which produced equal-density bands for infected stomach sam- ples and .... Baskerville, A., and D. G. Newell. 1988. ... Bell, G. D., and K. U. Powell. 1993.
CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Jan. 1996, p. 73–78 1071-412X/96/$04.0010 Copyright q 1996, American Society for Microbiology
Vol. 3, No. 1
A Sensitive and Specific PCR Method To Detect Helicobacter felis in a Conventional Mouse Model L. KONG,1* J. G. SMITH,1 D. BRAMHILL,2 G. K. ABRUZZO,1 C. BONFIGLIO,2 C. CIOFFE,3 A. M. FLATTERY,1 C. J. GILL,1 L. LYNCH,1 P. M. SCOTT,1 L. SILVER,1 C. THOMPSON,2 H. KROPP,1 AND K. BARTIZAL1 Antibiotic Discovery and Development,1 Department of Enzymology,2 and Department of Laboratory Animal Science,3 Merck Research Laboratories, Merck and Co., Inc., Rahway, New Jersey 07065-0900 Received 21 June 1995/Returned for modification 21 August 1995/Accepted 3 October 1995
Although many detection methods have been used to determine Helicobacter colonization in small animal models, the sensitivity and specificity of these detection methods are limited. To improve the Helicobacter felis conventional mouse model for accurate evaluation of therapeutic regimens, we developed a PCR for detection of, and a competitive PCR for quantitation of, H. felis in viral antibody-free (VAF) mice. The PCR was based on the H. felis 16S rRNA gene. An internal control DNA was used for competitive quantitation of the PCR. VAF conventional Swiss-Webster mice were infected with an H. felis culture by oral gavage. At various times after H. felis challenge and therapy, stomach mucosa was collected and evaluated by PCR. PCR detected approximately 50 to 100 H. felis cells per mouse stomach and showed no cross-reaction with other bacteria commonly found in mouse stomachs. Colonization of H. felis in the mouse stomach was confirmed by culture isolation from germfree mice and histological examination of VAF mice. Response to therapy in this H. felis model correlated well with results seen in human clinical trials with H. pylori. A model utilizing PCR detection which may be useful for discovering new antibiotics and/or vaccines against Helicobacter ulcer disease has been developed.
assays (9, 23, 26, 27). All of these techniques have some disadvantages. Culturing is often difficult for these fastidious, slow-growing organisms, which require a rich medium and special culture conditions. Histology and urease are too insensitive to confirm complete eradication of organism after treatment and may not be specific in the presence of contaminating microorganisms (10, 27). Serology may not differentiate active from past infection and thus cannot be used for short-term confirmation of clearance. PCR based on urease or 16S rRNA genes has been reported for diagnosis of H. pylori in human clinical studies and animal isolates (3, 10, 22) but not in animal models. Although PCR is a very powerful technique in terms of sensitivity and specificity, it is limited in terms of ability to perform accurate quantitation (11, 21). There remains a need for a sensitive and specific detection method, useful for small animal models of Helicobacter disease, that has a good positive correlation with human clinical therapeutic data and accurate quantitation of bacterial load in animals undergoing therapeutic and prophylactic treatment. In this paper, we present a 16S rRNA gene PCR methodology in an H. felis viral antibody-free (VAF) Swiss-Webster (SW) CV mouse model with increased sensitivity and specificity for detection over currently utilized methods. Also, when combined with an internal control DNA, this PCR method may be used for quantitation of tissue loads in H. felis-infected animals. This H. felis model may be useful for screening experimental H. pylori treatment regimens or assessment of vaccines with potential for prevention or eradication of H. pylori gastritis and ulcer disease.
Since 1983, Helicobacter pylori has been established as the most common cause of type B gastritis (25). It has been reported that 90% of duodenal ulcer patients and 70% of all gastric ulcer patients are infected with H. pylori (2). Although H. pylori is susceptible to many antibiotics in vitro, it is difficult to eradicate in vivo (20, 24). At the present time, there is no adequate monotherapy for H. pylori gastritis, and approved therapy regimens require combinations of two or three antibiotics taken over a prolonged period of time (2, 17, 20). Many animal models have been developed to investigate new antimicrobial compounds or novel intervention agents before human clinical trials are possible. Current H. pylori animal models utilize rhesus monkeys (1, 5), gnotobiotic and conventional (CV) piglets (6, 19), and germfree athymic or euthymic mice (12, 13). The H. pylori animal models are generally used in small-scale experiments because they are expensive and laborand time-intensive, may need special gnotobiotic facilities, and are not practical for routine therapeutic screening. Fortunately, a small animal model of Helicobacter felis infection has raised the possibility of an alternative approach (15). H. felis is a gram-negative, spiral-shaped bacterium originally isolated from the stomachs of cats and dogs (15). This organism has been shown by 16S rRNA gene sequencing to be very closely related to H. pylori (8). Also, H. felis has been shown to colonize gnotobiotic and CV mice and elicit host reactions and histopathology which are very similar to those seen in H. pylori infections of humans (15). The detection methods currently used in Helicobacter animal models rely on a combination of culture, histological examination, detection of specific antibodies, and rapid urease-based
MATERIALS AND METHODS Animals. Six- to eight-week-old female VAF CV and male and female germfree SW mice were obtained from Taconic Laboratories (Germantown, N.Y.). VAF SW mice were maintained in microisolator cages and supplied with sterilized food and water ad libitum. Germfree SW mice were maintained in sterilized microisolators and changed in a laminar air flow hood, using aseptic procedures.
* Corresponding author. Mailing address: RY80T-100, Antibiotic Discovery and Development, Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065-0900. Phone: (908) 594-1511. Fax: (908) 5945700. 73
74
KONG ET AL.
FIG. 1. Sequences of H. felis and H. muridarum primers and comparable sequences of related bacteria. The locations and orientations of the H. felis and H. muridarum primers within the 16S rRNA gene are shown. The sequence on which each of the primers is based (underlined) is compared with the corresponding regions of the 16S rRNA genes of the closely related organisms H. pylori and C. jejuni. All sequences given correspond to the ‘‘top’’ strand. Thus, primers HF-R and HM-R each have the sequence complementary to that underlined. The numbering refers to the H. pylori 16S rRNA gene.
All animal procedures were performed in accordance with the highest standards for the humane handling, care, and treatment of research animals and were preapproved by the Merck Institutional Animal Care and Use Committee. The care and use of research animals at Merck meet or exceed all applicable local, national, and international laws and regulations. H. felis. H. felis, a gift from James Fox (Division of Comparative Medicine, Massachusetts Institute of Technology, Boston, Mass.), was cultured on H. felis medium consisting of brucella agar (BBL Microbiology System, Becton Dickinson Co., Cockeysville, Md.) supplemented with 5% fetal bovine serum (GIBCO BRL, Grand Island, N.Y.), 5 mg of amphotericin B (Fungizone; E. R. Squibb, Princeton, N.J.) per liter, 6 mg of trimethoprim (Sigma Chemical Co., St. Louis, Mo.) per liter, and 5 mg of vancomycin (Sigma) per liter. Initially, H. felis was grown in batch cultures, aliquoted, and frozen to be used as a stock source for all experiments. Cultures were incubated under microaerophilic conditions in a GasPak jar with the CampyPak Plus system (BBL) at 378C for 3 to 5 days. H. felis cells were collected from the agar surface with sterile cotton swabs and resuspended in sterile saline. The bacterial concentration was determined by an optical density (660 nm) reading compared with a standard curve of titrated H. felis adjusted by hemacytometer cell count of H. felis in 10% buffered formalin. The concentration of the H. felis inoculum for animal challenge was adjusted to approximately 2 3 108 cells per ml. Challenge and sample collection. Mice were challenged with 0.5 ml of the H. felis inoculum (108 cells) by gavage two to three times within 1 week (at least 1 day separating each challenge). To assess infectivity at various times after challenge, animals were sacrificed by CO2 inhalation. Stomachs were excised and cut longitudinally with a sterile scalpel blade. Stomach contents were removed by washing with sterile saline, and mucosa was then separated from the stomach lining tissue by gently scraping the mucosa with the edge of a sterile microscope slide. For culture, the mucosal scrapings were spread onto H. felis medium plates which had been preequilibrated to microaerophilic conditions, and the plates were immediately placed into a microaerophilic environment for incubation as described above. For PCR analysis, samples were collected into 1.0 ml of TNE buffer (10 mM Tris, 10 mM NaCl, 10 mM EDTA), aliquoted, and stored at 2808C. For histology examination, the stomachs were collected, fixed with 10% buffered formalin, and stored in 10% buffered formalin. Then samples were sectioned and stained with hematoxylin and eosin by Microscopy Laboratories (Red Bank, N.J.). Primers for PCR. Two pairs of primers were designed to amplify two regions of the 16S rRNA genes such that each pair was specific to one or only a subset of known Helicobacter species. The location and orientation of the primers are indicated in Fig. 1. The GenBank accession numbers for the sequences used are M57398 (H. felis), U00679 (H. pylori), M80205 (H. muridarum), and L04315
CLIN. DIAGN. LAB. IMMUNOL. (Campylobacter jejuni). The sequence of H. felis primer HF-L was 59-ATGA CATGCCCTTTAGTTTGGGATAGCCA-39, and that of primer HF-R was 59CGTTCACCCTCTCAGGCCGGATACC-39. This primer set amplified a 169-bp fragment and was used for detection of H. felis in samples. Since H. muridarum had been isolated from mice (16), it had to be ruled out as a possible contaminant to verify the H. felis PCR results. The specific H. muridarum primers (control primers) were designed by comparing published 16S rRNA gene sequences of H. pylori, H. felis, H. muridarum, H. mustelae, Gastrospirillum hominis, C. jejuni, Proteus spp., and Escherichia coli. The sequence areas showing the least homology between species were identified and selected for H. muridarum primers (Fig. 1). The sequences of the control primers, which were specific for H. muridarum, are 59-GATATGGGAGTGCCACTTCTG-39 for HM-L and 59AGAGATTTGCTCCACCTCACGA-39 for HM-R. This primer set amplified a 277-bp fragment and was used for testing mouse samples which were positive in H. felis PCRs to ensure that none of the positive results were caused by H. muridarum. DNA extraction. The stomach mucosa (500 ml) was pelleted (5 min, 5,000 3 g) and resuspended in TNE buffer (570 ml) containing 1% Triton X-100 (Sigma) and 0.5 mg of lysozyme (Sigma) per ml in a 1.5-ml Eppendorf tube. The sample was well mixed and incubated at 378C for 30 min. To this was added proteinase K to a final concentration of 1 mg/ml (Boehringer GmbH, Mannheim, Germany), and the mixture was incubated at 658C for 2 h or at 378C overnight. DNA was isolated by extracting twice with an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1) and once with chloroform. DNA was precipitated with 0.3 M sodium acetate and 2 volumes of absolute ethanol and placed on dry ice for 20 min. DNA was then pelleted by centrifugation as described above, rinsed with 70% ethanol, and dried in a speed vacuum. The DNA pellet was dissolved in 50 ml of 0.13 TE buffer (13 TE is 10 mM Tris [pH 7.4], 0.1 mM EDTA) and stored at 48C. PCR amplification. Master mixtures were made in a laminar flow biological safety cabinet (LabGard) with pipettes reserved specifically for this purpose, filter pipette tips, and single-use aliquots of all reaction components. The reaction mixture for each PCR contained PCR master mixture (23 ml) and test sample (2 ml). One reaction mixture consisted of 2.5 ml of Taq 103 buffer (Stratagene Cloning Systems, La Jolla, Calif.), 200 mM deoxynucleoside triphosphate (Boehringer), 1 mM (each) primer, and 0.625 U of Taq polymerase (Stratagene). The master mix was dispensed at 23 ml per 200-ml PCR tube, which was placed in a 96-well tray. To each tube, 2 ml of test sample was added to bring the reaction volume to a total of 25 ml. The tray was briefly centrifuged to bring down the reagents and then placed in a thermal cycler (Perkin-Elmer 9600 system; The Perkin-Elmer Corp., Norwalk, Conn.). DNA was amplified for 35 cycles consisting of 20 s at 948C, 30 s at 558C, and 1 min at 728C. The last cycle was identical except that the 728C elongation time was extended to 10 min. Positive and negative control reactions, consisting of a tube of deionized H2O, tubes containing Helicobacter DNA in dilutions representing 1, 10, and 100 cells, and a tube of mouse tissue DNA (;2 mg) as templates, were performed with each PCR run. The PCR products were analyzed by 2% agarose gel electrophoresis with 50 mM ethidium bromide and a 100-bp DNA ladder (GIBCO) and visualized under UV light. Internal control DNA. To confirm a negative result in a PCR assay of a tissue sample or to quantitate tissue load of H. felis by competition, an internal control template was constructed. Pairs of synthetic oligonucleotides were annealed, 59-ATAACATGCCCTTTAGTTTGGGATAGCCA-39 with 59-GGCCTGGCT ATCCCAAACTAAAGGGCATGTTATAGCT-39 (providing EagI and SacI sticky ends) and 59-TCGAGGTATCCGGCCTGAGAGGGTGAACGGTAC-39 with 59-CGTTCACCCTCTCAGGCCGGATACC-39 (providing KpnI and XhoI sticky ends), and cloned into the vector pBluescript II SK1 (Stratagene). The upstream primer site was cloned between the EagI and SacI sites, and the downstream primer site was cloned between the KpnI and XhoI sites in the polylinker. Between these two regions of homology to the 16S gene, the 346-bp EcoRI-BamHI fragment of pBR322 was cloned. The resulting plasmid, H. felis internal plasmid (pHfe-IC), produces a 400-bp PCR product with the same primers as used for the H. felis 16S RNA gene, but the product can be easily distinguished from the actual 16S PCR product (169 bp) on agarose gel electrophoresis. Competitive PCR. Competitive PCR was employed for semiquantitative analysis of H. felis in infected tissue. This technique determines quantitation based on relative amounts of products produced in a reaction mixture containing both target sequence DNA and an introduced competitive internal control DNA. The H. felis target DNA and internal control DNA were amplified with the same H. felis primers. The PCR products, 400 bp for the internal control DNA (pHfe-IC) and 169 bp for H. felis, were differentiated on agarose-ethidium bromide gels. The relative intensities of the product bands in a given gel lane for both the target and internal control were visualized. The standard curve of competitive PCR was based on the reactions which contained relative equal intensity of the two product bands visualized in different combinations of H. felis DNA and pHfe-IC. To quantitate H. felis in mouse stomach mucosa, competitive quantitative PCR with the internal control was conducted. Three types of DNA templates were used: (i) H. felis in serial log10 dilutions from 2 3 105 to 2 3 1021 cells per ml mixed with mouse tissue DNA equivalent to the amount of tissue present in mouse samples for PCR; (ii) internal control DNA (pHfe-IC) in serial log10 dilutions from 3 3 1025 to 3 3 10210 mg/ml; and (iii) H. felis-infected mouse
PCR IN H. FELIS MOUSE MODEL
VOL. 3, 1996
75
FIG. 3. Competitive quantitation with titration of H. felis target DNA and pHfe-IC internal DNA. Lane 1, 100-bp DNA ladder; lanes 2 to 17, titration of H. felis DNA from 40,000, 20,000, 4,000, 2,000, 400, 200, and 40 cells with 7 fg of pHfe-IC DNA (lanes 2 to 8) and without pHfe-IC DNA (lanes 11 to 17) in the reaction mixture for PCR; lane 9, 7 fg of pHfe-IC DNA; lane 18, deionized H2O; lane 19, 100-bp DNA ladder. FIG. 2. Sensitivity of H. felis primers. (a) Sensitivity of H. felis PCR. Lane 1, 100-bp DNA ladder; lanes 2 and 3, deionized H2O negative control; lanes 4 to 11, H. felis DNA from 1, 2, 5, 10, 102, 103, 104, and 105 cells, respectively. (b) PCR sensitivity of H. felis DNA in the presence of mouse tissue DNA. Lane 1, 123-bp DNA ladder; lanes 2 and 19, deionized H2O control; lanes 3 and 11, 1.5 and 2.0 mg of mouse tissue DNA, respectively; lanes 4 to 10, 1.5 mg of mouse tissue DNA with H. felis DNA from 1, 2, 5, 10, 50, 100, and 500 cells, respectively; lanes 12 to 18, 2.0 mg of mouse tissue DNA with H. felis DNA from 1, 2, 5, 10, 50, 100, and 500 cells, respectively.
DNA in serial log10 dilutions. PCRs were set up in checkerboard combination, titrating two of the three DNA templates. DNA from pHfe-IC and H. felis or pHfe-IC and mouse samples was tested. Titrations of DNA from H. felis-spiked mouse tissue and pHfe-IC were used for generation of a standard curve of competitive points. Titrations of DNA from pHfe-IC and infected mouse tissue were used to quantitate the mucosal tissue load of H. felis by comparison to the standard curve. Therapy studies. To test the model’s application, single- and triple-antibiotic therapy studies were conducted. All drugs were solubilized in Sorensen’s buffer (pH 8.0). Amoxicillin (Goldline Laboratories, Ft. Lauderdale, Fla.) at a concentration of 1.5 mg/ml was used in single therapy. For triple-antibiotic therapy, the solution contained amoxicillin (1.5 mg/ml), metronidazole (1.35 mg/ml; Sigma), and bismuth subsalycilate (0.16 mg/ml; Pepto Bismol solution; Proctor and Gamble, Cincinnati, Ohio). The drugs were stored at 48C and used within 1 week of preparation. For oral dosing, mice received 0.5 ml of drug suspension by gavage. Therapy was initiated 4 weeks after oral challenge of the mice with H. felis. Studies included an efficacy comparison of amoxicillin single-antibiotic therapy with a triple-antibiotic therapy regimen (amoxicillin-metronidazole-bismuth). The H. felis-colonized mice were given either amoxicillin alone or a combination of amoxicillin, metronidazole, and bismuth orally once a day for 14 days. A saline-treated group was included as a control. The mice were sacrificed and stomach samples were evaluated by H. felis PCR at various times after termination of therapy to determine colonization reduction and relapse of infection.
RESULTS PCR specificity and sensitivity. To test the sensitivity of the PCR assay, DNA extracted from pure culture of H. felis and DNA from mouse stomach mucosa spiked with H. felis cells were tested by PCR. The results indicated that the PCR could detect DNA representing as little as 1 cell in a pure culture of H. felis (Fig. 2a) but not 0.1 cell (data not shown). In the presence of 2 mg of DNA extracted from mouse stomach mucosa (corresponding to about 1/50 to 1/100 of a whole stom-
ach sample), 2 cells of H. felis could be detected, which is equivalent to 50 to 100 H. felis cells per stomach (Fig. 2b). To identify possible cross-reactions, DNAs extracted from mouse stomach tissue and bacteria including H. pylori, H. felis, H. muridarum, C. jejuni, E. coli, and Proteus vulgaris were tested by PCR with the H. felis and H. muridarum primers. The H. felis primers amplified DNA from H. pylori, H. felis, H. muridarum, and the H. felis internal control template (pHfeIC) but not DNA from the non-Helicobacter species or other unidentified isolates found in CV mouse stomach microflora (Table 1). The minimum levels of detection by PCR with the H. felis primers were 1 cell for H. felis, 10 cells for H. muridarum, and 100 cells for H. pylori. The H. felis primers did not amplify DNA from uninfected mouse stomach mucosa. The H. muridarum-specific primers amplified only DNA from H. muridarum. Infected stomach mucosa from VAF CV mice which were positive by H. felis PCR were randomly selected and tested with the H. muridarum primers. All H. felis PCR-positive samples tested were negative by H. muridarum PCR. Competitive PCR quantitation. Competitive PCR between H. felis DNA and pHfe-IC is shown in Fig. 3. Due to the competition for primers, the sensitivity of PCR for target DNA was lower when the template combinations of H. felis and its competitor, pHfe-IC, were used in competitive PCR than when H. felis template alone was used in noncompetitive PCRs (as shown in Fig. 3). Table 2 shows the standard curve generated from the signal density comparison of pHfe-IC DNA and the H. felis DNA PCR product bands on agarose-ethidium bromide gels. From a checkerboard titration, relatively equal signals were produced in the reactions containing 4 3 104 H. felis cells plus 7 3 101 pg of pHfe-IC DNA, 4 3 103 H. felis cells plus 7 3 100 pg of pHfe-IC DNA, 4 3 102 H. felis cell DNA plus 7 3 1021 pg of pHfe-IC DNA, and so on. These reactions formed a linear range of equal signal density between H. felis and pHfe-IC from 4 to 4 3 104 for H. felis cells or 7 fg to 70 pg for pHfe-IC DNA.
TABLE 2. Competitive PCR quantitation: standard curve of quantitationa
TABLE 1. Primer specificity test Amplification with primer seta Template
H. felis E. coli Proteus sp. C. jejuni H. muridarum H. pylori H. felis internal DNA
H. felis
H. muridarum
1 (1 cell) 2 2 2 1 (100 cells) 1 (10 cells) 1
2 2 2 2 1 (1 cell) 2 2
a 1, positive; 2, negative. Value in parentheses indicates minimum number of cells detectable in PCR.
Internal control DNA (pg) H. felis DNA (no. of cells) 7 3 101 7 3 100 7 3 1021 7 3 1022 7 3 1023 7 3 1024
4.0 4.0 4.0 4.0 4.0 4.0 4.0
3 3 3 3 3 3 3
105 104 103 102 101 100 1021
1 1/p p p p p p
1 1 1/p p p p p
1 1 1 1/p p p p
1 1 1 1 1/p p p
1 1 1 1 1 1/p p
1 1 1 1 1 1 1
a 1, H. felis signal . pHfe-IC signal; p, pHfe-IC signal . H. felis signal; 1/p, equal-density signals of H. felis and pHfe-IC DNA.
76
KONG ET AL.
CLIN. DIAGN. LAB. IMMUNOL.
FIG. 4. Histology of stomach section from VAF Swiss mouse 4 weeks post-challenge, stained with hematoxylin and eosin, showing H. felis in crypts (a) (31,000) and evidence of gastritis (b) (3400).
PCR IN H. FELIS MOUSE MODEL
VOL. 3, 1996
77
TABLE 3. Competitive PCR quantitation: quantitation of H. felis in infected stomach mucosa tissue samplea Infected tissue (DNA dilution)
1.0 1.0 1.0 1.0 1.0
3 3 3 3 3
100 1021 1022 1023 1024
Internal control DNA (pg) 7 3 10
7 3 1021
7 3 1022
7 3 1023
1 1 1/p p p
1 1 1 1/p p
1 1 1 1 1/p
1 1 1 1 1
0
a 1, H. felis signal . pHfe-IC signal; p, pHfe-IC signal . H. felis signal; 1/p, equal-density signals of H. felis and pHfe-IC DNA.
To avoid PCR variation due to reagents or assay conditions, pools of infected mouse samples, randomly selected, were tested at the same time with the standard for quantitation of H. felis tissue load by competitive PCR. The titration reactions which produced equal-density bands for infected stomach samples and pHfe-IC were the PCR tubes containing a 1:10-diluted sample DNA plus 70 pg of pHfe-IC DNA, 1:100-diluted sample DNA plus 7 pg of pHfe-IC DNA, and so on (Table 3). In the PCR, 7 pg of pHfe-IC DNA showed equal competition with 1:100-diluted infected mouse stomach tissue samples corresponding to 4 3 103 H. felis cells in the standard curve. Therefore, the test sample contained approximately 105 H. felis cells. Since the undiluted sample represented 1/50 to 1/100 of the whole stomach sample, the tissue load of H. felis was estimated to be between 5 3 106 and 1 3 107 cells per whole mouse stomach. H. felis mouse model and therapy. To evaluate the infectivity of H. felis, VAF SW mice were orally challenged and mouse stomach mucosa samples were analyzed for H. felis by PCR at various times after infection (Table 4). The percentage of mice positive for H. felis by PCR at all time points ranged from 80 to 100%. In two long-term studies, 100% of mice sampled were positive for H. felis by PCR at 6 months and 1 year postchallenge. H. felis infection in mice was confirmed by culture reisolation of the organism from germfree mice orally challenged with H. felis. At 20 min, 1 week, and 5 weeks after oral challenge, the stomach mucosae were collected and plated onto prereduced H. felis agar plates. H. felis was successfully isolated at all three time points, and identification was confirmed by Gram stain, urease, oxidase, and H. felis PCR. In each case, the reisolation and PCR results of the infected stomach mucosa were in agreement. Histological examination of mouse stomach mucosa 4 weeks
TABLE 4. PCR detection of infectivity of H. felis in Swiss micea Time postchallenge (wk)
Sample size (no. of mice)
Infectivity (% PCR positive)
1 2 3 4 5 10 26 52
5 5 5 10 5 .100 5 10
100 100 80 100 100 99 100 100
a The number of mice tested at 10 weeks postchallenge resulted from the combined data of a time course study (10 mice) and 10 therapeutic studies (infection control, 10 mice per study). The number of mice tested at the other time points are from the infection time course study (5 or 10 mice per group).
FIG. 5. Percentage of mice PCR positive following treatment with amoxicillin or triple therapy (amoxicillin, metronidazole, and bismuth).
after challenge with H. felis showed spiral-shaped organisms in deep crypts of antral sections of stomachs when examined by hematoxylin and eosin stain (Fig. 4a). Evidence of chronic active gastritis was also present in hematoxylin-and-eosinstained sections (Fig. 4b). All stomach samples analyzed 4 weeks after challenge were positive for H. felis by PCR. Therapy. Studies with two therapeutic regimens, amoxicillin alone and amoxicillin-metronidazole-bismuth triple therapy, were conducted, and stomach mucosal tissue samples were analyzed by PCR at various times following termination of treatment (Fig. 5). Following therapy with amoxicillin alone, the percentage of mice positive for H. felis at 24 h was 10%. However, 1 week after termination of therapy, the infection relapsed and the percentage of mice positive by PCR increased to 90%. The percentage of mice positive remained high, resulting in 80, 100, and 100% PCR-positive results at 2, 3, and 4 weeks after termination of therapy, respectively. To test for possible reinfection or cross-infection during the experiment, infected and uninfected animals were allowed to cohabit in the same cage for 4 weeks. Stomach mucosal and fecal samples from test animals were examined by PCR. The results showed no cross-infection; all infected animals were PCR positive, and all uninfected animals were PCR negative (data not shown). There was no detectable H. felis in any fecal samples surveyed. Following triple-antibiotic therapy, none of the mice were positive by H. felis PCR at 24 h and later time points up to 4 weeks after termination of therapy, and no relapse of infection was detected. DISCUSSION We have shown that PCR permits the sensitive and specific detection of H. felis DNA directly from CV mouse stomach mucosa specimens. The detection limit of H. felis PCR was similar to or better than that of the H. pylori PCR methodology used in human H. pylori studies. The normal microflora and transient bacteria in the stomach of CV mice make culture of H. felis difficult. The urease assay as a Helicobacter detection method was not only insensitive, with a limit of detection of 106 cells per assay (from our unpublished data and as previously reported [27]), but also nonspecific, because other non-H. felis bacteria (such as Proteus spp.) in the stomach of CV mice are often urease producers that also give positive urease reactions. When examined by histology, similarities in morphology of Helicobacter spp. and other spiral-shaped bacteria could also
78
KONG ET AL.
CLIN. DIAGN. LAB. IMMUNOL.
be misinterpreted, and the pathologic inflammation responses observed could be a consequence of other bacteria or conditions. A common drawback of serological analysis is that it measures antibodies to H. pylori that persist long after the bacterium is eradicated by antibiotics. The value of a serological test lies in its initial detection of H. pylori antibodies, but it cannot confirm active infection or even monitor whether antibiotic treatment was successful. The enhanced sensitivity and specificity of PCR in this model allow rapid and accurate assessment of efficacy compared with other assay methods. We have also shown that the PCR methodology can be used for semiquantitation when combined with an internal DNA standard. In competitive PCR, each amplicon (target and internal DNA) accumulates in a constant ratio even in the plateau phase of amplification, and all calculations are based on the relative quantities of coamplified material. Therefore, the competitive PCR described here is semiquantitative and nonisotopic and can be widely applied to assessment of the quantity of DNA present in testing samples. Since the detection limit is 100 cells per mouse stomach and the tissue load of H. felis is 106 cells per stomach, there is a large window for the sensitive PCR assay to evaluate therapy regimens, which may be more predictive of organism reduction, eradication, or possible relapse of infection following termination of antibacterial therapy. Modification of this methodology may have application in human clinical testing for H. pylori therapeutic analysis. The therapy studies here examined the capabilities of the PCR methodology to assess antimicrobial activity of amoxicillin monotherapy and triple-antibiotic therapy in our H. felis mouse model (Fig. 5). Our PCR results of successful eradication of organisms following triple-antibiotic therapy, but not amoxicillin monotherapy, are similar to those reported for human clinical trials of H. pylori therapy (17, 20). The tripletherapy results are also comparable to those with other H. felis mouse models (4); however, amoxicillin monotherapy showed greater efficacy in the Dick-Hegedus and Lee mouse model than in our PCR model (67 versus 0% efficacy). The difference between the monotherapy results in the two models may be due to the increased sensitivity of PCR analysis for detection compared with culture and histology. Since possible transmission of infection from infected mice to uninfected animals was ruled out, H. felis detectable following amoxicillin monotherapy was due to bacterial rebound. This suggests that monotherapy with amoxicillin actually only reduced H. felis to levels below the limits of detection during and immediately after the termination of therapy (24 h). The results of rebound from unsuccessful treatment were similar to those reported in humans (18). These findings suggest that the PCR technique, when used in the VAF CV SW mouse model of H. felis, is a useful model for evaluating and comparing potential antimicrobial therapies or other modalities for the eradication of Helicobacter spp. This is especially important considering that in vitro susceptibility testing has been found to be nonpredictive of therapeutic success (9). REFERENCES 1. Baskerville, A., and D. G. Newell. 1988. Naturally occurring chronic gastritis and C. pylori infection in the rhesus monkey: a potential model for gastritis in man. Gut 29:465–472. 2. Bell, G. D., and K. U. Powell. 1993. Eradication of Helicobacter pylori and its effect in peptic ulcer disease. Scand. J. Gastroenterol. Suppl. 196:7–11. 3. Clayton, C., K. Kleanthous, and S. Tabaqchali. 1991. Detection and identi-
4. 5.
6. 7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17. 18. 19.
20.
21.
22.
23.
24. 25. 26.
27.
fication of Helicobacter pylori by the polymerase chain reaction. J. Clin. Pathol. 44:515–516. Dick-Hegedus, E., and A. Lee. 1991. Use of a mouse model to examine anti-Helicobacter pylori agents. Scand. J. Gastroenterol. 26:909–920. Dubois, A., N. Fiala, L. M. Heman-Ackah, E. S. Drazek, A. Tarnawski, W. N. Fishbein, G. I. Perez-Perez, and M. J. Blaser. 1994. Natural gastric infection with Helicobacter pylori in monkeys: a model for spiral bacteria infections in humans. Gastroenterology 106:1905–1917. Eaton, K. A., D. R. Morgan, and S. Krakowka. 1990. Persistence of Helicobacter pylori in conventionalized piglets. J. Infect. Dis. 161:1922–1301. Eckloff, B. W., R. P. Podzorski, B. C. Kline, and F. R. Cockerill III. 1994. A comparison of 16S ribosomal DNA sequences from five isolates of Helicobacter pylori. Int. J. Syst. Bacteriol. 44:320–323. Fox, J. G., L. L. Yan, F. E. Dewhirst, B. J. Paster, B. Shames, J. C. Murphy, A. Hayward, J. C. Belcher, and E. N. Menes. 1995. Helicobacter bilis sp. nov., a novel Helicobacter species isolated from bile, livers, and intestines of aged, inbred mice. J. Clin. Microbiol. 33:445–454. Goodwin, C. S., E. Blincow, G. Peterson, C. Sanderson, W. Cheng, B. Marshall, and R. McCulloch. 1987. Enzyme-linked immunosorbent assay for Campylobacter pyloridis: correlation with presence of C. pyloridis in the gastric mucosa. J. Infect. Dis. 205:484–488. Ho, S. A., J. A. Hoyle, F. A. Lewis, A. D. Secker, D. Cross, N. P. Mapstone, M. F. Dixon, J. I. Wyatt, D. S. Tompkins, G. R. Taylor, and P. Quirke. 1991. Direct polymerase chain reaction test for detection of Helicobacter pylori in humans and animals. J. Clin. Microbiol. 29:2543–2549. Kanangat, A., A. Solomon, and B. T. Rouse. 1992. Use of quantitative polymerase chain reaction to quantitate cytokine messenger RNA molecules. Mol. Immunol. 29:1229–1236. Karita, M., Q. Li, D. Cantero, and K. Okita. 1994. Establishment of a small animal model for human Helicobacter pylori infection using germ-free mice. Am. J. Gastroenterol. 89:208–213. Karita, M., Q. Li, and K. Okita. 1993. Evaluation of new therapy for eradication of H. pylori infection in nude mouse model. Am. J. Gastroenterol. 88:1366–1372. Karita, M., M. Tada, K. Okita, M. Tsuda, and T. Nakazawa. 1993. Evaluation of methylene blue and triple therapy for eradication of Helicobacter pylori infection in the nude mouse model. Eur. J. Gastroenterol. Hepatol. 5:S79–S83. Lee, A., J. G. Fox, G. Otto, and J. Murphy. 1990. A small animal model of human Helicobacter pylori active chronic gastritis. Gastroenterology 99:1320– 1323. Lee, A., M. W. Philips, J. L. O’Rourke, B. J. Paster, F. E. Dewhirst, G. J. Fraser, J. G. Fox, L. I. Sly, P. J. Romaniuk, T. J. Trust, and S. Kouprach. 1992. Helicobacter muridarum sp. nov., a microaerophilic helical bacterium with a novel ultrastructure isolated from the intestinal mucosa of rodents. Int. J. Syst. Bacteriol. 42:27–36. Malfertheiner, P. 1993. Compliance, adverse events and antibiotic resistance in Helicobacter pylori treatment. Scand. J. Gastroenterol. Suppl. 196:34–37. Rauws, E. A., and G. N. Tytgat. 1990. Cure of duodenal ulcer associated with eradication of Helicobacter pylori. Lancet 335:1233–1235. Rudmann, D. G., K. A. Eaton, and S. Krakowka. 1992. Ultrastructure study of Helicobacter pylori adherence properties in gnotobiotic piglets. Infect. Immun. 60:2161–2164. Tytagat, G. N. J. 1994. Treatments that impact favorably upon the eradication of Helicobacter pylori and recurrence. Aliment. Pharmacol. Ther. 8:359– 368. Ursi, J., D. Ursi, M. Leven, and S. R. Pattyn. 1992. Utility of internal control for the polymerase chain reaction. Acta Pathol. Microbiol. Immunol. Scand. 100:635–639. Valentine, J. L., R. R. Arthur, H. L. Mobley, and J. D. Dick. 1991. Detection of Helicobacter pylori by using the polymerase chain reaction. J. Clin. Microbiol. 29:689–695. Von Wuffen, H., J. Heesemann, G. H. Butzow, T. Loning, and R. Laufs. 1986. Detection of Campylobacter pyloridis in patients with antrum gastritis and peptic ulcers by culture, complement fixation test, and immunoblot. J. Clin. Microbiol. 24:716–720. Wagner, S., M. Gebel, and M. Manns. 1993. Therapy of Helicobacter pylori infection: current status. Z. Gastroenterol. 31:459–463. Warren, J. R., and B. Marshall. 1993. Unidentified curved bacilli on gastric epithelium in chronic active gastritis. Lancet i:1273–1275. (Letter.) Weiss, J., J. Mecca, E. D. Silva, and D. Gassner. 1994. Comparison of PCR and other diagnostic techniques for detection of Helicobacter pylori infection in dyspeptic patients. J. Clin. Microbiol. 32:1663–1668. Xia, H. X., C. T. Keane, and C. A. O’Morain. 1994. Pre-formed urease activity of Helicobacter pylori as determined by a viable cell count technique—clinical implications. J. Med. Microbiol. 40:435–439.
