Bacterial strains and growth conditions

Thirty three H. pylori strains (31 clinical isolates and 2 collection strains), that had their clarithromycin resistance profile determined in this study by sequencing and E-test (see method description below), were used. All strains were maintained on Columbia Agar Base (Liofilchem s.r.l., Roseto D.A., Italy) supplemented with 5% (vol/vol) defibrinated horse blood (Probiológica, Belas, Portugal). Single colonies were streaked onto fresh media every 2 or 3 days, and the plates were incubated in a CO₂ incubator (HERAcell 150^®; Thermo Electron Corporation, Waltham, MA, USA) set to 10% CO₂ and 5% O₂, at 37°C [21, 22].

Design of PNA oligonucleotide probes for the detection of clarithromycin resistance

PNA probes were designed by adapting the already existing DNA probes, targeting the region of the point mutations described for this antibiotic in H. pylori [2]. Since PNA probes usually present higher melting temperatures it was possible to design shorter sequences with 15 nucleotides. The selected probes were Hp1 (A2143G) 5'-GGG TCT CTC CGT CTT-3', Hp2 (A2142G) 5'-GGG TCT TCC CGT CTT-3' and Hp3 (A2142C) 5'-GGG TCT TGC CGT CTT-3'. An additional probe to detect wild type strains (Hpwt 5'-GGG TCT TTC CGT CTT-3') was also included. Afterwards, the selected sequences were synthesized (Panagene, Daejeon, South Korea). The N terminus of the Hp1, Hp2 and Hp3 oligomers was connected to Alexa Fluor 488, and that of the Hpwt connected to Alexa Fluor 594, all via a double AminoEthoxyEthoxy Acetyl linker.

Fluorescence in situhybridization

As a starting point for the optimization of hybridization conditions the protocol previously described was used [14, 21]. Since the different probes only differed in one nucleobase, and for multiplex purposes, a common hybridization temperature was expected for all probes. Based on the brightest signals and specificity of the results, the best performance was obtained at 70°C (data not shown). H. pylori sensitive or resistant strain suspensions were prepared in water and 20 μl of each suspension was dispensed in 8 mm well slides (Marienfeld, Lauda-Königshofen, Germany) and then allowed to air dry. For permeabilization and fixation of bacteria, 30 μl of 4% paraformaldehyde (wt/vol) were placed in the wells with care to cover the entire surface, followed by 50% (vol/vol) ethanol for 10 minutes each, and then allowed to air dry. Approximately 20 μl of hybridization solution containing a mixture of the four probes were added to the fixed smears, which were then covered with coverslips and incubated for 1 hour at 70°C. Each 1 ml of hybridization solution contained 200 nM of the probes mixture, 10% (wt/vol) dextran sulphate, 10 mM NaCl, 30% (v/v) formamide, 0.1% (wt/vol) sodium pyrophosphate, 0.2% (wt/vol) polyvinylpyrrolidone, 0.2% (wt/vol) FICOLL, 5 mM disodium EDTA, 0.1% (vol/vol) Triton X-100 and 50 mM Tris-HCl (all from Sigma-Aldrich, Sintra, Portugal, except disodium EDTA that was from Pronalab, Lisbon, Portugal). Subsequently, the slides were transferred to a Coplin jar containing prewarmed (70°C) washing solution, that consisted of 5 mM Tris Base, 15 mM NaCl and 1% (vol/vol) Triton X-100 (all from Sigma-Aldrich, Sintra, Portugal), where the coverslips were carefully removed. The washing step was carried out for 30 minutes at 70°C. The slides were allowed to air dry and mounted with one drop of mounting oil and covered with a coverslip.

Specificity and sensitivity of PNA probes

After optimizing hybridization conditions, experiments with the PNA-FISH were performed on the 33 available strains in order to confirm the practical specificity and sensitivity of the probes. These results were compared with the gold standard susceptibility culturing test (E-test) and with the presence/absence of mutations in the 23S rRNA gene.

Validation of the testing protocol in gastric biopsy slides for clinical application

To validate the method in the stomach tissue, thirty nine paraffin-embedded gastric biopsy specimens from patients with known resistance antibiotic profile by antibiogram were used. The study was in accordance with the institutional ethical standards. Informed consent was obtained from the patients. Three-micrometer thick paraffin cuts were deparaffinized and rehydrated in xylol and ethanol based on a protocol previously described [21]. Sections were emerged in xylol (Fisher Chemical, Leicestershire, U.K.) three times (firstly for 15 minutes, and then twice for 10 minutes each), absolute ethanol (Panreac, Barcelona, Spain) (twice for 7.5 minutes each) and ethanol decreasing concentrations (95%, twice for 7.5 minutes each; 80%, 10 minutes; 70%, 10 minutes; 50%, twice for 15 minutes each). Finally sections were immersed in 1% (vol/vol) Triton X-100 (Sigma-Aldrich, Sintra, Portugal) solution for 20 minutes at 70°C. Histological slides were then allowed to air dry and the hybridization protocol previously described for smears, with the exclusion of the fixation step, was used. The completion of the whole procedure takes 4 h 15 m.

Susceptibly test: E-test

In order to confirm the susceptibility profile, the minimal inhibitory concentration (MIC) of each strain was determined by the E-test, in accordance with the company instructions (AB Biodisk, Biomérieux, Portugal). Briefly, 2 day-old pure cultures were inoculated into Mueller-Hinton broth, supplemented with 5% (vol/vol) fetal calf serum [23] and the turbidity of the inoculum adjusted to McFarland standard 3 [7]. Agar plates containing Mueller-Hinton supplemented with 5% (vol/vol) defibrinated horse blood (Probiológica, Belas, Portugal) were inoculated by swabbing the surface with the inocula. One E-test strip was applied on the surface of the plate, after drying. The plates were incubated in a CO₂ incubator (HERAcell 150^®; Thermo Electron Corporation, Waltham, MA, USA) set to 10% CO₂ and 5% O₂ at 37°C for 72 h or until visible inhibition ellipse was seen [2, 7, 23]. Strains were considered susceptible when the MIC was < 1 μg/ml, and resistant when the MIC was > 1 μg/ml [9].

Assessment of clarithromycin resistance in gastric tissues by PCR and sequencing

Total DNA was extracted from biopsy samples after digestion with Proteinase K for at least 12 hours at 55°C. Proteinase K was inactivated by incubation at 95°C for 10 minutes. Ten microliters of the lysates were used for PCR amplification of H. pylori 23S rRNA gene as previously described [24]. PCR products were sequenced using BigDye Terminator v3.1 Cycle Sequencing Kits (Applied Biosystems, CA, USA) and run in an ABI Prism 3130 DNA automated sequencer (Applied Biosystems). In some H. pylori isolates, PCR and sequencing were used to characterize the 23S rRNA gene.

Microscopic visualization

Visualization of samples never exceeded 48 h after the experimental procedure. Smears or histological slides were observed using an epifluorescence microscope (BX51 Olympus, Hamburg, Germany) equipped with filters adapted to the Alexa Fluor (488 and 594) signalling molecules within the probes. The filters that were not sensitive for the reporter molecules were used as negative control.

Specificity and sensitivity of the PNA-FISH probes

From the three mutations, the one that is less frequent is the A2142C transversion [1, 12], and in this study we were only able to test one strain with that mutation. Nevertheless, the available strain was always detected when the Hp3 probe was present in the hybridization solution. Probes Hp1, Hp2, and Hp3, hybridized only with the resistant strains that had the corresponding point mutations conferring clarithromycin resistance and as such presented 100% sensitivity (95% CI, 79.9-100) and 100% specificity (95% CI, 71.6-100). The set of probes can discriminate the resistant and susceptible strains, even though they only have one mismatch. We next further tested the method using a mixture of the four probes simultaneously in a multiplex detection (figure 1; A-C). In this case, the detection of point mutations was even more robust, which is possibly due to the fact that all probes target the same locus, and as such there is a competition effect between them. However, with the mixture it is only possible to discriminate between clarithromycin resistant and clarithromycin sensitive strains, as opposed to the discrimination between point mutations that was conferred by using the probes separately. In practical terms and considering the application of the PNA-FISH to the clinical setting, the mixture of probes introduces an important simplification to the method.

Validation of the testing protocol in gastric biopsy slides for clinical application

Considering the application of the PNA-FISH method in clinical settings, we used the developed PNA probes to identify and differentiate clarithromycin resistant and susceptible H. pylori strains in histological slides of gastric biopsy samples. Results clearly show that it is possible to discriminate susceptible from resistant H. pylori strains and, in the latter group, to detect the three different mutations, using fluorescence microscopy (figure 1; D-K). Taking into consideration the antibiogram as the gold standard, the PNA-FISH method showed specificity and sensitivity of 90.9% (95% CI, 57.1-99.5) and 84.2% (95% CI, 59.5-95.8), respectively (data not shown). These can probably be explained by the existence of another mechanism of resistance apart from the three point mutations assessed in this study. In fact, association between A2142C, A2142G and A2143G mutations and clarithromycin resistance was defined as approximately 84% in a world wide data compilation [3].

Although, these results need to be confirmed in a larger series of cases, they indicate that PNA-FISH is applicable to clinical samples and may be useful for the selection of the most adequate antibiotic combination to be used for H. pylori eradication. Moreover, this method is simple to perform and the procedure is fast (4 h 15 m), indicating that results can be provided to clinicians simultaneously with the histological diagnosis.