Helicobacter pylori Infection in Infant Rhesus Macaque Monkeys is Associated with an Altered Lung and Oral Microbiome

of the


Background
Helicobacter pylori is a common cause of chronic gastritis in humans and can elicit a range of disease phenotypes depending upon a range of host, bacterial virulence, and environmental factors (reviewed in [1]).It is estimated that half of the world's population is infected with H. pylori, with prevalence varying by geographical location [2].In the United States, a recent survey of veterans indicated a positive diagnosis of H. pylori infection in over 25% of patients [3].Although colonization of the stomach imposes an increased risk of developing peptic ulcers (10-20%) and gastric cancer (1-2%), the majority of human subjects infected with H. pylori are asymptomatic [4].H. pylori is primarily acquired in early childhood and unless the infected individual is treated with antibiotics, colonization persists for life [5][6][7][8].Transmission of H. pylori is most likely through a buccal-gastro route, as the organism has been detected in both vomitus and buccal aspirates of infected children [9,10].
When H. pylori is present in the stomach, it dominates the gastric microbial community, decreasing overall taxonomic diversity and comprising up to 97% of reads in the culture-independent analysis [11].
As such, characterization of the human gastric microbiome has revealed two apparent phenotypes: one colonized with the gram-negative, microaerophilic bacteria H. pylori and one without.Experimental H. pylori infection also alters the community structure and microbial abundance at distant sites within the gastrointestinal tract in a murine model, although eradication treatment in humans has been reported to be successful in reverting to baseline alpha and beta diversity [12,13].To date, it is unknown whether gastric H. pylori colonization can in uence the lung microbiome on a short-term or persistent basis, but microaspiration may be a potential pathway for the direct introduction of the pathogen to the lung; H. pylori and associated virulence factors have been identi ed in the human adult lung [14][15][16].
While bacteria pathogens can be aspirated into the lung, it remains unclear how the commensal population of the lung can be altered by the host beyond antibiotic treatments.Several reports have suggested a link between the community composition of the lung microbiome and respiratory diseases, including asthma, COPD, and cystic brosis [14,[17][18][19][20].However, many chronic diseases of the airways in adults are thought to be established in early childhood, which suggests a more thorough understanding of the development of the normal pediatric lung microbiome is critical [21].A potential immunomodulatory role for H. pylori in the context of the gut-lung axis has been suggested by epidemiological ndings in human populations demonstrating an inverse association between the development of childhood asthma and gastric colonization with H. pylori [22,23].Correspondingly, introduction of H. pylori in a neonatal murine model has shown the induction of T regulatory cells via IL-18-producing dendritic cells is essential for immunosuppression of allergic airways disease development [24,25].
Human microbiome studies have provided support for the notion that commensal bacteria are essential for maintaining physiological homeostasis and in uencing disease development [26].To determine whether the presence of gastric H. pylori infection can alter the composition of the lung and oral microbiome during infancy, we used the rhesus macaque monkey as a model of early childhood development.Rhesus monkeys in captive environments become naturally infected with an H. pylori strain virtually indistinguishable from strains identi ed in humans, and socially housed animals acquire H.pylori within the rst year of life [27,28].Here, we assessed the relationship between the gastric H. pylori infection status, lung microbiome, and oral microbiome in infant monkeys.We also evaluated whether the expression of IL-8 protein as a mucosal or systemic biomarker of H. pylori colonization is associated with microbial communities in either the lung or oral cavity.

Animals
Infant rhesus macaque (Macaca mulatta) monkeys aged 4-7 months (n = 25) were born at the California National Primate Research Center (CNPRC).All animals were socially housed, reared, and breastfed by dams indoors in an environment-controlled facility with an ambient temperature of 21-25°C, relative humidity of 40-60%, and a 12h light/dark cycle.All animals were provided water and commercial monkey chow ad libitum.Before, during, and after treatment, care and housing of animals complied with the Institute of Laboratory Animal Resources provisions and conforms to practices established by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC International).All animal procedures were approved by the University of California, Davis, Institutional Animal Care and Use Committee.All possible efforts were made to minimize pain and distress following the recommendations of the National Research Council publication, Guide for the Care and Use of Laboratory Animals, Eighth Edition.

Biospecimen collection
Animals were fasted overnight before collection of blood, lung wash, buccal swab, and gastric biopsy.
Animals were sedated with ketamine anesthesia (10 mg/kg) administered intramuscularly during sample collection.To avoid injury to the infant monkey respiratory tract, lung washes were obtained using a sterile single-use endotracheal tube and instillation of sterile PBS.PBS was instilled at a volume of 2 ml per kilogram body weight.A swab of the buccal cavity was collected prior to the lung wash, along with a blood sample for plasma collection.Endoscopy was conducted using a pediatric bronchoscope to collect gastric corpus biopsies for H. pylori culture as previously described [28].Lung wash uids, buccal swabs, and plasma was stored at -80°C until used for analysis.

H. pylori culture
Gastric biopsy specimens were placed in vials containing Brucella broth and transported to the laboratory for immediate processing as previously described [28].All plates were incubated in an atmosphere of 5% O 2 , 7.6% CO 2 for up to 8 days.H. pylori was identi ed conventionally by colony morphology (pinheadsized translucent colonies), microscopy (gram-negative curved organisms), and biochemistry (oxidase, catalase, and urease positive).Colony forming units per gram of tissue specimen were calculated by enumerating colonies, adjusting for the dilution, and dividing by the tissue weight.
Cytokine ELISA IL-8 protein concentration in lung wash uid and plasma was measured by ELISA DuoSet kit (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions.The limit of detection for the IL-8 ELISA was 7.5 pg/mL.

Demographics of animals evaluated for lung and oral microbiome composition
Demographics of animals evaluated by 16S rRNA gene sequencing analysis are listed in Supplemental Table 1.Of the 25 infant rhesus macaques (17 females, 8 males) initially screened for H. pylori infection, a total of 14 animals were selected for 16S rRNA gene sequencing analysis (6-7 months old, 13 females, 1 male).Of the 14 animals evaluated for microbiome composition, 8 were H. pylori-negative and 6 were H. pylori-positive by gastric biopsy.Microbiome analysis of buccal swabs was limited to 12 out of 14 animals evaluated.Animals had a median weight within 0.1 kg and a median age within 0.8 months of each other.

16S rRNA gene amplicon sequencing library preparation
In brief, according to the manufacturer's instructions, DNA was isolated using a PowerSoil DNA Isolation Kit (MoBio, Carlsbad, CA).A nested PCR was used to amplify and barcode the V4 domain of the 16S ribosomal RNA (rRNA) gene.The rst PCR was conducted using 1µL of DNA template and primers 27F (AGAGTTTGATCMTGGCTCAG) and 1492R (GGTTACCTTGTTACGACTT) were used to amplify a large portion of the 16S rRNA gene.The second PCR was conducted using 1µL of the resulting amplicon with the following primers to amplify the V4 domain: 515F: (AATGATACGGCGACCACCGAGATCTACACNNNNNNNNTATGGTAATTGTGTGCCAGCMGCCGCGGTAA) 806R: (CAAGCAGAAGACGGCATACGAGATNNNNNNNNAGTCAGTCAGCCGGACTACHVGGGTWTCTAAT).
Each primer used in the second PCR contained a unique 8 nt barcode (N) a primer pad (underlined), a two-base linker (italicized) and the Illumina adaptor sequences (bold text).A unique combination of these primers was used to barcode each sample PCR reaction contained 1 Unit Kapa2G Robust Hot Start Polymerase (Kapa Biosystems), 1X Kapa2G buffer with MgCl2, 10mM NTPs and 10 pmol of each primer.The rst PCR step cycling conditions were an initial incubation at 95°C for 5 min, followed by 25 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 90 s and a nal extension of 72°C for 7 min.The second PCR conditions were an initial incubation at 95°C for 3 min, followed by 25 cycles of 95°C for 45 s, 50°C for 60 s, and 72°C for 90 s and a nal extension of 72°C for 10 min.The nal PCR product was quanti ed on the Qubit instrument using the Qubit high-sensitivity DNA kit (Invitrogen) and individual amplicons were pooled together at the same concentration.Library quality was assessed using the Agilent Bioanalyzer 2100 High Sensitivity DNA assay.

16S rRNA gene amplicon library sequencing and bioinformatics
All samples were sequenced on an Illumina MiSeq platform at the Genome DNA Technology Core at the University of California, Davis.Demultiplexing of raw FASTQ les and adapter trimming of sequences was performed using dbcAmplicons version 0.8.5.(https://github.com/msettles/dbcAmplicons).The unmerged forward and reverse reads were imported into QIIME2 [29] version 2020.8 (https://docs.qiime2.org/2020.8/),and amplicon sequencing variants (ASVs) were determined following the DADA2 analysis pipeline [30].Snakemake [31] was used as a work ow manager to manage the QIIME2 environment (https://github.com/nasiegel88/tagseq-qiime2-snakemake-1).Each sequence was assigned to its given samples based on the given barcode.Reads that did not match any barcode were discarded (failed to meet minimum quality thresholds).Barcoded forward and reverse sequencing reads were quality ltered and merged.Sequences only observed once or in a single sample were also discarded.Chimeras were detected and ltered from paired end reads.A comparison of clustered sequences was performed against SILVA 138.

Statistical analysis
All R packages used for analyses were installed in R 4.1.3Studio software unless otherwise stated.Statistical analysis of microbial communities was performed primarily using the Bioconductor package Phyloseq (version 1.38.0).Spearman and Pearson correlations were done in R using the Microeco R package (version 0.7.0).Differential abundance analyses were performed using Linear discriminant analysis Effect Size (LEfSe) as described in the literature [32,33].Cytokine data were evaluated using ANOVA (1-or 2-way), unpaired t-test, or linear regression using rstatix (version 0.7.0).Outlier tests were performed where appropriate.Statistical signi cance was set at a p-value < 0.05.

Results
The infant rhesus monkey lung and oral microbiomes have distinct pro les H. pylori infection is highly prevalent in socially housed captive colonies of rhesus macaque monkeys [27,34].We have previously detected H. pylori infection in infant rhesus monkeys as early as 12 weeks of age, with approximately 90% of animals becoming culture positive at one year of age [28].To assess whether gastric bacterial infection is associated with an altered commensal microbiome for the respiratory tract, we rst evaluated a cohort of 25 infant monkeys (4-7 months of age, 8 males and 17 females) by gastric biopsy, lung wash and blood collection (Table 1).From our initial pool of 25 study animals, a total of 11 infant monkeys (6 males and 5 females) were identi ed as culture positive for H. pylori from gastric biopsy samples; we did not nd differences in age between H. pylori culture positive versus negative animals (Fig. S1A, B).There was an inverse relationship between infant H. pylori infection and maternal age, with younger dams associated with offspring having a higher H. pylori gastric load (p = 0.007) (Fig. S2).The lung microbiome of infant monkeys for both H. pylori positive and negative animals was primarily composed of the phyla (in order of relative abundance) Proteobacteria, Firmicutes, Bacteroidota, Fusobacteriota, Campilobacterota and Actinobacteriota (Fig. 1A).The oral microbiome for infant monkeys was primarily composed of the phyla (in order of relative abundance) Proteobacteria, Firmicutes, Bacteroidota, and Fusobacteriota.At the genus level, Actinobacillus, Streptococcus, Rodentibacter, Fusobacterium, Porphyromonas, Campylobacter, Flavobacterium, and Veillonella were most abundant in the lung microbiome, while Actinobacillus, Streptococcus, Rodentibacter, Fusobacterium, Porphyromonas, and Veillonella were most abundant in the oral microbiome (Fig. 1B, S5).The lung microbiome of H. pylori infection positive and negative infant monkeys shared 17.5% of taxa (Fig. 1C).The oral microbiome was less affected by H. pylori infection status, with 44.9% of taxa shared between H. pylori positive and negative infant monkeys (Fig. 1D).

Effect of H. pylori infection on the infant monkey lung and oral microbiome
To assess the impact of H. pylori infection on the composition of the lung and oral microbial communities in infant monkeys, we examined the alpha and beta diversity metrics between H. pylori positive versus negative animals.The Shannon diversity index did not signi cantly differ between the lung and oral samples as well as relative to H. pylori infection status in our study animals (Fig. 2A, 2B).
Principal component analysis (PCA) of unweighted UniFrac distance showed distinct lung and oral microbiomes that did not cluster by H. pylori infection status (Fig. 2B, 2C).
Infant monkey lung IL-8 is associated with H. pylori gastric load and microbial composition Increased IL-8 protein in the gastric mucosa has been detected in children infected with H. pylori relative to healthy subjects, although circulating IL-8 levels may be reduced [35,36].To assess whether H. pylori colonization during rhesus monkey postnatal development is similarly associated with elevated IL-8 expression at other mucosal or systemic sites, we measured IL-8 protein in lung wash and plasma samples collected in conjunction with microbiome sampling.H. pylori gastric load signi cantly correlated with lung wash IL-8 protein concentration (r = 0.59, p = 0.04), although we did not detect a signi cant difference when just comparing H. pylori infection status (negative versus positive) (Fig. 4A, S4A).In contrast with lung washes, we did not observe an association between plasma IL-8 concentration and H. pylori gastric load or infection status (Fig. 4B, S4B).However, plasma but not lung IL-8 concentration did correlate with the chronologic age of animals (p = 0.02, R = 0.22) (Fig. S5).There were also no sex differences in lung or plasma IL-8 concentration for infant monkeys in this study (Fig. S6A, S6B) To determine whether there was a relationship between H. pylori infection, relative abundance of genera and IL-8 concentration, we grouped samples by location and H. pylori status.There were no signi cant correlations between H. pylori positive versus negative animals in the lung (Fig. S7A) or oral microbiome (Fig. S7B).We next sought to assess the relationship between IL-8 production and the most abundant bacteria in the lung and oral microbiota independent of gastric H. pylori status.Staphylococcus was one of the most abundant taxa in the lung wash microbiome and was positively correlated with lung IL-8 (r = 0.93, p = 0.02) but not plasma IL-8 (Fig. 5A).Additionally, Stenotrophomonas positively correlated with lung IL-8 (r = 0.88, p = 0.04).Conversely, no genera from the buccal swab correlated with lung IL-8.Porphyromonas (r = 0.78, p = 0.047) and Alysiella (r = 0.78, p = 0.047) positively correlated with plasma IL-8 (Fig. 5B).Lastly, we interrogated whether alpha diversity indices correlated with IL-8 protein concentration, including community richness (observed species, Chao1 and ACE), community diversity (Shannon, Simpson), and phylogenetic diversity.We found diversity measures including Chao1, ACE, Observed ASVs, Fisher, and phylogenetic diversity negatively correlated with lung but not plasma IL-8 for both the lung and oral microbiomes (Fig. 5C, 5D).

Discussion
Because the human intestinal microbiome is known to display a temporal progression toward greater diversity and stability during the rst 2-3 years of life, we hypothesized that H. pylori infection during infancy might alter the composition of commensal populations at mucosal sites distant from the gastric region.Using the infant rhesus macaque monkey to model neonatal development, we rst pro led the lung and oral microbiome by 16S rRNA gene sequencing in a cohort of animals that were 6-7 months of age.We subsequently determined the relationship of the infant monkey lung and oral microbiome relative to gastric H. pylori infection status as well as IL-8 protein expression.Independent of gastric H. pylori colonization, we found the lung microbiome of infant monkeys was dominated by the phyla Proteobacteria, Firmicutes, Bacteroidota, Fusobacteriota, Campilobacterota and Actinobacteriota while the oral microbiome was dominated by Proteobacteria, Firmicutes, Bacteroidota, and Fusobacteriota.In the lung, H. pylori infection status was associated with shifts in seven genera and species.Comparatively, there were fewer genera and species for the oral cavity relative to H. pylori infection status.Because increased gastric IL-8 expression has been reported to be associated with H. pylori colonization in human adults and children, we assessed whether airway or systemic IL-8 levels might a) re ect H. pylori infection status and b) correlate with microbial communities withing the lung or oral cavity [35,37].We found the lung, but not plasma IL-8 protein concentration associated with gastric H. pylori load, differences in the genera Stenotrophomonas and Staphyloccus, and microbial diversity for both lung and oral cavities.Plasma IL-8 protein concentration was only associated with differences in the genera Alysiella and Moraxella in the oral cavity.
Our ndings on the commensal microbiome detected in lung wash from infant monkeys are similar to reports of 16S rRNA gene sequencing conducted on lung lavage collected from adult cynomolgus or rhesus macaques.In a longitudinal assessment of lavage obtained by bronchoscopy from cynomolgus macaques infected with SIV-HIV, the phyla Proteobacteria, Bacteroidota, Fusobacteriota, Firmicutes, and Actinobacteriota were predominant [38].Consistent with ndings in adult cynomolgus macaques, a study of aging rhesus macaque monkeys also described predominating in lung lavage the phyla Proteobacteria, Bacteroidota, Fusobacteriota, Firmicutes, and Actinobacteriota [39].At the genus level, a comparison of our data indicates a substantial overlap between infant and adult rhesus monkey lung microbial communities, with a prevalence of Actinobacillus, Streptococcus, Rodentibacter, Fusobacterium, Porphyromonas, Veillonella, Staphylococcus, Moraxella, Gemella, Neisseria, and Prevotella for both age groups [39].A distinctive nding of prior lung lavage studies in adult cynomolgus and rhesus monkeys is the dominant contribution of Tropheryma whipplei in subsets of characterized animals, ranging from one fourth to two thirds of surveyed cohorts [38,39].Tropheryma whipplei has been observed in the respiratory tract of HIV infected subjects, but the basis for colonization is unknown [40].Tropheryma whipplei was not reported in a separate study of 26 adult cynomolgus macaque monkeys, therefore colonization with this species may vary by animal source [41].In our study of the infant rhesus monkey lung, we did not observe Tropheryma whipplei as a major constituent of the microbiome, however it is possible the use of a sterile endotracheal tube to collect a lung wash may not have captured a species that is located within more distal alveoli.Lavage from adult macaque monkeys also did not appear to contain the phyla Campilobacterota, which was found in lung wash from infant macaque monkeys.It is notable that we detected the genus Campylobacter in the lung but not oral cavity of animals in our study, suggesting a different anatomical route into the airways.We were unable to nd H. pylori in either the lung or oral cavity samples despite gastric infection; however, we did detect H. fennelliae in the lavage of 2 gastric H. pylori-negative animals (data not shown).
For the oral microbiome, Proteobacteria was the predominant phylum observed for infant rhesus monkeys, however we have previously found the phylum Firmicutes as the most predominant source of reads from the oral cavity of adult rhesus monkeys from the same breeding colony as our current study [42].In a separate study from Chen, et.al. the composition of the oral microbiome from a population of wild adult macaque monkeys consisting of rhesus (M.mulatta) and cynomolgus (M.fascicularis) hybrid monkeys closely resembled captive adult rhesus monkeys and adult humans, with Firmicutes followed by Proteobacteria and Bacteroidota as the most abundant phyla [43].Differences in relative abundance of phyla may be due to inter-animal variability, however the consistency of oral microbial communities from both captive and wild macaque monkeys suggests that chronological age of the infant monkeys in our study is a contributing factor to the observed microbial signature of the lung or oral cavity.Indeed, there is evidence from longitudinal assessment of saliva from humans starting at the rst week of life that uctuations in abundance and composition of the oral microbiome continues until ve years of age [44].
The oral microbiome of 18-month-old children signi cantly differs from their respective parents, which lends further support to the notion that adult microbial communities are progressively shaped by environmental exposures [45,46].
We sought to assess the impact of H. pylori infection on the lung and oral microbiome of the infant monkey due to its inverse association with asthma development in childhood.In ammatory mediators such as IL-8 have been shown to increase with H. pylori infection both in plasma and the gastric mucosa [36,47,48].When characterizing the IL-8 pro les of lung wash and plasma from the study cohort of 25 infant rhesus macaques, we found that lung, but not plasma, correlated with gastric H. pylori load.Yet while lung lavage IL-8 levels positively correlated with H. pylori load, we were unable to detect H. pylori in the infant lung or oral samples collected for this study.Microbes aspirated from the oral cavity are subject to rapid mucociliary clearance, therefore detection may be challenging.Another potential limitation of this study is that it is not possible to discern the disease state of animals; gastric cultures may provide some evidence of the state of colonization but children and rhesus macaques have been found to spontaneously clear infection [49,50].Our prior in vitro studies have demonstrated that H. pylori infection can directly elicit IL-8 secretion in rhesus monkey airway epithelium in an age-dependent manner, with a greater than four-fold induction in infant versus adult-derived cultures [51].As such, it is feasible that even brief exposure to H. pylori in the respiratory tract without colonization could be a mechanism for the increased IL-8 observed in lung wash samples from infant monkeys in our study.
Source communities for the adult lung microbiome are likely to be organisms from the oral cavity and the upper respiratory tract but the oral microbiome is distinct from the lung microbiome [52,53].Although we did not detect statistically signi cant differences in alpha diversity measures between the lung and oral microbiome of infant monkeys in our study, we found the compartmental source of IL-8 differentially correlated with genera within either anatomical site.In the lung, we found a signi cant positive correlation between lung IL-8 and the genera Staphylococcus and Stenotrophomonas.In comparison for the oral cavity, there was a signi cant positive correlation between plasma IL-8 and the genera Porphyromonas and Alysiella.Given that Staphylococcus aureus and Stenotrophomonas maltophilia are opportunistic pathogens in the respiratory tract and frequently detected in cystic brosis patients, it may be speculated that innate immune function in the lung mucosa could contribute to shifts in these genera [54].The pathogenicity of Alysiella is unknown but it is commonly found in the oral cavity of mammals, whereas Porphyromonas gingivalis is associated with periodontal disease [55,56].
To date, investigation of the respiratory microbiome in pediatric populations have focused on minimally invasive sampling sites, such as the nasopharynx.The composition of the nasopharyngeal microbiome has been associated with respiratory infections and asthma development [47].However, nasopharyngeal microbial communities are signi cantly different from the actual lung microbiome in adulthood and tend to be similar to the microbial composition of the skin [19,53].An alternative approach to assess the pediatric respiratory microbiome is voluntary sample production, such as sputum.Sputum analysis as permitted comparison of lung microbiomes from pediatric and adult cystic brosis patients, which detected few differences in core lung microbiota in association with age [57].It can be di cult to interpret sputum samples for lung microbial communities as this type of collection runs the risk of upper respiratory contamination and signi cant contamination from the oral cavity [58].A higher abundance of Proteobacteria in adult and pediatric asthma patients relative to control subjects has been observed in bronchoalveolar lavage, however the study was limited by variability in age, sex, and disease status [17].Due to technical and ethical challenges of an invasive sampling approach, the lung microbiome of healthy human infants is a substantial knowledge gap that is crucial for addressing causality 16S rRNA sequencing data collected during disease states in pediatric populations.
In summary, we assessed the impact of an early life gastric infection on the infant lung and oral microbiome by taking advantage of an endemic H. pylori colonization in a captive cohort of infant rhesus macaque monkeys.To the best of our knowledge, this is the rst study to characterize the normal infant lung microbiome in a primate species.Moreover, our data suggests the postnatal primate lung harbors a microbiome signature that is distinct from the oral cavity.Gastric H. pylori colonization was associated with a shift in the composition of microbial communities in the lung and oral cavity; future studies would be needed to determine if an altered microbiome persists as H. pylori infected animals mature into adulthood.The therapeutic value of H. pylori-derived components for reducing allergic airways in ammation is under investigation; given our present ndings it may be of interest to assess whether altering the lung microbiome during infancy could prevent development of childhood asthma [59,60]. Declarations H. pylori infection status correlates with increased IL-8 in the lung.A Linear t of gastric H. pylori load and lung wash IL8 concentration (p=0.04,R=0.59).B Linear t of gastric H. pylori load and plasma IL8 concentration (p=0.89,R=0.002).

Figure 1 The
Figure 1

Figure 3 The
Figure 3

Table 1
To investigate whether gastric H. pylori infection can in uence the composition of microbial communities in the lung, we conducted 16S rRNA gene sequencing of lung wash samples obtained from 14 infant monkeys (H.pylori positive n = 6; H. pylori negative n = 8) which were a subset of our original 25 animal cohort (Supplemental Table1).Buccal swabs were also obtained from 12 out of the 14 animals prior to lung washes to compare microbiome composition between the oral and lung mucosal surfaces.A median number of 23,145 reads were generated per lung wash sample (range = 10,693 − 65,361 reads) and a median number of 56,690 reads were generated per buccal swab sample (range = 36,991 − 80,043 reads).