Liquid culture medium for the rapid cultivation of Helicobacter pylori from biopsy specimens
- First Online:
- Cite this article as:
- Sainsus, N., Cattori, V., Lepadatu, C. et al. Eur J Clin Microbiol Infect Dis (2008) 27: 1209. doi:10.1007/s10096-008-0567-6
- 424 Views
The goal of this study was to develop a liquid culture medium for the rapid isolation, cultivation, identification and subsequent antibiotics susceptibility testing of Helicobacter pylori directly from biopsy specimens. Five liquid media were tested: Ham’s F-12, Brucella broth, tryptic soybroth, brain heart infusion broth and Mueller-Hinton broth. After optimisation of the medium, it was applied in order to investigate biopsy samples from 150 patients with gastro-duodenal disorders and compared with traditional culture methods, microscopy and an H. pylori-specific TaqMan real-time polymerase chain reaction (PCR). The most reliable and rapid growth of H. pylori, even at a small inoculum size, was obtained in Ham’s F-12 medium with 5% horse serum. The developed system allowed the primary isolation of H. pylori in clinical samples and provided 87% sensitivity and 100% specificity.
Helicobacter pylori is a causative agent of gastro-duodenal ulcer disease and an established carcinogen for gastric adenocarcinoma [1–3]. Poor sanitation, polluted groundwater and crowding are considered to be risk factors for H. pylori transmission [4–6]. The diagnosis and treatment of H. pylori infections are an important direction of public health related to large expenses . Traditional methods of H. pylori cultivation are cumbersome, time-consuming and require a special atmosphere-generating system. In the clinical practice, the non-invasive urease breath test can be used as an indicator for a urease-positive H. pylori infection. In addition, sensitive and H. pylori-specific polymerase chain reaction (PCR) have been developed, but these methods are rather expensive and are offered only by specialised laboratories. The definitive diagnosis of an H. pylori infection is usually still based on endoscopy followed by bacterial culture and/or histological analysis of gastric biopsy sections.
At present, physicians often prescribe empirical eradication therapy for the treatment of patients with diagnosed H. pylori infection. However, the increasing antibiotics resistance of H. pylori, particularly to clarithromycin and metronidazole, compromises the eradication and may, thus, worsen the clinical outcome and delay the recovery of the patient. Primary and secondary resistance rates reported worldwide vary from 1 to 58% for clarithromycin and from 5 to 76% for metronidazole [8–10]. Together with the growing resistance of H. pylori, the lack of convenient and rapid antibiotics susceptibility tests result not only in inefficient therapy, but also inadequate antibiotics-resistance surveillance . These problems could be circumvented if a low-tech, low-cost culture method for use in the clinical practice could be developed that permits not only to cultivate and specifically detect H. pylori, but also promptly test the antibiotics resistance of the cultured agent. Thus, the goal of the present study was to design and realise a combined three-in-one H. pylori test with a liquid culture/indication medium (CIM) that allows for: (1) selective primary isolation of H. pylori, (2) an indicator test and (3) a tentative antibiotics susceptibility test for H. pylori.
While developing this assay, several critical issues were considered. H. pylori is rather sensitive to transportation [12, 13]. Usually, at least 1 to 2 h passes between endoscopy and inoculation of the sample in the laboratory. The stress related to prolonged transportation can negatively affect H. pylori growth on solid media . Therefore, placing the biopsy specimen immediately after collection into a liquid culture medium seems to be an optimal solution.
Various non-specific liquid culture media (Brucella broth, Mueller-Hinton broth, brain heart infusion broth) have been used historically to culture H. pylori, but none of them exhibited any particular advantage [15–17]. Previous studies on H. pylori growth in complex and chemically defined liquid media demonstrated that this species is not as fastidious as it was formerly considered to be [18–22]. The addition of blood to the medium or the maintenance of a strict microaerophilic atmosphere was found to be not absolutely essential [23–27]. As shown by other researchers, the defined medium of Ham’s F-12 nutrient mixture has a unique property to provide the stable growth of H. pylori, even without the addition of serum [28, 29]. Nowadays, liquid media are usually used to accumulate large quantities of H. pylori cells and to transport and preserve H. pylori cultures.
In this study, a selective liquid media was formulated for the primary isolation of H. pylori from gastric mucosa biopsy samples. The selection criteria for the desired medium were: simple and stable composition, transparency and maintaining of growth at small initial inoculum sizes.
The rapid urease reaction can be used as an indicator test for the presence of urease-positive H. pylori in the culture. However, the addition of urea to liquid culture media may negatively affect H. pylori viability [30, 31]. The rapid rise of the pH above 7.2 caused by the urease activity is responsible for the dramatic decrease of H. pylori survival . Therefore, in the present study, we aimed to use a delayed urease test after 48 h of incubation of the liquid medium.
To evaluate the developed H. pylori test, it was applied to investigate gastric mucosa biopsy specimens collected from human patients diagnosed with gastro-duodenal ulcer, gastritis, gastric cancer and/or dyspepsia. The test was compared to an H. pylori-specific TaqMan real-time PCR assay and the sensitivity, specificity, as well as positive and negative predictive values of the developed test were determined.
In order to produce a low-cost system, we tested the feasibility of creating the necessary atmosphere composition directly in the container with the culture medium. Different atmosphere compositions with varying CO2 and O2 content have been successfully used for the growth of H. pylori in liquid media [27, 33, 34]. CO2 concentrations of 5 to 10% and 5 to 7% oxygen were found to be optimal . In our experiments, the necessary atmosphere was obtained by saturation of the liquid medium in the container with CO2. To validate this method, we compared the growth of H. pylori in CO2-enriched medium and in a standard atmosphere-generating system.
It was also found that the addition of antibiotics to the proposed liquid medium allowed for rapid qualitative susceptibility testing of H. pylori isolated from the biopsy specimens.
Materials and methods
Bacterial strains and growth
H. pylori strain CIP 103995 (identical to ATCC 43504) from the Centre de Ressources Biologiques de l’Institut Pasteur, France, and minimally passaged fresh clinical isolates were used in this study. These strains were maintained as a frozen stock at −70°C in brain heart infusion broth (BHIB; Oxoid, Hampshire, UK) with 20% glycerol. Brucella broth (BB; Fluka, Buchs, Switzerland) supplemented with 10% horse serum (HS) was used as the transport medium. The identification of H. pylori was done on the basis of colonial morphology (1–2 mm, small and translucent colonies), microscopically (the presence of curved Gram-negative cells) and by positive urease reaction. Gram stain with carbolfuchsin was used for microscopic examination.
Bacteria were grown on solid growth medium using Columbia agar (Fluka, Buchs, Switzerland) with 10% HS (Sigma-Aldrich, Seelze, Germany) under microaerophilic conditions at 37°C. The medium was supplemented with 10 mg/l of vancomycin, 5 mg/l of cefsulodin (Sigma-Aldrich), 2,500 U/ml of polymyxin B, 5 mg/l of trimetoprim and 5 mg/l of amphotericin B (Fluka, Buchs, Switzerland). The combination of urea (600 mg/l), phenol red (100 mg/l) and hydrochloric acid (HCl) was incorporated into the medium to obtain a presumptive identification of H. pylori presence by colour shift induced by urease activity. The final pH of the medium at 25°C was 6.0 after the dropwise addition of hydrochloric acid 1 N HCl. Unless stated otherwise, the chemicals were from Sigma-Aldrich (Seelze, Germany).
Preparation of H. pylori inoculums and counting of colony-forming units
For the production of H. pylori inoculums, pure colonies were picked from solid growth medium and homogenised in sterile sodium phosphate buffer (4%) to 108 colony-forming units (CFU)/ml, as determined by the McFarland nephelometer. The homogenate was further diluted in the tested medium to the necessary inoculum size. Inoculum sizes were confirmed by culture in quadruplicate on Columbia agar prepared as described above. After 5 to 7 days of incubation, the numbers of CFU on each plate were counted visually and the CFU/ml of the inoculum was calculated. The theoretical lower limit of bacterial counting accuracy was 300 CFU/ml. Concentration–time curves were constructed by plotting the log10 CFU/ml values versus time.
Selection of liquid culture medium
The growth of H. pylori (reference strain and three clinical strains) was tested in five liquid culture media: Ham’s F-12, brucella broth (BB), Mueller-Hinton broth (MHB), tryptic soy broth (TSB) and brain heart infusion broth (BHIB), and at two different inoculum sizes. The five liquid media were supplemented with 5% HS. Selectivity of the media was provided by the addition of vancomycin 10 mg/l, cefsulodin 5 mg/l, polymyxin B 2,500 U/l, trimetoprim 5 mg/l and amphotericin B 7.5 mg/l. H. pylori suspensions (10 ml) were incubated in a 50-ml tissue culture flask in a microaerobic environment at 37°C without shaking. The microaerophilic atmosphere was generated using a CampyGen system in an AnaeroJar (Oxoid, Hampshire, UK). Cultures were observed for 96 h. Samples were collected at 24-h intervals. Bacterial growth was measured by determining the CFU/ml on solid medium as described above.
The liquid medium which best supported the growth of H. pylori at both inoculum sizes was chosen for all further experiments.
Optimisation of the liquid culture medium
To enhance the growth capacity of the chosen medium (Ham’s F-12), it was diluted with distilled water (5%) to slightly reduce the concentration of sodium chloride. For the same reason, the medium was supplemented with additional iron, as ferric chloride hexahydrate, which was used at a quantity of 13.5 mg/l. Polymyxin B was supplemented to decrease the rate of contaminations, although it can inhibit some H. pylori strains. To reliably prevent the fungal growth, amphotericin B was used in a concentration higher than usual. To obtain a more intense colour, 0.01 g/l of phenol red was added. Finally, 1 N HCl was added drop-wise to the media as the colour changed from red to yellow/orange (final pH 6.0 at 22°C).
Evaluation of optimised liquid medium (CIM) with or without horse serum
The H. pylori reference strain was grown at 103 CFU/ml in the CIM, either with or without 5% HS. To avoid the inhibition of H. pylori growth by antibiotics in the absence of serum, the serum-free medium did not contain antibiotics, except for amphotericin B. Bacterial growth was measured by determining the CFU/ml on solid medium as described above. Bacterial forms were identified by microscopy after 4 days of culture. This experiment was repeated three times.
Patients and sample collection
A total of 120 patients were sampled from January 2006 to April 2007 in the Clinical Republican Hospital, the major tertiary hospital in Moldova. Adult patients (age range 18 to 79 years) with a diagnosis of gastro-duodenal ulcer, gastritis, gastric cancer and dyspepsia without organic findings were examined clinically and endoscopically. Antisecretory agents and antibiotics were not administered at least 4 weeks prior to the first gastroscopy. Altogether, 360 gastric biopsies (three adjacent specimens per patient) were collected from the antrum mucosa. One specimen was examined using the developed test with the optimised liquid medium, the second specimen was used for microscopy and culture on solid medium, and the third biopsy specimen was left for PCR analysis. Patients with positive tests were prescribed a 7-day triple eradication therapy with clarithromycin 500 mg twice daily, metronidazole 500 mg twice daily and lanzoprazol 30 mg once daily. The second gastroscopy with biopsy was recommended one month after treatment.
Processing of specimens, testing of bacterial growth and delayed urease test
Biopsy specimens were homogenised and incubated in the 4-ml test tubes described above in 3 ml of CIM at 37°C for 48 h. The microaerophilic atmosphere was generated using a CampyGen system in an AnaeroJar (Oxoid, Hampshire, UK). The growth of H. pylori was determined visually and confirmed by microscopy and culture on solid medium as described above. In addition to bacterial growth, urease activity was tested after 48 h of incubation. To this end, 0.9 ml of culture was extracted with a sterile syringe and mixed with 0.1 ml of a urea (100 mg/ml) solution in sterile saline. A reaction was judged positive if the colour of the suspension changed from yellow to pink within 5 min. Tests without visual growth but with positive urease reaction underwent an additional 48 h of microaerophilic incubation. Only samples with both positive culture and urease activity were considered to be H. pylori-positive.
Real-time TaqMan PCR
DNA was extracted from frozen human gastric biopsy samples using the Qiagen QIAamp Mini Kit. For the digestion, the biopsies were transferred in 180 μl of ATL buffer with the aid of a pipette tip “sucking” the biopsy through the tip, and digested for 4 h at 56°C after adding 20 μl of proteinase K. Elution of the DNA was performed using 100 μl of elution buffer. To detect potential cross-contaminations, extraction was performed concurrently on samples containing ATL buffer only. Successful extraction was assessed by analysing a 5-μl sample with the ABI 18S RNA gene real-time PCR system (Applied Biosystems) using the GeneAmp Fast PCR Mastermix (Applied Biosystems) on an ABI7500 fast cycler following the manufacturer’s instructions. The detection of H. pylori genomic DNA was based on a published real-time system specific for the 16S RNA gene . The reporter dye was FAM and the quencher TAMRA. A standard curve for absolute quantification was produced as described , starting from DNA from one patient using following primers flanking the real-time PCR system: HPy16SflankF 5’-gtgtgggagaggtaggtgga-3’, HPy16SflankR 5’-tgcgttagctgcattactgg-3’. The sensitivity of the system was one copy (0.5 genomes) per reaction. The quantification of H. pylori was performed by a comparison with the produced standard curve on a Rotor-Gene 6000 sequence detection system (Corbett Research, Sydney, Australia) as follows: a 5-μl samples were amplified with 6.0 μl of GeneAmp Fast PCR Mastermix (Applied Biosystems), 500 nM each primer, 250 nM probe in a 12-μl total reaction volume. An initial denaturation of 20 s at 95°C was followed by 45 cycles of 95°C for 3 s and 60°C for 30 s.
Comparison of the developed test system with other H. pylori detection systems
The developed test was compared to the H. pylori-specific real-time PCR assay. In addition, all samples were evaluated using culture on solid medium and microscopy. The sensitivity, specificity, as well as positive and negative predictive values of the developed test was determined.
Microaerophilic atmosphere in test tubes
In these experiments, optimised liquid medium was used without the addition of HCl. Thus, the initial pH of the medium was 7.4. Four-millilitre vials with screw-caps and PTFE/silicone septums (Supelco, Bellefonte, PA, USA) were used as the test tubes. The vials contained 3 ml of liquid medium. CO2 (1.5 cm3≈3 mg of CO2) was added through the cap. The pH of the medium was measured using a pH meter (Piccolo Plus, Hanna Instruments). This experiment was repeated ten times. To determine the influence of the addition of the specimen during which the test tube has to be opened, the dynamics of the pH of the medium was measured over time. After the addition of CO2 to the test tube, the tube was opened and the pH assessed immediately (time point 0), after 1 min and then every 5 min up to 30 min. This experiment was repeated three times.
To test whether the CO2-enriched test tubes support the growth of H. pylori, the tubes were inoculated with 105 CFU/ml of the reference strain and six clinical strains. The vials were incubated flat to increase the area of the gas–liquid interface. Bacterial growth was measured by determining the CFU/ml on solid medium as described above.
For antibiotics screening tests, clarithromycin (2 mg/l) or metronidazole (8 mg/l) were added to 3 ml of CIM in large 4-ml test tubes. Inoculum (0.1 ml) was prepared from 18 clinical isolates of 2-day H. pylori cultures grown in the same medium and the final concentration in the test was 105 CFU/ml. The tests were incubated in microaerophilic conditions in an AnaeroJar for 48 h fitted with loose caps. Inocula (0.1 ml) were prepared from 18 clinical isolates and a reference strain (CIP 103995) that had been grown in CIM for 48 h; the inoculum size was 105 CFU/ml.
In parallel, tubes with identical inoculums but without clarithromycin or metronidazole were cultured as positive growth controls. In the absence of growth in the test tubes (with clarithromycin or metronidazole) and the presence of growth in the control tubes, the strain was considered to be sensitive to clarithromycin or metronidazole, respectively. Urease tests and microscopy were used to confirm H. pylori in turbid cultures. In addition, the same bacterial strains at a concentration of 108 CFU/ml were inoculated with a swab onto Mueller-Hinton agar containing 5% horse blood and the Epsilometer test (Etest, AB Biodisk, Solna, Sweden) with clarithromycin and metronidazole was performed according to the recommendations of the British Society for Antimicrobial Chemotherapy (BSAC) . The duration of microaerophilic incubation for the Etest was 3 days.
Selection of liquid culture medium
Final formulation of the optimised culture/indication medium (CIM)
Concentration, per litre
Ham’s F-12 without glutamine
Ferric chloride hexahidrate
HCl 1 N
Comparison of the rapid test with CIM with other H. pylori detection methods using biopsy samples from patients
Sensitivity, specificity, positive (PPV) and negative predictive value (NPV) of H. pylori diagnostic assays: real-time PCR, culture and urease testing using the newly developed CIM, culture on solid medium and microscopy (n=120)
Culture and urease test using CIM
Culture on solid medium
H. pylori loads in biopsy specimens
The density of H. pylori as determined by quantitative TaqMan real-time PCR assay ranged from 10 to 2,051,610 genomes per biopsy specimen, with a mean of 152,073 genomes per biopsy. Among the infected patients, five (5.2%) had more than 106 bacterial genomes per biopsy and nine (9.3%) less than 100 genomes per biopsy. The analysis of stratified data revealed that 25 patients (25.8%) had from 1,000 to 10,000 genomes/biopsy, 22 patients (22.7%) had from 10,000 to 100,000 genomes/biopsy and 36 patients (37.1%) had more than 100,000 genomes/biopsy.
Patients with positive CIM (liquid culture and urease test) were considered to be infected and received antihelicobacter therapy. To assess the efficiency of the therapy, the second biopsy was performed one month after treatment. Eradication was obtained in 58 patients (75.3%) and therapy failure was shown in 19 cases (24.7%). Seven patients were lost to follow-up.
Test tube with microaerophilic atmosphere
To test whether the CO2-enriched test tubes the support growth of H. pylori, the reference strain and six clinical strains were inoculated. The CO2-enriched test tubes provided stable growth of all of the studied cultures within 48 h. By this time, bacterial growth was increased by 3 logs, which is comparable with the growth in microaerophilic conditions in an AnaeroJar, as presented in Fig. 1.
The proposed liquid culture medium can be used for qualitative antibiotic susceptibility testing
The H. pylori reference strain and 18 clinical isolates were tested for antibiotic susceptibility to clarithromycin and metronidazole using either the CIM or the Etest. There was no discrepancy between the two tests for clarithromycin. Two resistant strains were resistant according to both tests. However, from seven cases of metronidazole resistance detected by Etests (minimum Inhibitory Concentration [MIC]>32 mg/l), CIM showed resistance in only six cases. The seventh sample was further analysed: a subsequent Etest carried out using five separate colonies grown from this isolate on solid media produced susceptible and resistant bacteria, showing co-infection with different strains in this patient.
In the present study, a rapid H. pylori test with a liquid CIM was developed that allowed for the isolation, cultivation and identification of H. pylori in the biopsy specimens of patients with gastro-duodenal disorders. The addition of antibiotics to the proposed medium provides rapid qualitative susceptibility testing of H. pylori isolates from biopsy material.
Five different liquid media were tested for their ability to support the growth of H. pylori. All of the tested media provided vigorous growth, regardless of the tested strain, when large inoculum sizes were chosen. On the other hand, only the medium based on Ham’s F-12 was able to support the stable growth of H. pylori when small inoculum sizes had been used. This is in agreement with results from Testerman et al. . Thus, the liquid CIM used for further evaluation was based on a modified Ham’s F-12 medium.
According to our results, the liquid medium with 5% horse serum was superior to a serum-free formulation for the primary isolation of H. pylori from biopsy samples. Moreover, there is evidence that, in the absence of serum proteins, H. pylori becomes sensitive to antibiotics used as selective supplements—polymyxin B and trimetoprim .
There is no commonly used gold standard test for H. pylori infection. The most frequently used gold standard in the evaluation of indirect tests is culture and/or histology. Among the tests used in this study, the most specific and sensitive was the TaqMan real-time PCR. A high specificity and sensitivity of the TaqMan real-time PCR assays has been reported earlier for other targets [39–41]. However, these assays should only be used in specialised laboratories and with the appropriate controls in place. That is why they are usually costly. In addition, the high sensitivity may become questionable when it is used for clinical interpretation. The reaction detects even solitary or dead bacteria. It is not known, however, for how long traces of H. pylori remain in the gastric mucosa after a successful eradication therapy.
A minimal amount of 105 CFU/ml is necessary to obtain a positive result using the commercial rapid urease test CLOtest (Delta West, Bentley, West Australia), which contains 20 g of urea per litre . In our experiments, a delayed indication test was used for the measurement of urease activity after 48 h of incubation. By adding urea to the cultures only after 48 h of incubation, the negative effect that urea may have on the H. pylori viability was avoided [30, 31]. This made it possible to use as little as 103 CFU/ml of H. pylori as the starting material. For this reason, we assume that the novel test with the CIM can provide positive urease tests in patients even after antimicrobial or antisecretory treatment, when urease-positive H. pylori concentrations might be low. Due to the high concentration of urea used in our setup (100 mg/ml) and the high number of bacteria that can be expected after 48 h of incubation, the reaction is very rapid and can be judged within minutes.
The cultivation of H. pylori requires a specific atmosphere composition. Creation of this CO2-enriched atmosphere usually requires the placement of the cell culture flask with the inoculated broth in an incubator with 5% CO2 for 30–120 min. The flasks are then removed, the caps tightened and the flasks incubated with or without shaking for prolonged periods [33, 34]. In the herein proposed system, CO2 was introduced directly into the closed test vials via a PTFE/silicone septum prior to the inoculation of H. pylori. Thus, expensive preincubation with CO2 was not necessary. The CO2, which is heavier than oxygen, replaces the latter from the empty part of the opened vial, so the final O2 concentration is below 20%. The incubation of test tubes in a flat position is a rational and simple method to increase the culture–gas interface. Conflicting results have been reported by other workers in regard to the value of shaking for the growth of H. pylori in flasks with liquid media [17, 34, 43]. Using small-volume vials in our experiments, we considered shaking to be unnecessary.
The growing popularity of anti-helicobacter therapy in low-income countries and the uncontrolled use of generic antimicrobial agents lead to the inevitable appearance and spread of multiresistant strains of H. pylori . It is too expensive to perform routine monitoring of H. pylori antibiotics susceptibility using conventional methods recommended by the Clinical Laboratory Standards Institute  or by BSAC . In addition, simple and rapid antibiotic sensitivity tests will be largely required in treatment failures or, generally, before initiating therapy in patients with H. pylori infection. Thus, a screening susceptibility test for H. pylori based on a liquid medium can be of great value. The proposed tests for metronidazole and clarithromycin can be used after obtaining the liquid cultures in CIM from clinical material. Additional studies will be necessary to assess the accuracy of these tests. A case of discrepancy in the results observed in our study resulted from co-infection of the patient with susceptible and resistant strains of H. pylori. This confirms data from McNulty et al., who reported a 5–10% rate of this kind of co-infection .
In conclusion, the proposed liquid culture/indication test can be used for the primary isolation of H. pylori from gastric biopsy samples without the need for a transport medium. In the presence of serum proteins, CIM provides growth in 48 to 96 h with a minimal starting inoculum of 103 CFU/ml. Further studies are necessary in order to substantiate the value of the CO2-enriched CIM test for the cultivation of H. pylori from clinical samples. The herein described medium can be used for developing a screening test for the H. pylori antibiotics susceptibility testing of clarithromycin and metronidazole.
This study was performed with the financial support of the Swiss National Scientific Foundation (SNF) in the frameworks of the SCOPES 2005–2008 programme (IB73AO-110988/1).
We acknowledge the valuable contributions of Ion Cararus and Ion Chiriac for performing the endoscopic procedures.