Molecular Resistance Testing of Helicobacter pylori in Gastric Biopsies

Techniques for diagnosing Helicobacter pylori infection are classifiable as invasive or noninvasive. Invasive diagnostic methods require a gastric biopsy obtained via esophagogastroduodenal endoscopy and include histologic examination, microbiologic culture, or rapid urease testing (RUT). Rapid urease testing is an attractive diagnostic method because it is timely, affordable, and only requires visual interpretation following incubation.

Clarithromycin is a drug commonly used in eradication therapy for H pylori,1 but the increasing prevalence of macrolide resistance has been associated with treatment failure.2–5 The mechanism of clarithromycin resistance in H pylori is due to point mutations in the 23S ribosomal RNA (rRNA) gene, resulting in decreased binding of clarithromycin to 23S rRNA.6,7 Point mutations in 23S ribosomal DNA (rDNA) include transition (A2142G and A2143G) and transversion (A2143C) mutations. The A2142G mutation has been associated with an increased level of resistance and a higher minimum inhibitory concentration than the A2143G mutation.8 

Conventional antimicrobial susceptibility testing (AST) may be performed with H pylori isolates after documented therapy failure,3 but requires an additional endoscopy and subsequent bacteriologic culture. An agar dilution method has been officially approved by the National Committee on Clinical Laboratory Standards (NCCLS) for clarithromycin susceptibility testing of H pylori.9,10 In this study, the feasibility of molecular diagnosis of infection and molecular resistance testing (MRT) of H pylori using a single gastric biopsy was evaluated.

Twenty-six rapid urease-positive and 51 rapid urease-negative test kits (CLOtest, Ballard, Draper, Utah) were collected from the Microbiology Laboratories at the Massachusetts General Hospital (Boston, Mass). Gastric biopsy specimens were obtained as part of a discarded tissue protocol (No. 1999-P-009601/1) approved by Human Studies (Institutional Review Board) at the Massachusetts General Hospital. Biopsies of the gastric antrum, corpus, and duodenum were obtained by esophagogastroduodenal endoscopy using sterile cold forceps. After inoculation of each biopsy into a RUT kit, the kits were incubated at 37°C in ambient air, and color changes (yellow to pink/fuchsia) were observed at 1, 3, and 24 hours postincubation. All specimens demonstrating urease activity (pink/fuchsia color) were reported as positive for H pylori. Prior to DNA extraction, RUT kits were maintained at 4°C. In addition, 4 RUT biopsies were maintained at room temperature for up to 8 weeks prior to processing.

Following incubation at 37°C, biopsy material was placed into sterile 1.5-mL BeadBeater tubes (Biospec Products, Bartlesville, Okla) with 40 μL lysis buffer (0.05 mol/L sodium hydroxide, 0.25% sodium dodecyl sulfate), then incubated at 95°C for 5 minutes. Biopsy lysates were prepared by addition of 200 μL dH₂O and an equal volume of 0.1-mm glass beads (Biospec Products), followed by agitation in a Mini-BeadBeater (Biospec Products) at maximum velocity (5000 rpm) for 5 minutes at room temperature. Lysis was followed by centrifugation at 12 000 rpm for an additional 5 minutes, supernatant collection, and storage at −20°C.

A summary of oligonucleotides used to amplify various H pylori gene fragments is listed in Table 1. A polymerase chain reaction (PCR; final volume of 50 μL) was performed in a DNA Thermal Cycler 480 (Applied Biosystems, Foster City, Calif). Reactions included a mixture of biopsy lysate (1–5 μL), 0.5 units Taq DNA polymerase (Roche Molecular Biochemicals, Indianapolis, Ind), manufacturer-provided 10× reaction buffer (supplemented with 22 mmol/L magnesium chloride), 100 μg/mL bovine serum albumin (Roche), 100 mmol/L deoxynucleotide triphosphate (dNTP) solution (Amersham Pharmacia Biotech, Piscataway, NJ), and 100 ρmol of each primer. Cycling conditions were as follows: initial denaturation (94°C, 6 minutes); 40 cycles of 94°C for 30 seconds, 50°C for 1 minute, then 72°C for 3 minutes; and a final extension (10 minutes at 72°C).

A 1402-base pair (bp)-long fragment of the 23S rDNA gene of H pylori was amplified and gel-purified using Quantum Prep Freeze ’N Squeeze DNA Gel Extraction spin columns (Bio-Rad, Hercules, Calif). Nested PCR was performed to amplify a 696-bp internal portion of the 23S rRNA gene from the purified 1402-bp amplicon, using primers HPRP-1 and -2 (see Table 1). A protocol previously developed by Versalovic et al6 for detecting clarithromycin-associated point mutations in the 23S rRNA gene was modified as follows: enzymatic digestion of the internal 696-bp amplicon replaced that of the 1402-bp 23S rDNA fragment. Briefly, unincorporated dNTP and primers in the 696-bp nested PCR product were removed prior to restriction enzyme digestion using a PCR Clean-Up Kit (Roche). Digestion reactions contained 10 μL purified PCR product, manufacturer-provided buffer, and 1.5 units restriction enzyme. The purified PCR products were digested during a 14-hour incubation with BsaI and MboII (New England Biolabs, Beverly, Mass) at 50°C and 37°C, respectively. Heat inactivation of restriction enzymes at 65°C for 20 minutes followed overnight digestion. Restriction digestion products were separated by electrophoresis in a 3% agarose gel.

A total of 77 CLOtest specimens were collected and processed as indicated. Of 77 biopsies, 33.8% (26/77) were rapid urease-positive and 51 were rapid urease-negative after 24 hours of ambient air incubation at 37°C. Of the 26 patients with RUT-positive results, endoscopic studies yielded findings associated with inflammation in 22 patients. Histologic examination was performed with only 13 specimens, all of which yielded histopathologic findings consistent with gastritis or duodenitis. Eight (31%) of 26 rapid urease-positive specimens were obtained from patients with peptic ulcer disease.

Two genes were targeted for amplification, namely, a conserved region flanked by genus-specific primer binding sites in Helicobacter 16S rDNA and a species-specific sequence, glmM (ureC), encoding a phosphoglucosamine mutase of H pylori.11 The presence of H pylori DNA was demonstrated by amplification of both a 16S rDNA fragment (422 bp) and a glmM product (294 bp). Both gene targets, 16S rDNA and glmM, were amplified from all 26 urease-positive biopsies, indicating the presence of H pylori DNA. Helicobacter 16S rDNA segments were amplified from 7.8% (4/51) of biopsies negative by RUT, whereas glmM amplicons were generated from 3.9% (2/51) of gastric specimens negative by RUT. In addition to the 77 biopsies included in this study, 11 RUT kits yielded equivocal results (orange). Either the genus-specific 16S rDNA target or the H pylori–specific glmM segment were amplified from equivocal RUT specimens (data not shown).

Gastric biopsies positive for both urease activity and H pylori DNA were further tested for the presence of antimicrobial resistance determinants. Amplicons of a size similar to that of H pylori were not generated by PCR amplification of purified human genomic DNA under identical conditions (data not shown). These data demonstrated the specificity of primer sets complementary to H pylori glmM and 23S rDNA sequences. Purification and subsequent nested PCR of the 1402-bp fragment of the 23S rRNA gene produced the internal 696-bp amplicon, also supporting the identity of the product as the H pylori 23S rRNA gene. In addition, both 1402-bp and 696-bp amplicons from H pylori Sydney strain 1 (control) were sequenced to confirm the specificity of primers used to amplify 23S rDNA fragments. Restriction enzyme–mediated digestion with BsaI and MboII yielded specific, expected DNA fragments for the 24 samples analyzed. Figure 1 shows profiles obtained following restriction digestion and electrophoresis in 3% agarose for specimens lacking the A2142/3G mutation, as well as those with A2142G and A2143G mutations in the 23S rDNA. Enzyme restriction analyses of the 23S rDNA amplification products from gastric biopsies yielded evidence of clarithromycin-resistance mutations (A2142G or A2143G) in 17% (4/24) of specimens. A summary of PCR amplification and mutation analyses of RUT biopsy specimens is provided in Table 2.

In this study, H pylori DNA was successfully amplified from 100% (26/26) of urease-positive and 3.9% (2/51) of urease-negative gastric biopsies. Subsequent restriction enzyme–mediated digestion of 23S rDNA amplification products revealed that 17% (4/24) of urease-positive and H pylori DNA–positive biopsy specimens contained point mutations (A2142G or A2143G) associated with clarithromycin resistance. Helicobacter pylori DNA from gastric biopsies was successfully amplified 8 weeks following RUT.

To directly detect mutations conferring clarithromycin resistance, the presence of H pylori DNA in gastric biopsies was demonstrated by PCR amplification. Multiple genetic targets have been useful for detection of H pylori in gastric tissue, including cagA, vacA, ureA, glmM (ureC), and 16S rDNA.12–14 Polymerase chain reaction–based detection of vacA14 yielded maximal diagnostic sensitivity (99.4%) when compared with RUT (89%). Lu et al15 reported the highest sensitivities and specificities with 16S rDNA and glmM PCR, respectively. In this study, amplification of 16S rDNA was used to detect Helicobacter DNA,16 followed by confirmatory testing by amplification of the H pylori glmM segment.15 It must be emphasized that the selection of primers targeting glmM must be carefully considered. Stretches within the amplified 294-bp segment of glmM are polymorphic, albeit these substitution mutations are usually silent mutations.17 Molecular resistance testing was performed only with specimens that fulfilled biochemical and molecular criteria for the presence of H pylori (ie, urease positive and H pylori DNA positive). Several biopsies failed to meet these criteria. Two urease-equivocal specimens did not yield the glmM amplicon following PCR, although 1 specimen yielded the Helicobacter-specific 16S rDNA segment. This patient may be infected with a different gastric Helicobacter or related organism, and its identity is being pursued by 16S rDNA sequencing.

To our knowledge, this study represents the first report documenting the prevalence of antimicrobial resistance in H pylori in the northeastern United States. Although the sample size (n = 24) is limited, the documented 23S rDNA mutation prevalence (17%) in Boston is significantly higher than the reported prevalence of clarithromycin resistance in Europe (2%–10%)3,18 and elsewhere in the United States.19,20 In Texas, the prevalence of clarithromycin resistance exceeds 6%,19 while in the Midwest, an increase from 4% (1994) to 12.6% (1996) has been reported20 by conventional AST. Elevated resistance rates have been documented in patients with secondary resistance following treatment failure. For example, the MACH 2 study in Europe detected resistance in 12 (11.4%) of 105 patients after treatment failure.21 

Conventional AST of H pylori resulting in a minimum inhibitory concentration, as specified by the NCCLS,9 may result in underreporting of macrolide resistance as defined by mutational analyses. Maeda et al22 documented the increased detection sensitivity of MRT for H pylori from gastric juice specimens, when compared with AST by agar dilution methods. In their study, 9.7% (18/186) of patients tested harbored pure strains of H pylori with the A2143G mutation in the 23S rDNA. Additionally, 10% (19/186) of patients were infected with both resistant and susceptible cells (stable variants/mutants) existing in a mixed population. Of these 19 patients, agar dilution testing (conventional AST) identified only 9 patients as being infected by a population comprising both macrolide-susceptible and -resistant H pylori. Ten patients (5.4%) would have been misdiagnosed as being infected with a pure population of susceptible H pylori.

The ability to detect mixed infection by multiple strains of H pylori represents a key advantage of genotyping methods.23 Compared with culture-based methods, molecular detection of such mutations provides more accurate antimicrobial susceptibility information when multiple strains are present. Helicobacter pylori isolates can exhibit variable minimum inhibitory concentrations by conventional testing when mixed strains are present, potentially confounding susceptibility values and indicating therapy, whereas molecular approaches reveal the presence of wild-type (susceptible) and mutant 23S rDNA alleles in the same specimen.24 Evidence of a mixed-strain population was documented by genotyping of strains isolated from European patients.23 The presence of such infections had a negative impact on treatment efficacy.22 We also detected the presence of both the wild-type and 2142G 23S rDNA alleles in 1 specimen (data not shown).

Single point mutations (A2142G and A2143G) in H pylori 23S rDNA are necessary and sufficient to confer macrolide resistance in vitro.25 Furthermore, in vitro resistance and mutations in 23S rDNA have been associated with treatment failure in multiple studies.26 The presence of clarithromycin-resistant H pylori in infected individuals significantly diminishes efficacy when using clarithromycin in current treatment modalities. A recent meta-analysis found that the presence of clarithromycin-resistant H pylori reduced treatment efficacy by an average of 55%.5 If patients are treated with clarithromycin and a proton pump inhibitor, the presence of macrolide-resistant H pylori is nearly 100% predictive of treatment failure.27 In another study, 83% to 98% of patients infected with clarithromycin-susceptible strains were successfully treated, whereas only 25% to 50% of patients harboring strains resistant to clarithromycin were eradicated.3 

Empiric treatment regimens have been recommended in areas of low prevalence of clarithromycin resistance, whereas tailored regimens based on pretreatment AST (or MRT) are preferred in areas of high prevalence of resistance.28,29 Customized therapies, based on pretreatment AST (or MRT), may be more cost-effective with improved cure rates.30 Since recommendations for pretreatment antibiotic susceptibility testing depend on the prevalence of resistance in a given locale, routine surveillance of drug-resistant H pylori may prove beneficial. Molecular resistance testing could be utilized as a tool to monitor drug resistance in H pylori and may represent the only viable option in laboratories where single gastric biopsies are submitted for rapid urease testing. Biopsy material has been used previously to directly amplify H pylori genetic targets.31,32 Since CLOtests are incubated at 37°C for 1 to 24 hours prior to final reporting of urease status, recovery of H pylori from these biopsies for subsequent AST is limited. Helicobacter pylori isolates were not cultured in the current study owing to prolonged storage of biopsies in indicator media at room temperature. This concurs with findings that H pylori is rarely culturable from gastric biopsies submitted for RUT beyond 2 hours post–gel inoculation.33 

In this study, H pylori DNA was amplified from RUT-inoculated specimens kept at room temperature for up to 8 weeks postinoculation. Stored biopsies may be tested for clarithromycin-resistant H pylori in patients refractory to therapy. The ability to perform molecular analyses in RUT specimens stored for several weeks avoids the need for repeat endoscopy in patients failing treatment. Bacterial genotyping circumvents the need for additional medical procedures (esophagogastroduodenal endoscopy) by facilitating the diagnosis of infection and MRT from a single gastric biopsy. As with human immunodeficiency virus type 1 genotyping,34 customized therapies based on pretreatment molecular analysis may be cost-effective with an improvement in cure rates.

The purpose of this article is to document the ability to detect H pylori and perform MRT by PCR-based assays with gastric biopsies. In summary, we have demonstrated that H pylori 23S rDNA genotyping may detect macrolide-resistant H pylori in gastric biopsies. This assay can be performed prior to initiation of therapy or in patients refractory to anti–H pylori therapy. We propose an algorithm for the molecular diagnosis and resistance genotyping of H pylori in clinical laboratories (Figure 2). After biopsy specimens are obtained from patients with ‘alarm’ features (eg, age, weight loss, and gastrointestinal bleeding),35 RUT is performed, and specimens positive for urease activity are then stored at 4°C for possible MRT. Clinical laboratories may retain biopsies for surveillance purposes. In areas with low prevalence of clarithromycin resistance (less than 5%), therapy should commence immediately without susceptibility testing. If a patient fails therapy (documented by follow-up noninvasive testing, such as the urea breath test or fecal antigen test), MRT may be performed using the archived biopsy, and the regimen may be modified based on genotyping results. For patients who have either failed therapy previously or have been diagnosed with H pylori infection in an area with high resistance prevalence (exceeding 10%), MRT should be considered prior to antimicrobial therapy. Areas with intermediate resistance prevalence (5%–10%) represent locations where baseline H pylori genotyping would be justifiable.

This work was supported by funding from the Van Slyke Society of the American Association of Clinical Chemistry, Washington, DC (Dr Versalovic) and grant R01 AI37750 from the US National Institutes of Health, Bethesda, Md (Dr Fox). The authors also acknowledge the Massachusetts General Hospital Microbiology Laboratories, Boston, Mass, for the clinical specimens used in this study and James R. Lupski, MD, PhD, for providing purified human genomic DNA.