Introduction

The International Agency for Research on Cancer has characterized Helicobacter pylori as a type I carcinogen [1] that is responsible for gastritis, gastro-duodenal ulcers, and gastric malignancies in humans [2]. Although it is one of the most common infections in the world and is known to be transmitted in early childhood, the exact route of transmission is still unclear [3]. In a recent meta-analysis, a close relationship was found between H. pylori infection in the oral cavity and stomach; the authors concluded that H. pylori in the oral cavity was more difficult to eradicate than in the stomach, and may therefore be a source of gastric re-infections. However, the specific populated niche in the oral environment is unknown [4].

The majority of studies analysed specimens of dental plaque, saliva, or oral mucosa, and identified several H. pylori markers by various tests, such as the urea breath test, rapid urease test, Campylobacter-like organism test, or polymerase chain reaction (PCR). Although PCR studies have previously found H. pylori DNA in the oral cavity, reports of live H. pylori are extremely rare and highly inconclusive [4, 5]. Unequivocal identification of live H. pylori is only possible by direct culture, because erroneous PCR results can arise from transient H. pylori presence in the mouth via food or via the reflux of H. pylori or its DNA from the stomach to the mouth [6–8]; erroneous results can also arise from the misclassification of other urease-producing microorganisms. Thus, it is still unclear whether H. pylori can indeed survive in the oral environment. In this article we report two cases of successful isolation of live H. pylori from the oral cavity, particularly from root canal samples of teeth.

Methods

Patient selection and characteristics

We selected three consecutive pediatric patients who received dental treatment under general anesthesia because of severe early childhood caries. Table 1 summarizes data about the patients’ age and gender, as well as the tooth numbers of the extracted teeth. Altogether 10 teeth with pulp necrosis and chronic apical periodontitis were used for the analyses. The presence of gastric H. pylori in the children or their parents was not checked in the study, because there was no indication of gastric or abdominal problems. All parents gave their written consent for microbiological analyses of the teeth. The study protocol was reviewed and approved by the ethics committee of the University of Leipzig.

Table 1 Sample characteristics and Helicobacter pylori identification
Full size table

DNA isolation and H. pylori growth

Plaque and root canal samples were taken from each tooth. These samples were divided into three parts, one for conventional DNA isolation (DNA isolation kit; Qiagen, Hilden, Germany), a second for electron microscopy (see below), and a third for direct culturing. For culturing, the samples were incubated with 1 ml brain heart infusion medium by rigorous shaking at 200 rpm for 30 min, followed by growth on GC agar plates with 10 % horse serum (containing 10 μg/ml vancomycin, 5 μg/ml trimethoprim, 10 μg/ml nystatin, and 10 μg/ml colistin) for 7 days at 37 °C, using the Campygen gas-generating system (all reagents from Oxoid/Fisher Scientific, Dublin, Ireland) [9, 10]. Single bacterial colonies were further analyzed. and typical H. pylori strains (26695 and J99) were used as positive controls. To check for functional urease in H. pylori, the above GC agar plates were supplemented with phenol red (100 μg/ml) and urea (600 μg/ml) as described [11]. The molten agar was then acidified to pH 5 using 1 M HCl [11].

16S rRNA gene PCR and electron microscopy

For PCR amplification of the 16S rRNA gene in the genus Helicobacter, primers 5′-AGA GTT TGA TYM TGG C-3′ and 5′-TAC GGY TAC CTT GTT ACG A-3′ were used, and amplicons were sequenced as described [10]. For field-emission scanning electron microscopy (FESEM), tooth samples were fixed in a sterile solution containing 5 % formaldehyde and 2 % glutaraldehyde in cacodylate buffer (0.1 mM cacodylate, 0.01 mM CaCl2, 0.01 mM MgCl2, 0.09 mM sucrose, pH 6.9). The samples were subsequently covered with an approximately 10-nm-thick gold film by sputter coating and examined in a field-emission scanning electron microscope using an Everhart–Thornley secondary electron (SE) detector and in-lens detector in a 50:50 ratio at an acceleration voltage of 5.0 kV as described [10].

Protein profiling and western blotting

For protein profiling, pure plate-grown bacterial samples were run on 12 % sodium dodecylsulfate-polyacrylamide (SDS-PAGE) gels and analyzed by Coomassie blue staining or western blotting [9]. The following primary antibodies were used: mouse monoclonal anti-CagA antibody (Austral Biologicals, San Ramon, CA, USA), mouse polyclonal anti-urease antibodies [9], and polyclonal rabbit antibodies recognizing a series of other H. pylori proteins. These antibodies were raised against peptides corresponding to the following conserved amino-acid (aa) residues in H. pylori strain 26695: BabA (aa 126–140: CGGNANGQESTSSTT), SabA (aa 172–186: CAMDQTTYDKMKKLA), OipA (aa 275–288: NYYSDDYGDKLDYK), NapA (aa 105–118: EFKELSNTAEKEGD), Slt (aa 492–505: LRRWLESSKRFKEK), HtrA (aa 90–103: DKIKVTIPGSNKEY), FlaA (aa 93–106: KVKATQAAQDGQTT), VirB9 (aa 503–522: IKNYGELERVIKKLPLVRDK), VirB10/CagY (repeat region: VSRARNEKEKKE), and Cag3/Cagδ (aa 32–45: IKATKETKETKKEA). Rabbit anti-CagM, anti-CagN, and anti-VacA antibodies were raised against all the recombinant proteins. These antibodies were affinity-purified and prepared according to standard protocols by Biogenes (Berlin, Germany). Horseradish peroxidase-conjugated anti-mouse or anti-rabbit polyvalent sheep immunoglobulin was used as secondary antibody (DK-2600, DAKO, Glostrup, Denmark). Blots were developed with ECL Plus western blot reagents (GE Healthcare UK, Amersham, UK) as described [9].

Results

To investigate whether H. pylori was present in the 10 tooth samples, DNA isolated from both plaque and root canal samples was subjected to PCR to amplify a ~1.5-kb DNA fragment derived from a 16S rRNA gene region that is highly conserved in Helicobacter. The expected strong PCR products were produced in two root canal and four plaque samples, suggesting that H. pylori DNA may have been present in some but not all patients (Table 1). To isolate viable H. pylori, all samples were prepared and cultured for seven days on selective agar plates to suppress other bacteria. Single colonies were identified under microaerobic growth conditions in two of the 10 root canal samples (samples 4 and 5 from patient #2), but not from any plaque sample. These two root canal samples were then subjected to FESEM investigation to see if typical H. pylori bacteria could be visualized. FESEM indeed revealed various H. pylori-like spiral-shaped organisms in the two samples in close association with tooth debris (Fig. 1a, yellow arrows). These candidate H. pylori were approximately 0.2 μm in diameter and varied in length from 2 to 3 μm. Several monopolar flagella were also observed, typical of H. pylori [10]. In addition, and as expected, coccoid bacteria of an unknown nature, which could also represent H. pylori, were observed (Fig. 1a, blue arrows). These morphological data suggested the presence of live, spiral-shaped H. pylori in the root canal environment of teeth.

Fig. 1
figure 1

Morphological analyses of two root canal samples (4 and 5) by field-emission scanning electron microscopy (FESEM) and urease tests. a FESEM revealed Helicobacter pylori-like spiral-shaped bacteria (yellow arrows) that were approximately 0.2 μm in diameter and varied in length from approximately 2 to 3 μm. Coccoid bacteria 0.5–1 μm in diameter were also observed in large aggregates (blue arrows). Representative pictures are shown from two preparations. b Selection of bacteria producing functional urease on acidified agar supplemented with urea. Left samples root canal samples 4 and 5 and strain 26695; the observed color change from orange to red indicated that bacterial colonies were producing functional urease and growing. Right samples 26695ΔureA and plaque samples 4 and 5. Color change did not occur in the right samples, indicating that functional urease was not being produced

Full size image

To exclude artifacts, bacteria were grown on selective acidified agar plates supplemented with urea, the substrate of H. pylori urease [11]. These experiments yielded functional urease enzymes allowing urea hydrolyzation in root canal samples to a high extent, similar to that in H. pylori control strains, while retarded growth and no urea hydrolyzation was seen in ΔureA mutants or in any of the non-H. pylori samples from dental plaque (Fig. 1b). To unquestionably identify H. pylori, we determined the 16S rRNA gene sequences from the two root canal isolates, as described [10]. Both strains had completely identical sequences showing strong homology to that of several published H. pylori strains (Fig. 2a). To characterize our isolates further, we performed western blotting and confirmed the presence of several well-known H. pylori-specific pathogenicity factors as compared to the fully sequenced strains 26695 and J99. Specific antibodies revealed the presence of urease subunits A and B, as well as a major disease-associated factor, CagA (Fig. 2b, arrows). In agreement with the observation of flagella by FESEM, we also found that our isolated root canal strains expressed the flagellin component FlaA (Fig. 2b). Moreover, the presence of certain adhesins (BabA, SabA, and OipA), cag pathogenicity island encoded proteins (CagL, CagM, CagN, Cag3, VirB9, and VirB10), and other virulence factors (NapA, HtrA, Slt, and VacA) was also confirmed by western blotting, using specific antibodies (data not shown). Thus, our findings clearly indicate the successful isolation of live H. pylori from the root canals of teeth.

Fig. 2
figure 2

16S rRNA sequencing and western blotting analysis of H. pylori-specific pathogenicity factors. a Phylogenetic tree of 16S rRNA gene sequences of root canal samples and closest H. pylori strains. b Western blotting analysis of root canal samples for well-documented H. pylori proteins including urease A, urease B, CagA, and flagellin A (FlaA)

Full size image

Discussion

H. pylori can be cultured from human stomach biopsies, but attempts to identify other natural reservoirs for these organisms or the routes by which they are transmitted to the stomach have been unsuccessful [5, 12]. Here, live bacteria from two root canal samples were unequivocally identified as H. pylori. To our knowledge, this is the first report of the recovery of viable H. pylori from root canal samples, suggesting that this environment may be a reservoir for H. pylori survival and growth that could serve as a potential source for the organism’s transmission. It is possible that these bacteria are of gastric origin, and that patients carrying H. pylori in their dental root canals are also colonized by the same or different strains in the gastric mucosa. Colonization of the root canal may explain why eradication is often unsuccessful, as the antibiotic therapy used may not penetrate the root canal. Whether or not this environment represents a reservoir for H. pylori which facilitates transmission among humans is a pressing question for future studies.