- Email: [email protected]
Development and application of a novel screening PCR assay for direct detection of ‘Helicobacter heilmannii’-like organisms in human gastric biopsies in Southeast England
Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
www.elsevier.com/locate/diagmicrobio
Bacteriology
Development and application of a novel screening PCR assay for direct detection of ‘Helicobacter heilmannii’-like organisms in human gastric biopsies in Southeast England Stephanie A. Chisholm*, Robert J. Owen Laboratory of Enteric Pathogens, Central Public Health Laboratory, London NW9 5HT United Kingdom Received 6 August 2002; accepted 13 November 2002
Abstract A novel PCR assay (HHLO-16) to screen for presence of ‘Helicobacter heilmannii’-like organisms (HHLO) direct from gastric biopsies is described. As ‘H. heilmannii’ is generally uncultivable, diagnosis of infection is reliant on histology; thus prevalence may be underestimated. Analysis of an HHLO histology-positive human gastric biopsy and 15 gastric biopsies from domestic cats demonstrated that the HHLO-16 assay was more sensitive than an alternative available species-specific PCR assay. Further testing of 131 gastric biopsies from dyspeptic patients demonstrated an HHLO prevalence rate of 2.3% in Southeast England. Subsequent combination of the HHLO-16 assay with a H. pylori-specific PCR assay in a multiplex format (HpHh assay), and repeat analysis of the 131 biopsies showed the HpHh assay was as sensitive as each individual test. This novel PCR assay provides simple concomitant testing of dyspeptic patients for both HHLOs and H. pylori, thereby rapidly identifying individuals requiring eradication therapy. © 2003 Elsevier Inc. All rights reserved. Keywords: ‘Helicobacter heilmannii-like organisms’; Helicobacter pylori; Multiplex PCR assay; Biopsy; Detection
1. Introduction Infections with ‘Helicobacter heilmannii’, previously named ‘Gastrospirillum hominis’ in man (McNulty et al., 1989), are found in a wide range of other animal hosts that include cats, dogs and pigs with reported prevalence rates of 78-100% (De Groote et al., 1999; Eaton et al., 1996; Neiger et al., 1998). Although the nomenclature implies a defined species, 16S rDNA phylogenetic analyses demonstrate that ‘H. heilmannii’ comprises two related but distinct types (O’Rourke et al., 2001; Solnick et al., 1993). Type 1 is closely related to ‘Candidatus H. suis’, found in pigs (De Groote et al., 1999), whereas Type 2 is more closely related to the H. bizzozeronii/H. felis/H. salomonis subgroup that most commonly infect cats and dogs (Jalava et al., 1997; Solnick et al., 1993). These morphologically complex, often non-culturable bacteria, from various human and animal hosts represent part of an ill-defined taxonomic group of
* Corresponding author. Tel.: ⫹44-20-8200-4400; fax: ⫹44-20-89059929. E-mail address: [email protected] (S.A. Chisholm). 0732-8893/03/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0732-8893(02)00547-3
closely related forms which, for convenience, have been referred to as ‘Helicobacter heilmannii’-like organisms (HHLO). Human HHLO infections possibly occur via a zoonotic route, and may lead to the development of peptic ulcer disease (Hilzenrat et al., 1995; Jhala et al., 1999) or low-grade mucosal associated lymphoid tissue (MALT) lymphoma (Morgner et al., 2000). While infections with H. pylori, the well established human gastric pathogen, are world-wide and prevalence rates exceed 40% in many countries, the reported incidence of ‘H. heilmannii’ infection based on histology is low (0.12%) in dyspeptic adult (Ierardi et al., 2001; Jhala et al., 1999; Svec et al., 2000) and pediatric patients (Mention et al., 1999; Svec et al., 2000). As HHLOs from human tissue remain uncultivable with rare exceptions (Andersen et al., 1996), diagnosis is primarily reliant on histology, although touch cytology is also described (Debongnie et al., 1994). Recently, another microscopy-based system, fluorescent in situ hybridization (FISH), was described that detected all known types of ‘H. heilmannii’ and several new forms direct from paraffin embedded gastric biopsies (Trebesius et al., 2001). PCR-based studies have identified HHLOs in both human (Morgner et al., 2000) and animal (Norris et al.,
2
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
1999) gastric biopsies by amplification of Helicobacter genus-specific fragments of 16S rDNA followed by sequencing. One PCR detection assay, based on the ureB gene was described for detection of ‘H. heilmannii’ in domestic cat biopsies (Neiger et al., 1998), but specific PCR has only been applied for detection of human infection in a single patient (Dieterich et al., 1998). A number of histologically based studies have provided valuable information on the clinical features of HHLO infections (Holck et al., 1997; Jhala et al., 1999; Mention et al., 1999; Stolte et al., 1994) and have suggested that when compared with H. pylori, HHLO distribution tends to be focalised in the antrum (Jhala et al., 1999) as well as being patchy and sparse (Holck et al., 1997). Microscopy-based methods of detection can be relatively insensitive when bacterial numbers are low (Holck et al., 1997; Megraud 1995). In the original case report of this infection, it was stated that slides had to be examined for at least five minutes, to minimize possible false negative results (McNulty et al., 1989). It is thus possible that the incidence of human HHLO infection may be under-reported. The aim of this study was to develop a novel PCR-based assay to screen for HHLOs direct from human gastric biopsies to facilitate enhanced surveillance in England. The PCR test was then combined with a H. pylori-specific PCR assay in a multiplex format for rapid detection of both HHLO and H. pylori infections of the human gastric mucosa.
2. Materials and methods 2.1. Bacterial isolates and culture conditions Twenty-two strains of Helicobacter representing 12 species, mostly obtained as lyophilized cultures from the National Collection of Type Cultures (NCTC, London) were included in this study. These were H. pylori (NCTC 11637T and NCTC 11638, strains 26695, J99, and H566), H. fennelliae (NCTC 11612T and NCTC 11613), H. cinaedi (NCTC 11611 and NCTC 11614), H. pullorum (NCTC 12825 and NCTC 12826), H. canis (NCTC 12739T and NCTC 12745), H. acinonychis (NCTC 12688 and NCTC 12689), H. felis (NCTC 12436T), H. hepaticus (NCTC 12886T), H. nemestrinae ⫽ H. pylori (NCTC 12491T), H. muridarum (NCTC 12714T), H. pametensis (NCTC 12888 and NCTC 12889) and H. mustellae (NCTC 12032); T indicates the type strain of the species. Isolates were cultured on Columbia agar base (Oxoid) containing 10% (v/v) defibrinated horse blood under microaerobic conditions (4% O2, 5% CO2, 5% H2, 86% N2) in a variable atmospheric incubator (VAIN) (Don Whitley Scientific Ltd, Shipley, UK) at 37°C for 72h.
2.2. Clinical specimens Gastric biopsies were from 208 patients attending open access endoscopy clinics for routine investigation of dyspeptic symptoms. Human antral gastric biopsies included in this study comprised: 131 biopsies collected prospectively (Broomfield Hospital, Chelmsford and Kings College Hospital, London) for which H. pylori status had been previously determined by culture and by histologic examination of paired biopsies, this group included one biopsy for which there was histologic evidence of an HHLO infection (A. Price, S. M. Dobbs and R. J. Dobbs, personal communication); a stored (⫺20°C) collection of 39 biopsies known to be H. pylori-positive by culture and histology (Peters et al., 1997); and 38 biopsies, confirmed as H. pylori-positive at the time of endoscopy by the rapid urease (CLO) test (North Middlesex University Hospital, London). Biopsies in the latter group were retained in batches at the hospital and periodically forwarded to the laboratory at ambient temperature in CLO test strips for culture and molecular testing. Biopsies from Chelmsford PHL, were frozen in DENTS transport medium (⫺20°C) after primary culture, and subsequently transported frozen to the laboratory for molecular analyses. Gastric biopsies from 15 domestic cats under investigation for various symptoms at a veterinary clinic in north London were collected (unfrozen), cultured for HHLOs and other Helicobacters, examined microscopically by Gram stain and then stored (⫺20°C) until required for further analysis. 2.3. DNA extraction Genomic DNA was extracted as described previously from all gastric biopsies and from bacterial cultures (Marais et al., 1999; Chisholm et al., 2001). All DNA extracts were stored at ⫺20°C until required. 2.4. PCR-based detection of HHLOs A novel PCR assay (HHLO-16) targeting 16S rDNA to detect HHLOs was designed based on in silico alignments in GeneBase version 1.0 (Applied Maths, Kortjivik, Belgium) of 51 sequences of 16S rDNA genes from 19 species of Helicobacter held in GenBank that included: ‘H. heilmannii’ (accession numbers AF25625, AF058777, AF058770, AF058771, Y18028, AF058773, AF058772, AF058774, AF058775, AF058776, L10079, L10080), H. felis (M57398, M37643, U51872, U51871, U51870, AF103882, AF103881, AF 103880, AF103879), H. bizzozeronii (Y09404, AF103883, AF302107), H. salomonis (U89351) and Candidatus H. suis (AF127028). A 112 bp fragment of the 16S rRNA gene was then amplified from 100 ng culture-extracted DNA or from 5 L biopsy DNA extract in a reaction mix (50 L final volume) containing 200 M (each) deoxynucleoside trisphosphates
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
3
Fig. 1. PCR products generated by ‘H. heilmannii’ specific (*H) and by H. felis specific (†F) ureB assays (Neiger et al., 1996) and by assay HpHh (‡M) amplifying 16S rDNA and vacA fragments of HHLOs and H. pylori respectively. Human biopsy: negative for ‘H. heilmannii’ (ureB) (lane 1), and H. felis (ureB) (lane 3) but positive by histology and by HpHh assay (lane 5). Feline biopsy positive by both ureB assays (lanes 2 & 4) and by HpHh assay (lane 6); Lanes 7 & 8, HHLO positive human biopsies; Lanes 9 & 10, H. pylori positive human biopsies; Lanes 11 & 12, HHLO positive biopsies spiked with H. pylori DNA; Lane 13, 123-bp size marker.
(Invitrogen, UK), 0.4 M (each) primer HeilF (5⬘ AAG TCG AAC GAT GAA GCC TA 3⬘) and HeilR (5⬘ ATT TGG TAT TAA TCA CCA TTT C 3)’ (MWG Biotech Ltd, Germany), and 1 U of Taq polymerase (Invitrogen) in 20 mM Tris-HCl, pH 8.4, 1.5 mM MgCl2, 50 mM KCl, 0.2% (v/v) glycerol. Target DNA was denatured at 95°C for 5 min then amplified by 35 cycles of 1 min each at a denaturation temperature of 95°C, an annealing temperature of 53°C and an elongation temperature of 72°C, extension was completed by a 5-min incubation at 72°C. PCR products were electrophoresed (100V/1hr) in a 1% w/v agarose gel (Ultra Pure, Invitrogen) in TBE buffer (90 mM Tris-HCl, 90 mM boric acid, 2.0 mM EDTA), and ethidium bromide stained (0.5 g/ml). In addition, two separate species-specific PCR assays amplifying the urease B subunit (ureB) gene were applied to gastric biopsies to detect ‘H. heilmannii’ and H. felis following the protocols described previously (Neiger et al., 1998). 2.5. Multiplex PCR assay (HpHh) An H. pylori-specific PCR assay targeting the vacuolating cytotoxin (vacA) gene (Chisholm et al., 2001) was combined with the new HHLO-16 assay to form the multiplex assay designated HpHh. Either vacA or 16S rDNA were amplified in a 50 L reaction containing 5 L extracted DNA, PCR buffer, 1.5 mM MgCl2, 200 M (each) dNTP, 0.6 M (each) primer HeilF and HeilR, 0.28 M (each) primer VAC3624F (5⬘-GAG CGA GCT ATG GTT ATG AC-3⬘) and VAC4041R (5⬘-CAT TCC TAA ATT
GGA AGC GAA-3⬘) (MWG Biotech Ltd) and 1 U Taq polymerase (Invitrogen). Following denaturation at 95°C (5 min), specific targets were amplified in the reaction mix through 35 cycles as follows: 94°C for 30 sec, 54°C for 30 sec and 72°C for 45 sec each; extension was completed at 72°C for 5 min. Aliquots of each PCR product were electrophoresed in a 2% w/v agarose gel in TBE buffer and stained in ethidium bromide.
3. Results 3.1. Assessment of sensitivity and specificity of the HHLO16 assay Analysis of genomic DNA from 12 species of Helicobacter demonstrated that the HHLO-16 assay did not amplify the 112-bp 16S rDNA fragment from H. pylori or any species of Helicobacter other than H. felis. The latter result was consistent with the findings of a GenBank BLAST search that indicates the HHLO-16 assay would theoretically detect members of the H. felis, H. bizzozeronii, H. salomonis subgroup (⫽‘H. heilmannii’ type 2). However, DNA from the latter two species was not available for testing. Application of the HHLO-16 assay to a human gastric biopsy for which there was histologic evidence of an HHLO infection generated the expected 112-bp product (Fig. 1). Sensitivity of the HHLO-16 assay was assessed by application to gastric biopsies from domestic cats, where the
4
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
Table 1 Comparison of performances of the novel HHLO-16 assay, and two ‘H. heilmannii’* or H. felis* specific ureB PCR assays when applied to human and to feline gastric biopsies Gastric biopsy
PCR Assay
(n)
HHLO-16 ⫹ (†)
Human (113) Feline (15)
3 (1) 13 (2)
⫺ 110 2
formation, such as contact with animals, was available for these patients. 3.3. Evaluation of HpHh assay for detection of HHLOs and H. pylori
‘H. heilmannii’*
H. felis*
⫹(†)
⫹(†)
⫺
‡
na 14
⫺
1 (0) 13 (2)
112 2
na 1 (1)
The multiplex HpHh assay was evaluated initially on DNA from nine gastric biopsies. The H. pylori-specific vacA PCR product was generated from DNA of the five H. pylori positive biopsies tested, while specific fragments of 16S rDNA were amplified in all four DNA extracts from HHLO-positive biopsies of both human (n ⫽ 2) and feline (n ⫽ 2) origin. In addition, analysis of DNA from the HHLO-positive biopsy spiked with 100 ng H. pylori DNA showed that each target was successfully co-amplified (Figure 1). Application of the two PCR assays independently demonstrated that for biopsies previously tested by culture and histology, 59/170 were H. pylori and 3/170 were HHLO PCR-positive, respectively, while vacA and 16S rDNA was amplified in 27/38 and 0/38 CLO test positive biopsies, respectively (Table 2). Subsequent repeat testing of human biopsies demonstrated that both assays in a multiplex format were as sensitive as the individual PCR assays (Table 2). However the HpHh test gave one false-negative as 12/15 feline gastric biopsies were HHLO-positive compared with 13/15 positive by single testing with the HHLO-16 assay (Table 2).
* Published assays targeting the urease B subunit (ureB) (Neiger et al., 1998) † Number confirmed microscopically (histology or Gram stain) ‡ H. felis assay not applied to human gastric biopsies
incidence of infection with HHLOs was expected to be high according to reported prevalence rates (Norris et al., 1999). The HHLO-16 assay generated a specific product in 13/15 feline biopsies, two of which were confirmed as containing spiral organisms by Gram stain (Table 1). Further analyses of the cat biopsies with the two ureB PCR assays of Neiger et al. (1998) demonstrated that all 13 of these were positive for ‘H. heilmannii’ (⫽type 2) while only one biopsy was also positive for H. felis (Table 1). Spiral organisms were observed in that particular biopsy by Gram staining and microscopy. 3.2. Determination of prevalence of infection with HHLOs in man
3.4. Investigation of the frequency of H. pylori/HHLO co-infections
Application of the HHLO-16 assay to DNA extracted from biopsies, collected prospectively from dyspeptic patients in mid-Essex (England) over a 3-month period in 1999, generated specific fragments of 16S rDNA from 3/131 biopsy extracts giving a prevalence rate of 2.3%. There did not appear to be any familial or geographical link between the three HHLO positive patients. No further in-
The HpHh multiplex PCR assay was applied to 101 human gastric biopsies that were positive for H. pylori, either by culture and/or histology (n ⫽ 63) or by CLO test (n ⫽ 38). The results of the assay showed vacA and 16S rDNA fragments in 86/101 and 0/101 samples, respectively, indicating no co-infections.
Table 2 Comparison of sensitivities of the vacA and HHLO-16 assays when applied individually and combined in a multiplex format (HpHh) PCR assays H. pylori (vacA) Gastric Biopsies
HHLO-16 (16S rRNA) % Sensitivity†
% Sensitivity
Origin (n)
Status*
⫹
⫺
Uniplex
Multiplex
⫹
Cat (15) Human (170) Human (38)
C/M ⫹ C/M ⫺ C/H ⫹ C/H ⫺ U⫹
0 0 59
0 15 4
na
na
94.0
94.0
27
11
77.5
77.5
0 13 0 3 0
12‡
⫺
Uniplex
Multiplex‡
0 2 3‡ 63 104 38
100.0
92.8
100.0
100.0
Na
na
* H. pylori status determined by culture (C), Gram stain microscopy (M), histology (H) or rapid urease (CLO) test (U). Sensitivity based on the observation that no PCR negative biopsies were histology positive. ‡ Results generated for the HpHh multiplex assay. †
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
4. Discussion A novel PCR based assay (HHLO-16) was developed to amplify fragments of the HHLO genomes directly from gastric biopsies, and its performance was compared with that of an alternative PCR assay targeting the ureB gene (Neiger et al., 1998). We aimed to develop an assay that could detect all known HHLO infections so the primers were designed by multiple alignment of 11 16S rDNA sequences of ‘H. heilmannii’ types 1 and 2, as well as Candidatus H. suis, H. bizzozeronii, H. felis and H. salomonis sequences currently held in GenBank. Application of the HHLO-16 assay to 12 different species of Helicobacter confirmed that only members of the HHLO group would be amplified. Other assays have been described for use on antral biopsies. For example, a ‘Candidatus H. suis’-specific PCR assay has been described that amplified product from porcine gastric biopsies (De Groote et al., 2000), while another assay for the H. bizzozeronii, H. felis, H. salomonis group was reported recently that successfully amplified specific product from canine gastric biopsies (De Groote et al., 2001). Our BLAST analysis of each of the primer pairs demonstrated that the former assay would amplify only ‘H. heilmannii’ type 1 sequences held in GenBank while the latter assay would amplify type 2 only. In contrast the HHLO-16 assay will amplify both types 1 and 2. As each type has been reported in the human stomach (Trebesius et al., 2001), our assay is more suitable for application to human gastric biopsies as it provides a rapid method of screening for all HHLOs in a single reaction. Comparison of the HHLO-16 assay with the ureB assay by Neiger et al. (1998) demonstrated that both assays generated product in 13/15 feline biopsies, but while three human biopsies were HHLO-16 PCR positive, only one human biopsy was positive by the ureB assay, and the histologically confirmed positive biopsy was PCR negative. This may indicate sequence variation in ureB between human and feline strains and suggests that the HHLO-16 assay is more sensitive for detection of HHLO infection in human samples. Moreover, our GenBank BLAST search of the ureB primer sequences suggest that only 4/16 strains held in GenBank would be amplified by this assay. It seems likely that this ureB assay is ‘H. heilmannii ’ type 2-specific and thus lacks the broad specificity needed for initial testing of human infections. The HHLO-16 assay gave sharper single bands than the ureB assay that often generated non-specific bands. Prospective application of HHLO-16 to 131 gastric biopsies demonstrated a HHLO prevalence rate of 2.3% in the series of patients in Southeast England. The incidence was higher than reported previously in other European countries (Ierardi et al., 2001; Mention et al., 1999) and the USA (Jhala et al., 1999; Mention et al., 1999). This observation may indicate either a higher geographical prevalence, or alternatively histology-based studies may have underestimated the true rate of infection in the particular dyspeptic
5
population examined. If the latter were true, then some patients, who would benefit from specific eradication therapy, are not being identified-thus a PCR-based approach for detection may be a more sensitive alternative to histology. Our results show that the incidence of HHLO infection is significantly lower than that of H. pylori in the study population (2.3% vs 16.8%) so it would not be cost effective to perform routine PCR tests for just HHLOs. The HHLO-16 assay was therefore combined in a multiplex format with another modified assay that amplified a fragment of the H. pylori-specific vacA gene (Chisholm et al., 2001). Evaluation of the HpHh assay on human biopsies of known H. pylori and HHLO status demonstrated that the sensitivity and specificity of the HpHh assay was comparable to those of the individual tests. In addition, spiking experiments demonstrated that both targets could be amplified simultaneously; thus this system could potentially detect co-infections with both HHLOs and H. pylori. Analysis of the feline biopsies demonstrated that sensitivity of the HHLO-16 assay was only slightly reduced, with one false negative result, in the multiplex format. Comparison of sensitivity of detection of serially diluted H. felis DNA (ranging from 100 ng –1 fg) for both assay formats demonstrated that both uniplex and multiplex assays detected as little as 10pg DNA, but the band intensity at the highest dilution was lower for the product generated in the multiplex format (data not shown). Thus the multiplex assay is only slightly less sensitive than the uniplex format. PCR reagent concentrations and cycling conditions were re-optimized to facilitate co-amplification of both targets, reduced efficiency of the HHLO-16 assay may have resulted for that sample from sequence variation at the primer binding sites. However, we suggest the benefit of combining the two assays in a multiplex format to provide a more diagnostically useful test outweighs the slight reduction in sensitivity. Although human H. pylori and HHLO co-infection has been identified by histology (Ierardi et al., 2001), neither the incidence nor the possible significance of this in terms of disease development has been determined. In the present study we found no evidence of co-infection. One hundred and one confirmed H. pylori-positive biopsies were tested by assay HpHh, but generation of specific HHLOs was not observed in any of the biopsies that were PCR positive for H. pylori. Given the observed low incidence of HHLO infection in the population investigated, it could be concluded that the incidence of co-infections with H. pylori is comparatively rare in Southeast England. A set of human biopsies included in this study had been identified as H. pylori-positive by rapid urease testing (CLO test). However, it should be noted that urease activity of HHLOs can also cause a CLO test color change (McNulty et al., 1989), and consequently CLO test positive results may be misinterpreted. These CLO-test biopsies were analyzed by the HpHh assay not only to investigate prevalence of co-infections but also to establish if any HHLO infections had been misdiagnosed as H. pylori positive. The sensitivity
6
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
of detection for the vacA assay was lower than had been found for the culture and/or histology-positive human biopsies tested. The possibility of false negative results due to PCR inhibitors in biopsy DNA extracts had been eliminated previously (data not shown) by an internal control PCR assay using primer pair H6A/H6B that amplifies the human mitochondrial cytochrome oxidase subunit 3 gene (Cadieux et al., 1993). As these biopsies had been transported at room temperature in the original CLO test from the primary diagnostic laboratory to our laboratory, DNA degradation may have occurred during transit-a phenomenon that we have observed previously when transporting using conventional media (Chisholm et al., 2001). While we did not identify any biopsies that were CLO test positive due to HHLO infection, the possibility of false negative results in 12 biopsies where specific target DNA may have degraded cannot be fully excluded. In addition, as densities of HHLO infections have been reported to be patchy and sparse relative to H. pylori (Holck et al., 1997; McNulty et al., 1989), it would be useful also to examine CLO tests that have generated weak or equivocal urease positive results for possible HHLO infection, however these were not available for the present investigation. A recently reported fluorescent in situ hybridization (FISH) system demonstrated that ‘H. heilmannii’ type 1 was the most prevalent in human infection, while type 2 and other novel types (3, 4 and 5) were less common, with mixed type infections also observed (Trebesius et al., 2001). Combination of our novel assay with the FISH technology would provide a comprehensive strategy for better characterization of HHLO infection in humans. The development of the HpHh assay provides a sensitive screening method to identify those biopsies positive for HHLOs that require additional, more complex testing by FISH to enable strain differentiation and characterization. A recent report that identified H. cinaedi in the gastric biopsies of two patients (Pena et al., 2002) has raised the possibility that a broader range of Helicobacter spp may be capable of colonising the human gastric mucosa than had been previously appreciated. As there is only one such report to date, further investigations aimed at detecting other Helicobacters including H. cinaedi will be necessary to establish the extent and significance of such infections at that site. In conclusion, the development and evaluation of a novel multiplex PCR assay for detection of HHLOs and H. pylori provides a means of direct detection of the principal bacterial pathogens causing gastric infection and disease in man. Routine use of this test would enable more accurate assessment of the prevalence of HHLOs in the dyspeptic population and would facilitate studies to improve understanding of the clinical significance and sources, such as domestic pets, of those infections in man, as well as enabling further investigation of HHLO/H. pylori co-infections. Future prospective analyses of gastric biopsies will now be possible to investigate larger study populations.
Acknowledgments We thank Dr E.L. Teare (Chelmsford PHL), Dr S. Saverymuttu (Broomfield Hospital, Chelmsford), Dr B. Patel (North Middlesex University Hospital, London) and Dr S.M. Dobbs and Dr R.J. Dobbs (King’s College Hospital, London) for kindly providing routine gastric biopsies. We are also grateful to Professor A. Price (Northwick Park Hospital, Harrow) for providing histologic details, and to Mrs M. Forster-Van Hijfte (The Village Veterinary Practice, Belsize Park, London) for providing gastric biopsies from domestic cats.
References Andersen, L.P., Norgaard, A., Holck, S., Blom, J., & Elsborg, L. (1996). Isolation of a “Helicobacter heilmannii”-like organism from the human stomach. Eur J Clin Microbiol Infect Dis, 15, 95–96. Cadieux, N., Lebel, P., & Brousseau, R. (1993). Use of a triplex polymerase chain reaction for the detection and differentiation of Mycoplasma pneumoniae and Mycoplasma genitalium in the presence of human DNA. J Gen Microbiol, 139, 2431–2437. Chisholm, S. A., Owen, R. J., Teare, E. L., & Saverymuttu, S. (2001). PCR-based diagnosis of Helicobacter pylori infection and real-time determination of clarithromycin resistance directly from human gastric biopsy samples. J Clin Microbiol, 39, 1217–1220. De Groote, D., Ducatelle, R., van Doorn, L. J., Tilmant, K., Verschuuren, A., & Haesebrouck, F. (2000). Detection of “Candidatus Helicobacter suis” in gastric samples of pigs by PCR: comparison with other invasive diagnostic techniques. J Clin Microbiol, 38, 1131–1135. De Groote, D., Haesebrouck, F., van Doorn, L. J., Vandamme, P., & Ducatelle, R. (2001). Evaluation of a group-specific 16S ribosomal DNA-based PCR for detection of Helicobacter bizzozeronii, Helicobacter felis, and Helicobacter salomonis in fresh and paraffin-embedded gastric biopsy specimens. J Clin Microbiol, 39, 1197–1199. De Groote, D., van Doorn, L. J., Ducatelle, R., Verschuuren, A., Haesebrouck, F., Quint, W. G., Jalava, K., & Vandamme, P. (1999). ‘Candidatus Helicobacter suis’, a gastric helicobacter from pigs, and its phylogenetic relatedness to other gastrospirilla. Int J Syst Bacteriol, 49, 1769 –1777. Debongnie, J. C., Mairesse, J., Donnay, M., & Dekoninck, X. (1994). Touch cytology. A quick, simple, sensitive screening test in the diagnosis of infections of the gastrointestinal mucosa. Arch Pathol Lab Med, 118, 1115–1118. Dieterich, C., Wiesel, P., Neiger, R., Blum, A., & Corthesy-Theulaz, I. (1998). Presence of multiple “Helicobacter heilmannii” strains in an individual suffering from ulcers and in his two cats. J Clin Microbiol, 36, 1366 –1370. Eaton, K. A., Dewhirst, F. E., Paster, B. J., Tzellas, N., Coleman, B. E., Paola, J., & Sherding, R. (1996). Prevalence and varieties of Helicobacter species in dogs from random sources and pet dogs: animal and public health implications. J Clin Microbiol, 34, 3165–3170. Hilzenrat, N., Lamoureux, E., Weintrub, I., Alpert, E., Lichter, M., & Alpert, L. (1995). Helicobacter heilmannii-like spiral bacteria in gastric mucosal biopsies. Prevalence and clinical significance. Arch Pathol Lab Med, 119, 1149 –1153. Holck, S., Ingeholm, P., Blom, J., Norgaard, A., Elsborg, L., Adamsen, S., & Andersen, L. P. (1997). The histopathology of human gastric mucosa inhabited by Helicobacter heilmannii-like (Gastrospirillum hominis) organisms, including the first culturable case. APMIS, 105, 746 –756. Ierardi, E., Monno, R. A., Gentile, A., Francavilla, R., Burattini, O., Marangi, S., Pollice, L., & Francavilla, A. (2001). Helicobacter hei-
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7 lmannii gastritis: a histological and immunohistochemical trait. J Clin Pathol, 54, 774 –777. Jalava, K., Kaartinen, M., Utriainen, M., Happonen, I., & Hanninen, M. L. (1997). Helicobacter salomonis sp. nov., a canine gastric Helicobacter sp. related to Helicobacter felis and Helicobacter bizzozeronii. Int J Syst Bacteriol, 47, 975–982. Jhala, D., Jhala, N., Lechago, J., & Haber, M. (1999). Helicobacter heilmannii gastritis: association with acid peptic diseases and comparison with Helicobacter pylori gastritis. Mod Pathol, 12, 534 –538. Marais, A., Monteiro, L., Occhialini, A., Pina, M., Lamouliatte, H., & Megraud, F. (1999). Direct detection of Helicobacter pylori resistance to macrolides by a polymerase chain reaction/DNA enzyme immunoassay in gastric biopsy specimens. Gut, 44, 463– 467. McNulty, C. A., Dent, J. C., Curry, A., Uff, J. S., Ford, G. A., Gear, M. W., & Wilkinson, S. P. (1989). New spiral bacterium in gastric mucosa. J Clin Pathol, 42, 585–591. Megraud, F. (1995). Diagnosis of Helicobacter pylori. Baillieres Clin Gastroenterol, 9, 507–518. Mention, K., Michaud, L., Guimber, D., Martin, D. L., Vincent, P., Turck, D., & Gottrand, F. (1999). Characteristics and prevalence of Helicobacter heilmannii infection in children undergoing upper gastrointestinal endoscopy. J Pediatr Gastroenterol Nutr, 29, 533–539. Morgner, A., Lehn, N., Andersen, L. P., Thiede, C., Bennedsen, M., Trebesius, K., Neubauer, B., Neubauer, A., Stolte, M., & Bayerdorffer, E. (2000). Helicobacter heilmannii-associated primary gastric lowgrade MALT lymphoma: complete remission after curing the infection. Gastroenterology, 118, 821– 828. Neiger, R., Dieterich, C., Burnens, A., Waldvogel, A., Corthesy-Theulaz, I., Halter, F., Lauterburg, B., & Schmassmann, A. (1998). Detection
7
and prevalence of Helicobacter infection in pet cats. J Clin Microbiol, 36, 634 – 637. Norris, C. R., Marks, S. L., Eaton, K. A., Torabian, S. Z., Munn, R. J., & Solnick, J. V. (1999). Healthy cats are commonly colonized with “Helicobacter heilmannii” that is associated with minimal gastritis. J Clin Microbiol, 37, 189 –194. O’Rourke, J. L., Grehan, M., & Lee, A. (2001). Non-pylori Helicobacter species in humans. Gut, 49, 601– 606. Pena, J. A., McNeil, K., Fox, J. G., & Versalovic, J. (2002). Molecular evidence of Helicobacter cinaedi organisms in human gastric biopsy specimens. J Clin Microbiol, 40, 1511–1513. Peters, T. M., Owen, R. J., Teare, E. L., & Goodbourn, C. (1997). Detection of Helicobacter pylori by PCR on culture- and histology-negative gastric biopsies from seropositive patients. PHLS Microbiol Digest, 14, 227–229. Solnick, J. V., O’Rourke, J., Lee, A., Paster, B. J., Dewhirst, F. E., & Tompkins, L. S. (1993). An uncultured gastric spiral organism is a newly identified Helicobacter in humans. J Infect Dis, 168, 379 –385. Stolte, M., Wellens, E., Bethke, B., Ritter, M., & Eidt, H. (1994). Helicobacter heilmannii (formerly Gastrospirillum hominis) gastritis: an infection transmitted by animals. Scand J Gastroenterol, 29, 1061– 1064. Svec, A., Kordas, P., Pavlis, Z., & Novotny, J. (2000). High prevalence of Helicobacter heilmannii-associated gastritis in a small, predominantly rural area: further evidence in support of a zoonosis? Scand J Gastroenterol, 35, 925–928. Trebesius, K., Adler, K., Vieth, M., Stolte, M., & Haas, R. (2001). Specific detection and prevalence of Helicobacter heilmannii-like organisms in the human gastric mucosa by fluorescent in situ hybridization and partial 16S ribosomal DNA sequencing. J Clin Microbiol, 39, 1510 –1516.
www.elsevier.com/locate/diagmicrobio
Bacteriology
Development and application of a novel screening PCR assay for direct detection of ‘Helicobacter heilmannii’-like organisms in human gastric biopsies in Southeast England Stephanie A. Chisholm*, Robert J. Owen Laboratory of Enteric Pathogens, Central Public Health Laboratory, London NW9 5HT United Kingdom Received 6 August 2002; accepted 13 November 2002
Abstract A novel PCR assay (HHLO-16) to screen for presence of ‘Helicobacter heilmannii’-like organisms (HHLO) direct from gastric biopsies is described. As ‘H. heilmannii’ is generally uncultivable, diagnosis of infection is reliant on histology; thus prevalence may be underestimated. Analysis of an HHLO histology-positive human gastric biopsy and 15 gastric biopsies from domestic cats demonstrated that the HHLO-16 assay was more sensitive than an alternative available species-specific PCR assay. Further testing of 131 gastric biopsies from dyspeptic patients demonstrated an HHLO prevalence rate of 2.3% in Southeast England. Subsequent combination of the HHLO-16 assay with a H. pylori-specific PCR assay in a multiplex format (HpHh assay), and repeat analysis of the 131 biopsies showed the HpHh assay was as sensitive as each individual test. This novel PCR assay provides simple concomitant testing of dyspeptic patients for both HHLOs and H. pylori, thereby rapidly identifying individuals requiring eradication therapy. © 2003 Elsevier Inc. All rights reserved. Keywords: ‘Helicobacter heilmannii-like organisms’; Helicobacter pylori; Multiplex PCR assay; Biopsy; Detection
1. Introduction Infections with ‘Helicobacter heilmannii’, previously named ‘Gastrospirillum hominis’ in man (McNulty et al., 1989), are found in a wide range of other animal hosts that include cats, dogs and pigs with reported prevalence rates of 78-100% (De Groote et al., 1999; Eaton et al., 1996; Neiger et al., 1998). Although the nomenclature implies a defined species, 16S rDNA phylogenetic analyses demonstrate that ‘H. heilmannii’ comprises two related but distinct types (O’Rourke et al., 2001; Solnick et al., 1993). Type 1 is closely related to ‘Candidatus H. suis’, found in pigs (De Groote et al., 1999), whereas Type 2 is more closely related to the H. bizzozeronii/H. felis/H. salomonis subgroup that most commonly infect cats and dogs (Jalava et al., 1997; Solnick et al., 1993). These morphologically complex, often non-culturable bacteria, from various human and animal hosts represent part of an ill-defined taxonomic group of
* Corresponding author. Tel.: ⫹44-20-8200-4400; fax: ⫹44-20-89059929. E-mail address: [email protected] (S.A. Chisholm). 0732-8893/03/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0732-8893(02)00547-3
closely related forms which, for convenience, have been referred to as ‘Helicobacter heilmannii’-like organisms (HHLO). Human HHLO infections possibly occur via a zoonotic route, and may lead to the development of peptic ulcer disease (Hilzenrat et al., 1995; Jhala et al., 1999) or low-grade mucosal associated lymphoid tissue (MALT) lymphoma (Morgner et al., 2000). While infections with H. pylori, the well established human gastric pathogen, are world-wide and prevalence rates exceed 40% in many countries, the reported incidence of ‘H. heilmannii’ infection based on histology is low (0.12%) in dyspeptic adult (Ierardi et al., 2001; Jhala et al., 1999; Svec et al., 2000) and pediatric patients (Mention et al., 1999; Svec et al., 2000). As HHLOs from human tissue remain uncultivable with rare exceptions (Andersen et al., 1996), diagnosis is primarily reliant on histology, although touch cytology is also described (Debongnie et al., 1994). Recently, another microscopy-based system, fluorescent in situ hybridization (FISH), was described that detected all known types of ‘H. heilmannii’ and several new forms direct from paraffin embedded gastric biopsies (Trebesius et al., 2001). PCR-based studies have identified HHLOs in both human (Morgner et al., 2000) and animal (Norris et al.,
2
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
1999) gastric biopsies by amplification of Helicobacter genus-specific fragments of 16S rDNA followed by sequencing. One PCR detection assay, based on the ureB gene was described for detection of ‘H. heilmannii’ in domestic cat biopsies (Neiger et al., 1998), but specific PCR has only been applied for detection of human infection in a single patient (Dieterich et al., 1998). A number of histologically based studies have provided valuable information on the clinical features of HHLO infections (Holck et al., 1997; Jhala et al., 1999; Mention et al., 1999; Stolte et al., 1994) and have suggested that when compared with H. pylori, HHLO distribution tends to be focalised in the antrum (Jhala et al., 1999) as well as being patchy and sparse (Holck et al., 1997). Microscopy-based methods of detection can be relatively insensitive when bacterial numbers are low (Holck et al., 1997; Megraud 1995). In the original case report of this infection, it was stated that slides had to be examined for at least five minutes, to minimize possible false negative results (McNulty et al., 1989). It is thus possible that the incidence of human HHLO infection may be under-reported. The aim of this study was to develop a novel PCR-based assay to screen for HHLOs direct from human gastric biopsies to facilitate enhanced surveillance in England. The PCR test was then combined with a H. pylori-specific PCR assay in a multiplex format for rapid detection of both HHLO and H. pylori infections of the human gastric mucosa.
2. Materials and methods 2.1. Bacterial isolates and culture conditions Twenty-two strains of Helicobacter representing 12 species, mostly obtained as lyophilized cultures from the National Collection of Type Cultures (NCTC, London) were included in this study. These were H. pylori (NCTC 11637T and NCTC 11638, strains 26695, J99, and H566), H. fennelliae (NCTC 11612T and NCTC 11613), H. cinaedi (NCTC 11611 and NCTC 11614), H. pullorum (NCTC 12825 and NCTC 12826), H. canis (NCTC 12739T and NCTC 12745), H. acinonychis (NCTC 12688 and NCTC 12689), H. felis (NCTC 12436T), H. hepaticus (NCTC 12886T), H. nemestrinae ⫽ H. pylori (NCTC 12491T), H. muridarum (NCTC 12714T), H. pametensis (NCTC 12888 and NCTC 12889) and H. mustellae (NCTC 12032); T indicates the type strain of the species. Isolates were cultured on Columbia agar base (Oxoid) containing 10% (v/v) defibrinated horse blood under microaerobic conditions (4% O2, 5% CO2, 5% H2, 86% N2) in a variable atmospheric incubator (VAIN) (Don Whitley Scientific Ltd, Shipley, UK) at 37°C for 72h.
2.2. Clinical specimens Gastric biopsies were from 208 patients attending open access endoscopy clinics for routine investigation of dyspeptic symptoms. Human antral gastric biopsies included in this study comprised: 131 biopsies collected prospectively (Broomfield Hospital, Chelmsford and Kings College Hospital, London) for which H. pylori status had been previously determined by culture and by histologic examination of paired biopsies, this group included one biopsy for which there was histologic evidence of an HHLO infection (A. Price, S. M. Dobbs and R. J. Dobbs, personal communication); a stored (⫺20°C) collection of 39 biopsies known to be H. pylori-positive by culture and histology (Peters et al., 1997); and 38 biopsies, confirmed as H. pylori-positive at the time of endoscopy by the rapid urease (CLO) test (North Middlesex University Hospital, London). Biopsies in the latter group were retained in batches at the hospital and periodically forwarded to the laboratory at ambient temperature in CLO test strips for culture and molecular testing. Biopsies from Chelmsford PHL, were frozen in DENTS transport medium (⫺20°C) after primary culture, and subsequently transported frozen to the laboratory for molecular analyses. Gastric biopsies from 15 domestic cats under investigation for various symptoms at a veterinary clinic in north London were collected (unfrozen), cultured for HHLOs and other Helicobacters, examined microscopically by Gram stain and then stored (⫺20°C) until required for further analysis. 2.3. DNA extraction Genomic DNA was extracted as described previously from all gastric biopsies and from bacterial cultures (Marais et al., 1999; Chisholm et al., 2001). All DNA extracts were stored at ⫺20°C until required. 2.4. PCR-based detection of HHLOs A novel PCR assay (HHLO-16) targeting 16S rDNA to detect HHLOs was designed based on in silico alignments in GeneBase version 1.0 (Applied Maths, Kortjivik, Belgium) of 51 sequences of 16S rDNA genes from 19 species of Helicobacter held in GenBank that included: ‘H. heilmannii’ (accession numbers AF25625, AF058777, AF058770, AF058771, Y18028, AF058773, AF058772, AF058774, AF058775, AF058776, L10079, L10080), H. felis (M57398, M37643, U51872, U51871, U51870, AF103882, AF103881, AF 103880, AF103879), H. bizzozeronii (Y09404, AF103883, AF302107), H. salomonis (U89351) and Candidatus H. suis (AF127028). A 112 bp fragment of the 16S rRNA gene was then amplified from 100 ng culture-extracted DNA or from 5 L biopsy DNA extract in a reaction mix (50 L final volume) containing 200 M (each) deoxynucleoside trisphosphates
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
3
Fig. 1. PCR products generated by ‘H. heilmannii’ specific (*H) and by H. felis specific (†F) ureB assays (Neiger et al., 1996) and by assay HpHh (‡M) amplifying 16S rDNA and vacA fragments of HHLOs and H. pylori respectively. Human biopsy: negative for ‘H. heilmannii’ (ureB) (lane 1), and H. felis (ureB) (lane 3) but positive by histology and by HpHh assay (lane 5). Feline biopsy positive by both ureB assays (lanes 2 & 4) and by HpHh assay (lane 6); Lanes 7 & 8, HHLO positive human biopsies; Lanes 9 & 10, H. pylori positive human biopsies; Lanes 11 & 12, HHLO positive biopsies spiked with H. pylori DNA; Lane 13, 123-bp size marker.
(Invitrogen, UK), 0.4 M (each) primer HeilF (5⬘ AAG TCG AAC GAT GAA GCC TA 3⬘) and HeilR (5⬘ ATT TGG TAT TAA TCA CCA TTT C 3)’ (MWG Biotech Ltd, Germany), and 1 U of Taq polymerase (Invitrogen) in 20 mM Tris-HCl, pH 8.4, 1.5 mM MgCl2, 50 mM KCl, 0.2% (v/v) glycerol. Target DNA was denatured at 95°C for 5 min then amplified by 35 cycles of 1 min each at a denaturation temperature of 95°C, an annealing temperature of 53°C and an elongation temperature of 72°C, extension was completed by a 5-min incubation at 72°C. PCR products were electrophoresed (100V/1hr) in a 1% w/v agarose gel (Ultra Pure, Invitrogen) in TBE buffer (90 mM Tris-HCl, 90 mM boric acid, 2.0 mM EDTA), and ethidium bromide stained (0.5 g/ml). In addition, two separate species-specific PCR assays amplifying the urease B subunit (ureB) gene were applied to gastric biopsies to detect ‘H. heilmannii’ and H. felis following the protocols described previously (Neiger et al., 1998). 2.5. Multiplex PCR assay (HpHh) An H. pylori-specific PCR assay targeting the vacuolating cytotoxin (vacA) gene (Chisholm et al., 2001) was combined with the new HHLO-16 assay to form the multiplex assay designated HpHh. Either vacA or 16S rDNA were amplified in a 50 L reaction containing 5 L extracted DNA, PCR buffer, 1.5 mM MgCl2, 200 M (each) dNTP, 0.6 M (each) primer HeilF and HeilR, 0.28 M (each) primer VAC3624F (5⬘-GAG CGA GCT ATG GTT ATG AC-3⬘) and VAC4041R (5⬘-CAT TCC TAA ATT
GGA AGC GAA-3⬘) (MWG Biotech Ltd) and 1 U Taq polymerase (Invitrogen). Following denaturation at 95°C (5 min), specific targets were amplified in the reaction mix through 35 cycles as follows: 94°C for 30 sec, 54°C for 30 sec and 72°C for 45 sec each; extension was completed at 72°C for 5 min. Aliquots of each PCR product were electrophoresed in a 2% w/v agarose gel in TBE buffer and stained in ethidium bromide.
3. Results 3.1. Assessment of sensitivity and specificity of the HHLO16 assay Analysis of genomic DNA from 12 species of Helicobacter demonstrated that the HHLO-16 assay did not amplify the 112-bp 16S rDNA fragment from H. pylori or any species of Helicobacter other than H. felis. The latter result was consistent with the findings of a GenBank BLAST search that indicates the HHLO-16 assay would theoretically detect members of the H. felis, H. bizzozeronii, H. salomonis subgroup (⫽‘H. heilmannii’ type 2). However, DNA from the latter two species was not available for testing. Application of the HHLO-16 assay to a human gastric biopsy for which there was histologic evidence of an HHLO infection generated the expected 112-bp product (Fig. 1). Sensitivity of the HHLO-16 assay was assessed by application to gastric biopsies from domestic cats, where the
4
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
Table 1 Comparison of performances of the novel HHLO-16 assay, and two ‘H. heilmannii’* or H. felis* specific ureB PCR assays when applied to human and to feline gastric biopsies Gastric biopsy
PCR Assay
(n)
HHLO-16 ⫹ (†)
Human (113) Feline (15)
3 (1) 13 (2)
⫺ 110 2
formation, such as contact with animals, was available for these patients. 3.3. Evaluation of HpHh assay for detection of HHLOs and H. pylori
‘H. heilmannii’*
H. felis*
⫹(†)
⫹(†)
⫺
‡
na 14
⫺
1 (0) 13 (2)
112 2
na 1 (1)
The multiplex HpHh assay was evaluated initially on DNA from nine gastric biopsies. The H. pylori-specific vacA PCR product was generated from DNA of the five H. pylori positive biopsies tested, while specific fragments of 16S rDNA were amplified in all four DNA extracts from HHLO-positive biopsies of both human (n ⫽ 2) and feline (n ⫽ 2) origin. In addition, analysis of DNA from the HHLO-positive biopsy spiked with 100 ng H. pylori DNA showed that each target was successfully co-amplified (Figure 1). Application of the two PCR assays independently demonstrated that for biopsies previously tested by culture and histology, 59/170 were H. pylori and 3/170 were HHLO PCR-positive, respectively, while vacA and 16S rDNA was amplified in 27/38 and 0/38 CLO test positive biopsies, respectively (Table 2). Subsequent repeat testing of human biopsies demonstrated that both assays in a multiplex format were as sensitive as the individual PCR assays (Table 2). However the HpHh test gave one false-negative as 12/15 feline gastric biopsies were HHLO-positive compared with 13/15 positive by single testing with the HHLO-16 assay (Table 2).
* Published assays targeting the urease B subunit (ureB) (Neiger et al., 1998) † Number confirmed microscopically (histology or Gram stain) ‡ H. felis assay not applied to human gastric biopsies
incidence of infection with HHLOs was expected to be high according to reported prevalence rates (Norris et al., 1999). The HHLO-16 assay generated a specific product in 13/15 feline biopsies, two of which were confirmed as containing spiral organisms by Gram stain (Table 1). Further analyses of the cat biopsies with the two ureB PCR assays of Neiger et al. (1998) demonstrated that all 13 of these were positive for ‘H. heilmannii’ (⫽type 2) while only one biopsy was also positive for H. felis (Table 1). Spiral organisms were observed in that particular biopsy by Gram staining and microscopy. 3.2. Determination of prevalence of infection with HHLOs in man
3.4. Investigation of the frequency of H. pylori/HHLO co-infections
Application of the HHLO-16 assay to DNA extracted from biopsies, collected prospectively from dyspeptic patients in mid-Essex (England) over a 3-month period in 1999, generated specific fragments of 16S rDNA from 3/131 biopsy extracts giving a prevalence rate of 2.3%. There did not appear to be any familial or geographical link between the three HHLO positive patients. No further in-
The HpHh multiplex PCR assay was applied to 101 human gastric biopsies that were positive for H. pylori, either by culture and/or histology (n ⫽ 63) or by CLO test (n ⫽ 38). The results of the assay showed vacA and 16S rDNA fragments in 86/101 and 0/101 samples, respectively, indicating no co-infections.
Table 2 Comparison of sensitivities of the vacA and HHLO-16 assays when applied individually and combined in a multiplex format (HpHh) PCR assays H. pylori (vacA) Gastric Biopsies
HHLO-16 (16S rRNA) % Sensitivity†
% Sensitivity
Origin (n)
Status*
⫹
⫺
Uniplex
Multiplex
⫹
Cat (15) Human (170) Human (38)
C/M ⫹ C/M ⫺ C/H ⫹ C/H ⫺ U⫹
0 0 59
0 15 4
na
na
94.0
94.0
27
11
77.5
77.5
0 13 0 3 0
12‡
⫺
Uniplex
Multiplex‡
0 2 3‡ 63 104 38
100.0
92.8
100.0
100.0
Na
na
* H. pylori status determined by culture (C), Gram stain microscopy (M), histology (H) or rapid urease (CLO) test (U). Sensitivity based on the observation that no PCR negative biopsies were histology positive. ‡ Results generated for the HpHh multiplex assay. †
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
4. Discussion A novel PCR based assay (HHLO-16) was developed to amplify fragments of the HHLO genomes directly from gastric biopsies, and its performance was compared with that of an alternative PCR assay targeting the ureB gene (Neiger et al., 1998). We aimed to develop an assay that could detect all known HHLO infections so the primers were designed by multiple alignment of 11 16S rDNA sequences of ‘H. heilmannii’ types 1 and 2, as well as Candidatus H. suis, H. bizzozeronii, H. felis and H. salomonis sequences currently held in GenBank. Application of the HHLO-16 assay to 12 different species of Helicobacter confirmed that only members of the HHLO group would be amplified. Other assays have been described for use on antral biopsies. For example, a ‘Candidatus H. suis’-specific PCR assay has been described that amplified product from porcine gastric biopsies (De Groote et al., 2000), while another assay for the H. bizzozeronii, H. felis, H. salomonis group was reported recently that successfully amplified specific product from canine gastric biopsies (De Groote et al., 2001). Our BLAST analysis of each of the primer pairs demonstrated that the former assay would amplify only ‘H. heilmannii’ type 1 sequences held in GenBank while the latter assay would amplify type 2 only. In contrast the HHLO-16 assay will amplify both types 1 and 2. As each type has been reported in the human stomach (Trebesius et al., 2001), our assay is more suitable for application to human gastric biopsies as it provides a rapid method of screening for all HHLOs in a single reaction. Comparison of the HHLO-16 assay with the ureB assay by Neiger et al. (1998) demonstrated that both assays generated product in 13/15 feline biopsies, but while three human biopsies were HHLO-16 PCR positive, only one human biopsy was positive by the ureB assay, and the histologically confirmed positive biopsy was PCR negative. This may indicate sequence variation in ureB between human and feline strains and suggests that the HHLO-16 assay is more sensitive for detection of HHLO infection in human samples. Moreover, our GenBank BLAST search of the ureB primer sequences suggest that only 4/16 strains held in GenBank would be amplified by this assay. It seems likely that this ureB assay is ‘H. heilmannii ’ type 2-specific and thus lacks the broad specificity needed for initial testing of human infections. The HHLO-16 assay gave sharper single bands than the ureB assay that often generated non-specific bands. Prospective application of HHLO-16 to 131 gastric biopsies demonstrated a HHLO prevalence rate of 2.3% in the series of patients in Southeast England. The incidence was higher than reported previously in other European countries (Ierardi et al., 2001; Mention et al., 1999) and the USA (Jhala et al., 1999; Mention et al., 1999). This observation may indicate either a higher geographical prevalence, or alternatively histology-based studies may have underestimated the true rate of infection in the particular dyspeptic
5
population examined. If the latter were true, then some patients, who would benefit from specific eradication therapy, are not being identified-thus a PCR-based approach for detection may be a more sensitive alternative to histology. Our results show that the incidence of HHLO infection is significantly lower than that of H. pylori in the study population (2.3% vs 16.8%) so it would not be cost effective to perform routine PCR tests for just HHLOs. The HHLO-16 assay was therefore combined in a multiplex format with another modified assay that amplified a fragment of the H. pylori-specific vacA gene (Chisholm et al., 2001). Evaluation of the HpHh assay on human biopsies of known H. pylori and HHLO status demonstrated that the sensitivity and specificity of the HpHh assay was comparable to those of the individual tests. In addition, spiking experiments demonstrated that both targets could be amplified simultaneously; thus this system could potentially detect co-infections with both HHLOs and H. pylori. Analysis of the feline biopsies demonstrated that sensitivity of the HHLO-16 assay was only slightly reduced, with one false negative result, in the multiplex format. Comparison of sensitivity of detection of serially diluted H. felis DNA (ranging from 100 ng –1 fg) for both assay formats demonstrated that both uniplex and multiplex assays detected as little as 10pg DNA, but the band intensity at the highest dilution was lower for the product generated in the multiplex format (data not shown). Thus the multiplex assay is only slightly less sensitive than the uniplex format. PCR reagent concentrations and cycling conditions were re-optimized to facilitate co-amplification of both targets, reduced efficiency of the HHLO-16 assay may have resulted for that sample from sequence variation at the primer binding sites. However, we suggest the benefit of combining the two assays in a multiplex format to provide a more diagnostically useful test outweighs the slight reduction in sensitivity. Although human H. pylori and HHLO co-infection has been identified by histology (Ierardi et al., 2001), neither the incidence nor the possible significance of this in terms of disease development has been determined. In the present study we found no evidence of co-infection. One hundred and one confirmed H. pylori-positive biopsies were tested by assay HpHh, but generation of specific HHLOs was not observed in any of the biopsies that were PCR positive for H. pylori. Given the observed low incidence of HHLO infection in the population investigated, it could be concluded that the incidence of co-infections with H. pylori is comparatively rare in Southeast England. A set of human biopsies included in this study had been identified as H. pylori-positive by rapid urease testing (CLO test). However, it should be noted that urease activity of HHLOs can also cause a CLO test color change (McNulty et al., 1989), and consequently CLO test positive results may be misinterpreted. These CLO-test biopsies were analyzed by the HpHh assay not only to investigate prevalence of co-infections but also to establish if any HHLO infections had been misdiagnosed as H. pylori positive. The sensitivity
6
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7
of detection for the vacA assay was lower than had been found for the culture and/or histology-positive human biopsies tested. The possibility of false negative results due to PCR inhibitors in biopsy DNA extracts had been eliminated previously (data not shown) by an internal control PCR assay using primer pair H6A/H6B that amplifies the human mitochondrial cytochrome oxidase subunit 3 gene (Cadieux et al., 1993). As these biopsies had been transported at room temperature in the original CLO test from the primary diagnostic laboratory to our laboratory, DNA degradation may have occurred during transit-a phenomenon that we have observed previously when transporting using conventional media (Chisholm et al., 2001). While we did not identify any biopsies that were CLO test positive due to HHLO infection, the possibility of false negative results in 12 biopsies where specific target DNA may have degraded cannot be fully excluded. In addition, as densities of HHLO infections have been reported to be patchy and sparse relative to H. pylori (Holck et al., 1997; McNulty et al., 1989), it would be useful also to examine CLO tests that have generated weak or equivocal urease positive results for possible HHLO infection, however these were not available for the present investigation. A recently reported fluorescent in situ hybridization (FISH) system demonstrated that ‘H. heilmannii’ type 1 was the most prevalent in human infection, while type 2 and other novel types (3, 4 and 5) were less common, with mixed type infections also observed (Trebesius et al., 2001). Combination of our novel assay with the FISH technology would provide a comprehensive strategy for better characterization of HHLO infection in humans. The development of the HpHh assay provides a sensitive screening method to identify those biopsies positive for HHLOs that require additional, more complex testing by FISH to enable strain differentiation and characterization. A recent report that identified H. cinaedi in the gastric biopsies of two patients (Pena et al., 2002) has raised the possibility that a broader range of Helicobacter spp may be capable of colonising the human gastric mucosa than had been previously appreciated. As there is only one such report to date, further investigations aimed at detecting other Helicobacters including H. cinaedi will be necessary to establish the extent and significance of such infections at that site. In conclusion, the development and evaluation of a novel multiplex PCR assay for detection of HHLOs and H. pylori provides a means of direct detection of the principal bacterial pathogens causing gastric infection and disease in man. Routine use of this test would enable more accurate assessment of the prevalence of HHLOs in the dyspeptic population and would facilitate studies to improve understanding of the clinical significance and sources, such as domestic pets, of those infections in man, as well as enabling further investigation of HHLO/H. pylori co-infections. Future prospective analyses of gastric biopsies will now be possible to investigate larger study populations.
Acknowledgments We thank Dr E.L. Teare (Chelmsford PHL), Dr S. Saverymuttu (Broomfield Hospital, Chelmsford), Dr B. Patel (North Middlesex University Hospital, London) and Dr S.M. Dobbs and Dr R.J. Dobbs (King’s College Hospital, London) for kindly providing routine gastric biopsies. We are also grateful to Professor A. Price (Northwick Park Hospital, Harrow) for providing histologic details, and to Mrs M. Forster-Van Hijfte (The Village Veterinary Practice, Belsize Park, London) for providing gastric biopsies from domestic cats.
References Andersen, L.P., Norgaard, A., Holck, S., Blom, J., & Elsborg, L. (1996). Isolation of a “Helicobacter heilmannii”-like organism from the human stomach. Eur J Clin Microbiol Infect Dis, 15, 95–96. Cadieux, N., Lebel, P., & Brousseau, R. (1993). Use of a triplex polymerase chain reaction for the detection and differentiation of Mycoplasma pneumoniae and Mycoplasma genitalium in the presence of human DNA. J Gen Microbiol, 139, 2431–2437. Chisholm, S. A., Owen, R. J., Teare, E. L., & Saverymuttu, S. (2001). PCR-based diagnosis of Helicobacter pylori infection and real-time determination of clarithromycin resistance directly from human gastric biopsy samples. J Clin Microbiol, 39, 1217–1220. De Groote, D., Ducatelle, R., van Doorn, L. J., Tilmant, K., Verschuuren, A., & Haesebrouck, F. (2000). Detection of “Candidatus Helicobacter suis” in gastric samples of pigs by PCR: comparison with other invasive diagnostic techniques. J Clin Microbiol, 38, 1131–1135. De Groote, D., Haesebrouck, F., van Doorn, L. J., Vandamme, P., & Ducatelle, R. (2001). Evaluation of a group-specific 16S ribosomal DNA-based PCR for detection of Helicobacter bizzozeronii, Helicobacter felis, and Helicobacter salomonis in fresh and paraffin-embedded gastric biopsy specimens. J Clin Microbiol, 39, 1197–1199. De Groote, D., van Doorn, L. J., Ducatelle, R., Verschuuren, A., Haesebrouck, F., Quint, W. G., Jalava, K., & Vandamme, P. (1999). ‘Candidatus Helicobacter suis’, a gastric helicobacter from pigs, and its phylogenetic relatedness to other gastrospirilla. Int J Syst Bacteriol, 49, 1769 –1777. Debongnie, J. C., Mairesse, J., Donnay, M., & Dekoninck, X. (1994). Touch cytology. A quick, simple, sensitive screening test in the diagnosis of infections of the gastrointestinal mucosa. Arch Pathol Lab Med, 118, 1115–1118. Dieterich, C., Wiesel, P., Neiger, R., Blum, A., & Corthesy-Theulaz, I. (1998). Presence of multiple “Helicobacter heilmannii” strains in an individual suffering from ulcers and in his two cats. J Clin Microbiol, 36, 1366 –1370. Eaton, K. A., Dewhirst, F. E., Paster, B. J., Tzellas, N., Coleman, B. E., Paola, J., & Sherding, R. (1996). Prevalence and varieties of Helicobacter species in dogs from random sources and pet dogs: animal and public health implications. J Clin Microbiol, 34, 3165–3170. Hilzenrat, N., Lamoureux, E., Weintrub, I., Alpert, E., Lichter, M., & Alpert, L. (1995). Helicobacter heilmannii-like spiral bacteria in gastric mucosal biopsies. Prevalence and clinical significance. Arch Pathol Lab Med, 119, 1149 –1153. Holck, S., Ingeholm, P., Blom, J., Norgaard, A., Elsborg, L., Adamsen, S., & Andersen, L. P. (1997). The histopathology of human gastric mucosa inhabited by Helicobacter heilmannii-like (Gastrospirillum hominis) organisms, including the first culturable case. APMIS, 105, 746 –756. Ierardi, E., Monno, R. A., Gentile, A., Francavilla, R., Burattini, O., Marangi, S., Pollice, L., & Francavilla, A. (2001). Helicobacter hei-
Chisholm and Owen / Diagnostic Microbiology and Infectious Disease 46 (2003) 1–7 lmannii gastritis: a histological and immunohistochemical trait. J Clin Pathol, 54, 774 –777. Jalava, K., Kaartinen, M., Utriainen, M., Happonen, I., & Hanninen, M. L. (1997). Helicobacter salomonis sp. nov., a canine gastric Helicobacter sp. related to Helicobacter felis and Helicobacter bizzozeronii. Int J Syst Bacteriol, 47, 975–982. Jhala, D., Jhala, N., Lechago, J., & Haber, M. (1999). Helicobacter heilmannii gastritis: association with acid peptic diseases and comparison with Helicobacter pylori gastritis. Mod Pathol, 12, 534 –538. Marais, A., Monteiro, L., Occhialini, A., Pina, M., Lamouliatte, H., & Megraud, F. (1999). Direct detection of Helicobacter pylori resistance to macrolides by a polymerase chain reaction/DNA enzyme immunoassay in gastric biopsy specimens. Gut, 44, 463– 467. McNulty, C. A., Dent, J. C., Curry, A., Uff, J. S., Ford, G. A., Gear, M. W., & Wilkinson, S. P. (1989). New spiral bacterium in gastric mucosa. J Clin Pathol, 42, 585–591. Megraud, F. (1995). Diagnosis of Helicobacter pylori. Baillieres Clin Gastroenterol, 9, 507–518. Mention, K., Michaud, L., Guimber, D., Martin, D. L., Vincent, P., Turck, D., & Gottrand, F. (1999). Characteristics and prevalence of Helicobacter heilmannii infection in children undergoing upper gastrointestinal endoscopy. J Pediatr Gastroenterol Nutr, 29, 533–539. Morgner, A., Lehn, N., Andersen, L. P., Thiede, C., Bennedsen, M., Trebesius, K., Neubauer, B., Neubauer, A., Stolte, M., & Bayerdorffer, E. (2000). Helicobacter heilmannii-associated primary gastric lowgrade MALT lymphoma: complete remission after curing the infection. Gastroenterology, 118, 821– 828. Neiger, R., Dieterich, C., Burnens, A., Waldvogel, A., Corthesy-Theulaz, I., Halter, F., Lauterburg, B., & Schmassmann, A. (1998). Detection
7
and prevalence of Helicobacter infection in pet cats. J Clin Microbiol, 36, 634 – 637. Norris, C. R., Marks, S. L., Eaton, K. A., Torabian, S. Z., Munn, R. J., & Solnick, J. V. (1999). Healthy cats are commonly colonized with “Helicobacter heilmannii” that is associated with minimal gastritis. J Clin Microbiol, 37, 189 –194. O’Rourke, J. L., Grehan, M., & Lee, A. (2001). Non-pylori Helicobacter species in humans. Gut, 49, 601– 606. Pena, J. A., McNeil, K., Fox, J. G., & Versalovic, J. (2002). Molecular evidence of Helicobacter cinaedi organisms in human gastric biopsy specimens. J Clin Microbiol, 40, 1511–1513. Peters, T. M., Owen, R. J., Teare, E. L., & Goodbourn, C. (1997). Detection of Helicobacter pylori by PCR on culture- and histology-negative gastric biopsies from seropositive patients. PHLS Microbiol Digest, 14, 227–229. Solnick, J. V., O’Rourke, J., Lee, A., Paster, B. J., Dewhirst, F. E., & Tompkins, L. S. (1993). An uncultured gastric spiral organism is a newly identified Helicobacter in humans. J Infect Dis, 168, 379 –385. Stolte, M., Wellens, E., Bethke, B., Ritter, M., & Eidt, H. (1994). Helicobacter heilmannii (formerly Gastrospirillum hominis) gastritis: an infection transmitted by animals. Scand J Gastroenterol, 29, 1061– 1064. Svec, A., Kordas, P., Pavlis, Z., & Novotny, J. (2000). High prevalence of Helicobacter heilmannii-associated gastritis in a small, predominantly rural area: further evidence in support of a zoonosis? Scand J Gastroenterol, 35, 925–928. Trebesius, K., Adler, K., Vieth, M., Stolte, M., & Haas, R. (2001). Specific detection and prevalence of Helicobacter heilmannii-like organisms in the human gastric mucosa by fluorescent in situ hybridization and partial 16S ribosomal DNA sequencing. J Clin Microbiol, 39, 1510 –1516.