Impact of Helicobacter pylori eradication therapy on gastric microbiome 1

13 Background: Helicobacter pylori (Hp) eradication therapy has been used in clinical 14 practice for many years. Yet, the effect of this therapy on existing gastric microflora has not 15 been well understood. In this study, we explored the effect of eradication therapy on the 16 microbial community in the stomach and the specific recovery after the successful 17 eradication therapy. 18 Methods: Among the 89 included patients, 23, 17, 40, and 9 were enrolled into the Hp- 19 negative, Hp-positive, Successful eradication, and Failed eradication groups, respectively. 20 Four subgroups were further divided according to disease status (Hp-negative chronic 21 gastritis [N-CG], Hp-negative atrophic gastritis [N-AG], successful-eradication chronic 22 gastritis [SE-CG], and atrophic gastritis with successful eradication [SE-AG]). During the 23 endoscopic examination, one piece of gastric mucosa tissue was obtained from the lesser 24 curvature side of the gastric antrum and gastric corpus, respectively. 16S rRNA gene 25 sequencing was used to analyze the gastric mucosal microbiome. Results: In Hp negative group, the gastric microbiota was dominated by five phyla: 27 Firmicutes , Proteobacteria , Actinobacteria , Bacteroidetes, and Fusobacteria . Two Hp- 28 related genera were selected as potential biomarkers: Curvibacter and Acinetobacter . After 29 successfully eradicating Hp, the bacterial flora in the stomach recovered to a considerable 30 extent, and the failure of eradication was almost unchanged compared with Hp positive 31 subjects. SE-CG was characterized by an increase in Firmicutes taxa and a decrease in 32 Proteobacteria taxa compared with N-CG. SE-AG was characterized by a decrease in 33 Firmicutes relative to N-AG. Finally, no differences were found in pairwise comparisons of 34 nitrate and nitrite reductase functions among the four subgroups. Conclusions: After Hp infection, the diversity and relative abundance of gastric microflora were significantly decreased. Yet, gastric microbiota could be partially restored to the Hp- negative status after eradication; however, this effect was incomplete and might contribute 38 to the long-term risks.


48
The gastrointestinal microecological balance has an important role in digestion, absorption, 49 metabolism, immunity, and inhibition of pathogen colonization. The disorder of its structure 50 or function can lead to many diseases. Helicobacter pylori (Hp) is the most important and 51 most studied microorganism in the stomach. It has been associated with various 52 gastrointestinal diseases, such as chronic gastritis, peptic ulcer, gastric mucosa-associated 53 lymphoid tissue lymphoma, and gastric cancer. Numerous clinical studies have shown that 54 Hp eradication reduces the incidence of gastric cancer, and this benefit becomes more 55 pronounced with increasing age [1][2][3]. Currently, eradication therapy has been used in many 56 regions to prevent the development of stomach cancer [4,5]. 57 The core of Hp eradication treatment is the acid-suppressive effect of PPIs and the 58 bactericidal effect of antibiotics. Antibiotics have a direct and strong effect on the bacteria 59 in the stomach [6]. The strong acid inhibitory effect of PPIs can sharply increase the 60 stomach's pH value, thereby reducing gastric acid's effect on the removal of guest bacteria, 61 which is not conducive to digestion and leads to various changes in substrate levels [7,8]. 62 Combined with existing research, the drug itself and Hp's elimination have a potential effect 63 on the gastric flora [9,10]. 64 Through molecular methods, the main phyla detected in the stomach are Proteobacteria, 65 Firmicutes, Bacteroidetes, Actinobacteria, and Fusobacteria. The most abundant phyla after 66 Hp infection in the stomach are Proteobacteria, Firmicutes, and Actinobacteria [11]. The 67 gastric cancer model study of INS-GAS mice showed that the non-Hp flora could promote 68 the occurrence of tumors [12,13]. Other evidence in clinical research revealed that microbial 69 diversity changes with the health of the gastric mucosal epithelium [14,15]. Researches on 70 the gastric cancer flora found that Niche-specific microbial networks may reflect the 71 disease-specific microenvironment, and disease-associated bacteria can form a cooperative 72 network, which additionally contributes to the disease [16,17]. In addition, previous 73 research has mainly focused on the gastric microbiome of patients with gastric cancers rather 74 than precancerous lesions such as gastritis atrophy (AG). In this study, we analyzed the 75 influence of eradication treatment on gastric flora and evaluated patients' recovery with 76 successful eradication under different mucosal states. one piece of gastric mucosa tissue was obtained from the antrum's lesser curvature side and 85 another piece from the lesser curvature side of the corpus. Each specimen was placed in a 86 separate sterile cryopreservation tube. Another mucosal tissue of the gastric antrum was used 87 for histological biopsy to assess gastric mucosa and Hp infection status.

88
Exclusion criteria were as follows: (1) patients who have taken proton pump inhibitors,

89
H2 receptor antagonists or other antacids, probiotics, mucosal protective agents or 90 antibiotics within recent four weeks; (2) history of gastric adenoma, gastric cancer, or 91 mucosa-associated lymphoid tissue lymphoma; (3) patients who underwent gastrectomy; (4) 92 patients who underwent Hp eradication therapy and were again Hp positive.

93
Current Hp infection was defined as a positive result from one of the following three 94 tests: (1) C-urea breath test, (2) histologic examination, (3) Hp culture. Furthermore, 95 according to some previous studies, those samples with < 1% of Hp relative abundance were 96 excluded from the analysis to obtain higher representativeness [16,18]. For patients with a 97 history of eradication, we selected those with a completion time of one year. We combined 98 past and current Hp infection status to confirm eradication situation. Only those whose 99 gastric mucosa status was judged by endoscopy, which was further confirmed by 100 pathological biopsy results, were classified into subgroups.   Omega, Inc., USA) according to manufacturer 's instructions. Polymerase chain reaction 124 (PCR) was used to amplify the 16S rRNA V3-V4 region, using the 341F 5'-125 CCTACGGGNGGCWGCAG-3' and 805R 5'-GACTACHVGGGTATCTAATCC-3' primers.

126
The 5' ends of the primers were tagged with specific barcodes per sample and sequencing

154
A linear discriminant analysis (LDA) and effect size ( To evaluate alterations in the microbiota structure between the gastric antrum and corpus, 171 we measured microbial alpha diversity and beta diversity. Alpha diversity showed a high 172 degree of similarity ( Figure 1A). Beta diversity revealed no significant differences between

197
To identify the potential biomarkers with Hp infection, we conducted LEfSe analysis 198 between the paired groups (group N versus group P, group SE versus group FE). We selected 199 the common bacteria at different levels through the Wilcoxon rank-sum test. Shannon index were found in SE patients ( Figure 1B). No obvious differences were found 218 between P and FE patients. Beta diversity analysis revealed no remarkable differences in 219 microbial diversity between N and SE patients (ANOSIM R = -0.0269, P = 0.871, Figure   220 2B), as well as P and FE patients (ANOSIM R = 0.0537, P = 0.132, Figure 2C). When 221 enrolling four groups together, we distinguished two pairs of groups from sample 222 distribution (ANOSIM R = 0.3127, P = 0.001, Figure 2D). As the dominant phyla, the   In order to further explore the differences between CG and AG, we next analyzed four 257 subgroups (N-CG, N-AG, SE-CG, and SE-AG). We found that N-CG had greater richness 258 and diversity than the other three subgroups (except for Simpson's index comparing N-CG 259 and SE-CG, P=0.06), and there were no differences between the three subgroups ( Figure   260 1C). The sum of the relative abundance of Firmicutes and Proteobacteria exceeded 50% in 261 each subgroup. In group N, gastric mucosal atrophy showed an increase in Firmicutes 262 (P<0.001, Figure 4B) and was accompanied by a relative decrease in Proteobacteria 263 (without statistical significance, Figure 4B). However, the same trend was not observed in 264 group SE (P=0.528, P=0.430, Figure 4B). Elevated levels of Firmicutes taxa: genera 265 unclassified Lachnospiraceae and Romboutsia were detected in SE-CG patient samples, 266 whereas Proteobacteria and Acidobacteria taxa were depleted in these samples ( Figure 5B). indicating that corresponding changes occurred at lower taxonomic levels as well.  Figure 6B). Additionally, we used correlation heatmaps to investigate the 282 association between differential genera and KEGG pathways. Genera unclassified 283 Alphaproteobacteria was positively correlated with cell motility, while genera unclassified 284 Lachnospiraceae was negatively correlated with cell motility (Figure 6A-B). Interestingly, 285 except for four genera (Bifidobacterium,Bacillus,unclassified Aminicenantales,and 286 Rhodococcus), all negatively correlated genera are of the phylum Firmicutes, while all 287 positively correlated genera are of the phylum Proteobacteria (Figure 6A-B). 288 289 Figure 6. Associations of microbiota with predicted KEGG functions evaluated by Spearman 290 correlation coefficients between 33 genera and differential KEGG pathways in N-CG versus N-AG 291

Kyoto encyclopedia of genes and genomes. 293
The pathological changes from chronic gastritis, precancerous lesions to gastric cancer 294 is a long process. Other non-Hp bacteria with specific functions are likely to be involved.

295
The existing hypothesis is that nitrate-reducing bacterial species are associated with 296 increased risks in gastric carcinoma [10]. Eradication therapy may have a potential impact 297 on this function by changing the flora in the stomach. Therefore, we evaluated four 298 subgroups and compared the results. Pairwise comparisons revealed that all nitrate and 299 nitrite reductase functions had no significant differences ( Figure S1).

302
In the present study, we identified differential bacterial taxa and metagenomics functions 303 before and after successful Hp's eradication. Based on our results, the bacterial composition 304 between the paired gastric antrum and corpus was highly similar, which was consistent with 305 previous results [21][22][23]. 306 Previous research suggested that Hp had the greatest impact on gastric bacterial 307 composition and diversity [21]. In this study, we found that the abundance and diversity of 308 bacteria after Hp colonization were significantly lower than those of non-infected Acinetobacter that might be associated with Hp infection still belong to Proteobacteria. 314 Curvibacter, which is a common part of oral microflora, is prevalent in patients with 315 atherosclerotic plaques [25]. Earlier studies reported that atrophic gastritis was 316 accompanied by a reduction in Hp colonization [26,27].  that the OmpA-like protein from Acinetobacter spp. could stimulate the production of 318 gastrin and IL-8 cytokine, which suggests they can cause gastritis or participate in the 319 transformation towards atrophic gastritis [28]. Interestingly, Acinetobacter spp. might have 320 a pathogenic role when Hp is decreased, which suggests they are not only participants but 321 also activating factors.

322
The richness, diversity, and structure of the bacterial communities in the FE and P 323 samples were highly similar. These results showed that once Hp abnormal occupied the 324 stomach, the gastric flora was difficult to be disturbed. It also implied that Hp eradication 325 drugs had a little long-term impact on the gastric flora. After successful eradication of Hp, 326 the phylum and genus composition of the gastric flora could be restored to levels close to 327 those of Hp-negative subjects, and the bacterial diversity index increased, which was 328 consistent with previous reports [15,21]. However, there was still a significant difference 329 between the SE and N groups, revealing an outcome of limited recovery. We assumed these 330 differences might due to some irreversible changes after Hp colonization. To confirm this, 331 we further evaluated whether gastric mucosal atrophy could affect the intragastric flora 332 through four subgroups (N-CG, N-AG, SE-CG, SE-AG). The results showed that the 333 richness and diversity of Hp-negative CG patients were significantly higher than those of 334 the other subgroups, and no significant differences were observed among them (N-AG, SE-335 CG, and SE-AG). This further confirmed our hypothesis that regardless of gastric mucosal 336 atrophy development, the gastric flora of patients with successful eradication was closer to seem to be at a higher risk than the normal population.

347
Recent studies showed that in gastric carcinoma microbiota, increased nitrate reductase 348 and nitrite reductase functions were considered as drivers of cancer development [17,31,349 32]. To assess this risk, we next addressed the functional features of the microbiota. However, 350 our results did not reveal this trend in atrophy patients. We speculate that the risk of 351 dysbacteriosis in this regard is relatively low due to the successful elimination of Hp.

352
To sum up, this article focused on the impact of eradication on the flora and assessment 353 of the recovery of patients with successful eradication. Moreover, we described the effects 354 of gastric mucosal atrophy on changes in gastric microbiota. Our study verified that in the 355 presence of Hp, the gastric flora was quite stable and, therefore, difficult to alter by 356 antibiotics and highly effective acid suppressants. Hp is the initiating factor and a key link 357 in Correa's cascade [10]. Moreover, even when advanced precancerous lesions occur, the 358 successful removal of Hp is of great importance, especially in East Asia [33]. It seems that 359 the risk of gastric cancer in patients with successful eradication has been greatly reduced, 360 which has been confirmed by large-scale clinical research [34]. However, for those who 361 already experienced precancerous lesions such as atrophy, the risk is still higher than in the 362 normal aging stomach [35]. Consistent with this, our study reported that people who 363 successfully eradicated Hp were closer to those with Hp-negative gastric mucosal atrophy, 364 which represented a smaller bacterial community. Interestingly, this change may not have a 365 profound impact.

366
This study has a few limitations. First, this was a single-center cross-sectional study 367 with small sample size, especially considering those enrolled in group P and group FE.

368
However, in this study, we implemented strict screening criteria and eliminated samples with 369 < 1% of the Hp sequence in order to obtain higher representativeness. Secondly, we did not 370 obtain mucosal samples from the same subject before and after Hp eradication treatment to 371 achieve self-control. Third, the bacterial community is continuous and dynamic, so it was 372 not possible to determine the causal relationship between these changes and different states.

373
In addition, this study did not use PCR quantification techniques to quantify individual 374 bacteria in different samples; thus, the analysis could only be based on the relative 375 abundance of different bacteria. Therefore, further studies are still needed to verify and 376 clarify the influence of eradication therapy and precancerous lesions on the gastric flora.

379
Hp eradication drug itself has little effect on gastric flora. After successful eradication, the 380 gastric microenvironment could be partially restored to levels close to Hp-negative gastric 381 mucosal atrophy patients. We also identified some biomarkers that may synergize with Hp 382 to promote disease progression. The specific mechanisms and pathways underlying these 383 changes will be explored in future research.