Multiple and mixed Helicobacter pylori infections: Comparison of two epidemiological situations in Tunisia and France

Infection, Genetics and Evolution 37 (2016) 43–48

Contents lists available at ScienceDirect

Infection, Genetics and Evolution journal homepage: www.elsevier.com/locate/meegid

Research paper

Multiple and mixed Helicobacter pylori infections: Comparison of two epidemiological situations in Tunisia and France Khansa Ben Mansour a, Chedlia Fendri a, Hajer Battikh a, Martine Garnier b, Meriem Zribi a, Asma Jlizi a, Christophe Burucoa b,⁎ a b

Laboratoire de Microbiologie, CHU La Rabta, Tunis, Tunisia EA 4331 LITEC, Université de Poitiers, CHU de Poitiers, Laboratoire de Bactériologie-Hygiène, Poitiers, France

a r t i c l e

i n f o

Article history: Received 29 June 2015 Received in revised form 16 October 2015 Accepted 26 October 2015 Available online 28 October 2015 Keywords: Helicobacter pylori Genetic diversity Mixed infection Heteroresistance RAPD Multiple infection

a b s t r a c t Individuals can be infected by either a single or multiple strains of Helicobacter pylori. Multiple infection with genetically different isolates and particularly mixed infection with both antibiotic-susceptible and resistant isolates are difficult to detect and should impact the effectiveness of eradication treatment. It is largely assumed that multiple infections are more frequent in developing countries but an actual comparison developing/developed using a single methodology has never been reported. To compare the prevalence of multiple and mixed H. pylori infection in Tunisia and France, we conducted a prospective study including 42 H. pylori-culture positive infected patients (21 Tunisian and 21 French) never previously treated for H. pylori infection. One gastric biopsy was collected from antrum. Three to eleven (mean = 9) colonies were isolated from each biopsy. A total of 375 different isolates were genotyped using RAPD fingerprinting and antimicrobial susceptibility testing was performed on amoxicillin, clarithromycin, ciprofloxacin, rifampicin, tetracycline and metronidazole with E-tests. Multiple infection was defined by different RAPD fingerprintings among the different isolates from a single patient. Mixed infection was defined by different resistance profiles among the different isolates from a single patient. Multiple H. pylori infection is more prevalent in Tunisia than in France. It occurred in ten (48%) Tunisian patients and in one (5%) French patient (p b 0.001). Mixed infection is common (24%), it occurred in 4 (19%) Tunisian patients and in 6 (29%) French patients (p = 0.46) and was mainly (8/10) due to genetically related clones in single infection. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Helicobacter pylori infection is one of the most common chronic bacterial infections in the world and has been established as a major cause of gastritis, peptic ulcer disease, and gastric cancer (Suerbaum and Michetti, 2002). H. pylori is one of the most genetically diverse bacterial species (Suerbaum et al., 1998). It chronically infects the gastric mucosa. Infection is acquired in early childhood and persists over decades, if not lifelong, in the absence of specific treatment. About 50% of the world's population is colonized with a prevalence of over 80% in many developing countries and a rapid decline in developed countries, where prevalence is now estimated at around 20–30% (Frenck and Clemens, 2003; Ben Mansour et al., 2010a). In Tunis, prevalence of H. pylori infection was estimated in a multicenter serological study performed in 2006– 2007 (Ben Mansour et al., 2010a). The seroprevalence in blood donors was 64% and 99% in patients referred for gastroduodenal endoscopy

⁎ Corresponding author at: Laboratoire de Bactériologie-Hygiène, BP 577, 86021 Poitiers, France. E-mail address: [email protected] (C. Burucoa).

http://dx.doi.org/10.1016/j.meegid.2015.10.028 1567-1348/© 2015 Elsevier B.V. All rights reserved.

but since this study was published we have observed a progressive decrease in the prevalence of H. pylori infection diagnosed by culture of biopsy specimens now reaching 50% in Tunis (personal data). In France, where there has been no recently published study, we can estimate the prevalence of H. pylori infection by the prevalence of H. pylori infection diagnosed by culture of French patients' biopsies, which is now only 22% in Poitiers (personal data). This global decrease of H. pylori infection has already been described and is believed to be due to improved hygiene, and active elimination by antibiotics during childhood, both of which have contributed to declining transmission risk (den Hollander et al., 2015). Individuals can be infected by either a single or multiple strains of H. pylori. A multiple infection is defined as the infection of a single patient by two or more genetically distinct isolates. The genetic distinction of isolates can be performed using genotyping of virulence genes (cagA, vacA, iceA) (Audibert et al., 2000; Ben Mansour et al., 2010b; Morales-Espinosa et al., 1999; Patra et al., 2012), random amplified polymorphism DNA (RAPD) (Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Ren et al., 2012; Sheu et al., 2009; Hua et al., 1999; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Toita et al., 2013; Enroth et al., 1999; Marshall

44

K. Ben Mansour et al. / Infection, Genetics and Evolution 37 (2016) 43–48

et al., 1995), Multilocus Sequence Typing (MLST) (Suerbaum et al., 1998) or whole genome sequencing (Kenneman et al., 2011; Krebes et al., 2014). The prevalence of such multiple infections has been reported to vary (0–100%) (Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Ren et al., 2012; Sheu et al., 2009; Hua et al., 1999; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Toita et al., 2013; Enroth et al., 1999; Marshall et al., 1995; Dore et al., 1998) depending on geographical region, whether in a developed or developing country (low and high overall infection risk, respectively), but probably also according to the methods used to sample isolates from the stomach or on account of the discriminatory power of the strain typing method used in different studies. It is largely assumed that multiple infections are more frequent in developing countries but an actual comparison developing/developed using a single methodology has never been reported. In addition to multiple infection, which is characterized by genetic diversity, an individual can be infected by both antibiotic-susceptible and resistant H. pylori strains (Wong et al., 2001; Yakoob et al., 2001; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Raymond et al., 2010; Ben Mansour et al., 2010c). These strains can be either derived from the same genetic isolate or genetically totally different (multiple infection). In the first case it is assumed that the resistant clone emerges from the susceptible one under selective pressure due to antibiotic consumption. In the second case, the patient has been infected by two different strains, one susceptible and the other resistant. Mixed infection is defined by the infection of a single patient with two or more isolates presenting different antibiotic susceptibility profiles. Mixed infection is usually detected by the presence of a subpopulation of resistant colonies in the inhibitory area of a disk or E-test diffusion test or by discordant results of susceptibility testing in two biopsies from the same patient (Wong et al., 2001; Kao et al., 2014). Mixed infection is difficult to detect and should impact the effectiveness of eradication treatment if the resistant strain is not detected. For precise determination of both multiple and mixed infections it is important to work on pure isolates obtained from isolated colonies and not from the mixture of different colonies harvested together on the first culture plate (Wong et al., 2001). Inworks on mixtures of different colonies or by PCR directly on DNA extracted from a single biopsy, multiple infection can be suspected when vacA genotyping yields a mixture of the different alleles m1 and m2, s1 and s2 (Audibert et al., 2000; Ben Mansour et al., 2010b). In Tunisia, in a previous study, the prevalence of multiple infections was estimated at 31.4% for the vacA genotyping performed directly on DNA extracted from biopsy (Ben Mansour et al., 2010b). In France, in a previous study performed in Poitiers, the prevalence of multiple infections was estimated at 2.6% for vacA genotyping performed directly on DNA extracted from biopsy (Audibert et al., 2000). Virulence gene genotyping is not sensitive enough to detect multiple infections and certainly underestimates this phenomenon. RAPD has been proven to achieve an excellent discriminatory power allowing precisely distinguishing H. pylori isolates (Burucoa et al., 1999) and has been widely used to identify different genotypes among and within individual patients (Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Ren et al., 2012; Sheu et al., 2009; Hua et al., 1999; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Toita et al., 2013; Enroth et al., 1999; Marshall et al., 1995). The aim of this study was to compare the prevalence of multiple and mixed H. pylori infection between two locations in different epidemiological situations: Tunis in a developing country with high prevalence of infection and Poitiers in a developed country with low prevalence of infection. With this in mind we chose to work on multiple isolated colonies from 21 Tunisian and 21 French H. pylori infected patients in view of genotyping isolates using RAPD analysis and of determining antibiotic susceptibility on isolated colonies.

2. Material and methods 2.1. Patients This prospective study was conducted from June to October 2009 in two University Hospitals: CHU La Rabta in Tunis, Tunisia and CHU La Milétrie in Poitiers, France. Inclusion criteria were, for each center, the first 21 consecutive adult patients of both genders referred for gastroduodenal endoscopy, detected H. pylori-infected by a positive H. pylori culture from the antrum biopsy specimen and never previously treated for H. pylori infection. One gastric biopsy was routinely collected for bacteriological diagnosis of H. pylori infection in the antrum during upper gastrointestinal endoscopy. The demographic characteristics, including age, gender, results of endoscopic examination and histology were collected for each patient. 2.2. Strains Biopsy was crushed and culture was performed in the same manner in the two centers following the same recommendations (Burucoa and Mégraud, 2011). Cultures were carried out on Columbia agar plates supplemented with 10% horse blood and Skirrow's supplement (trimethoprim 5 mg/L, vancomycin 10 mg/L, polymyxin B 2500 IU/L, Oxoid, France) at 37 °C in micro-aerophilic conditions for 3–7 days (CampyGen*, Oxoid, France). Depending on the number of colonies obtained in primary culture, three to eleven single colonies per patient (mean = 9) were picked from the primary culture plates and were subcultured so as to obtain pure isolates. H. pylori isolates were identified as positive for urease activity and spiral shape morphology on Gram stain. 2.3. DNA extraction and RAPD genotyping Genomic DNA was extracted from each isolate of H. pylori using QIAamp DNA mini-kit (Qiagen, France) according to the manufacturer's instructions. Random amplified polymorphism DNA reaction (RAPDPCR) was carried out as previously described (Burucoa et al., 1999). RAPD-PCR was performed in a Perkin-Elmer GeneAmp PCR system 2400 thermal cycler (Perkin-Elmer Cetus, USA) in 100 μL containing 1 μL of chromosomal DNA (~20 ng), 3 mM MgCl2, each primer at a concentration of 0.2 μM, 2.5 U of Eurotaq DNA polymerase (Eurogentec, France), each dinucleotide triphosphate (Eurogentec, France) at a concentration of 0.2 μM, 10 mM Tris–HCl (pH 8.3), and 50 mM KCl (Burucoa et al., 1999). Two arbitrary primers were used: primer 1254 (5′-CCGCAGCCAA-3′) and 1247 (5′-AAGAGCCCGT-3′) (Burucoa et al., 1999). The cycling program was 1 cycle of 94 °C for 2 min, 37 °C for 1 min, and 72 °C for 4 min and 29 cycles of 94 °C for 2 min, 37 °C for 3 min, and 72 °C for 7 min. After PCR, 20 μL of the amplification products was electrophoretically separated in 2% agarose gels. 2.4. Definition of multiple infection using RAPD genotyping For each patient, genomic diversity among the different colonies (3 to 11, mean = 9) of H. pylori isolates was analyzed by RAPD genotyping. Criteria defined by Tenover et al. were used for interpreting RAPD banding patterns (Tenover et al., 1995). Indistinguishable and closely related RAPD banding patterns in the isolates of one patient were defined as single-strain infection. Bacterial infection with more than one RAPD banding pattern found in the isolates of one patient was defined as multiple infection. 2.5. Antimicrobial susceptibility testing Susceptibility testing for amoxicillin, clarithromycin, tetracycline, metronidazole, ciprofloxacin and rifampicin was performed with the E-test method (bioMérieux, France) on Columbia agar plates

K. Ben Mansour et al. / Infection, Genetics and Evolution 37 (2016) 43–48

supplemented with 10% horse blood (Oxoid, France) as recommended (Burucoa and Mégraud, 2011). Inocula were prepared from 2-day-old agar plates, colonies were harvested in 2 mL of Brucella broth. The final inoculum was adjusted to an opacity equivalent to a 3 Mcfarland turbidity standard (approximately 108 cfu/mL) and was spread on agar plates. Minimal inhibitory concentration cut-offs to define resistance were 0.12 mg/L for amoxicillin, 0.25/0.50 mg/L for clarithromycin, 1 mg/L for tetracycline, rifampicin and ciprofloxacin, and 8 mg/L for metronidazole. 2.6. Definition of mixed infection Mixed infection was defined by the presence among the different isolates from a single patient of susceptible and resistant isolates for the different antibiotics tested. 2.7. Statistics Categorical data were analyzed by the Chi2 test and Fisher's exact test, while continuous variables were analyzed by the Student t test. p-Values lower than 0.05 were considered statistically significant. 3. Results The 21 Tunisian H. pylori-infected patients were included in 5 months among 45 consecutively biopsied patients referred for gastroduodenal endoscopy in Tunis (47%). The 21 French H. pylori-infected patients were included in 2 months among 104 consecutively biopsied patients referred for gastro-duodenal endoscopy in Poitiers (20%). The prevalence of H. pylori-positive culture patients among patients referred for gastro-duodenal endoscopy is significantly higher in Tunis than in Poitiers (p = 0.0009). Patient baseline characteristics were not statistically different according to inclusion center (Table 1). Raw data for 375 isolates from the 21 Tunisian biopsies and the 21 French biopsies are presented in Table 2. The two primers used for RAPD analysis generated DNA fingerprinting for all 375 H. pylori isolates examined (typeability = 100%) (Figs. 1 and 2). Among the 375 isolates tested from the 42 enrolled patients, RAPD analysis yielded 57 distinguishable patterns. The two primers used discriminated the 375 isolates in the same manner although with different patterns, confirming the reliability of the discrimination. The RAPD profiles of isolates from a particular patient were distinct from the RAPD profiles of isolates from other patients (Table 2). Two (8), three (2), or four (1) different patterns were observed among the isolates of 11 patients indicating a multiple infection (Fig. 2). Multiple H. pylori infection occurred in 10/21 (48%) Tunisian patients and in 1/21 (5%) French patient (p b 0.001). The demographic characteristics, including age, female/male ratio and the proportion of patients with peptic ulcer, were not significantly different between the 11 patients with a multiple infection and the 31 patients with a single-strain infection. All 375 successfully cultured H. pylori isolates from the 42 patients, naïve to any eradication therapy before, were assayed against amoxicillin, clarithromycin, metronidazole, ciprofloxacin, tetracycline, and Table 1 Patient's baseline characteristics according to inclusion center. Hospital

Mean age in years

CHU La Rabta Tunis 52.8 Tunisie CHU La Milétrie Poitiers 54.5 France

Gender N° M/N° F

Ulcer disease N° (%)

N° of patients

13/8

4 (19%)

21

11/10

3 (14%)

21

All characteristics and results reported in this table were not statistically different between the two inclusion centers.

45

Table 2 Genotyping and susceptibility typing of the 375 isolates from 21 Tunisian and 21 French patients. N° biopsy

Country

Age in years

Gender

Ulcer disease

N° of colony

RAPD profile

Cla

Mtz

Cip

1

T

66

M

N

2

T

54

F

N

3 4

T T

60 34

M M

N N

5 6

T T

65 55

F M

N N

7 8 9

T T T

67 30 52

F F M

N N Y

10

T

69

M

N

11

T

43

F

N

12

T

54

F

N

13 14

T T

50 40

M M

N Y

15

T

82

M

N

16 17 18

T T T

38 37 49

F F M

N N N

19 20

T T

64 50

M M

Y Y

21

T

51

M

N

22 23 24 25 26

F F F F F

33 42 57 56 87

M F F M F

N N Y N N

27 28

F F

29 28

M M

Y N

29 30 31

F F F

53 70 46

M F F

N N N

32 33

F F

43 83

F F

N N

34 35 36 37 38 39

F F F F F F

62 46 75 58 45 50

M F F M M M

N N N N N Y

40 41 42

F F F

69 88 25

M F M

N N N

3 1 5 3 2 3 4 4 8 4 3 3 10 10 8 2 5 4 1 7 3 6 4 10 7 3 7 3 10 10 6 3 9 6 4 9 1 8 10 8 10 6 4 1 9 3 3 1 6 8 7 2 10 6 1 7 7 7 9 10 6 5 10 10 6 2 1 1

T1 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T16 T16 T17 T18 T19 T20 T21 T22 T23 T24 T25 T26 T27 T28 T29 T30 T31 T32 T33 T33 T34 T35 T36 T37 T38 T38 T38 T39 T40 T40 T40 T41 T42 T43 T43 T44 T45 T45 T46 T47 T48 T49 T50 T51 T51 T52 T53 T54 T55 T56 T57

S S S S S S S S S S S S S S S S R R S S S S S S S S S S S S S S S R R S R R R R S R S R S S R R S S S S S S S S S S R R S R S S R R R S

S R S S R S S S S S S S S S S S R S S S S S S S S S S S S S S S S R R S S R R R R R R R R S S R R S S R R S R R R S R R R R S S R R R S

S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S R S S S S S R R R S S S R S S S S S S S S S S S S

T: Tunisia, F: France, M: male, F: female, N: no, Y: yes, Cla: clarithromycin, Mtz: metronidazole, Cip: ciprofloxacin, S: susceptible, R: resistant.

46

K. Ben Mansour et al. / Infection, Genetics and Evolution 37 (2016) 43–48

(8/10) in cases of single infection. Metronidazole resistance was involved in 7 cases, clarithromycin resistance in 6 cases and ciprofloxacin resistance in 1 case. 4. Discussion

Fig. 1. Random amplified polymorphism DNA (RAPD) banding patterns of 11 isolates from the Tunisian patient n° 26 were obtained with the primer 1254. Isolates A–F were clarithromycin resistant, metronidazole resistant, and ciprofloxacin susceptible. Isolates G–J were clarithromycin susceptible, metronidazole resistant, and ciprofloxacin susceptible. Isolate K was clarithromycin resistant, metronidazole resistant, and ciprofloxacin resistant.

rifampicin (Table 2). In none of the patients, resistance to amoxicillin, tetracycline or rifampicin was detectable. Primary resistance to clarithromycin was detected in 12 patients (29%), 9 French patients (43%) and 3 Tunisian patients (14%) (p = 0.04). Primary resistance to metronidazole was detected in 21 patients (50%), 17 French patients (81%) and 4 Tunisian patients (19%) (p = 0.0001). Primary resistance to ciprofloxacin was detected in 4 patients (10%), 4 French patients (19%) and no Tunisian patients (0%) (p = 0.035). A different antibiotic susceptibility pattern between the different isolates (3 to 11 per patient), defining mixed infection, was detected in 10 patients (24%). Mixed infection concerning two different antibiotics was detected in 3 patients, while 7 patients showed a discordant susceptibility against only one antibiotic agent only. Mixed infection occurred in 4 (19%) Tunisian patients and in 6 (29%) (p = 0.46) French patients, mainly

This study is unique, in that two countries with radically different epidemiological situations were compared using the same methodology. To our knowledge, it is the first to compare multiple and mixed infection prevalences in a developing country with high prevalence of H. pylori infection (Tunisia) and in a developed country of low, declining prevalence (France) using the same methods to sample isolates from the stomach, to genotype, and to determine the antibiotic susceptibility of these isolates. In this study, three to eleven single colonies from 42 patients (all in all, 375 single colonies, a mean of 9 colonies per patient) were isolated with the same protocol in the two centers. The relatively low number of studied patients is explained by the material limitation to test 375 isolates with 2 different RAPD assays and to determine the MIC of these isolates to 6 antimicrobial agents with the E-test. All these tests are time-consuming and expensive for a developing country. When regarding the 10 published studies working on isolated colonies of H. pylori, 12 to 46 patients were studied (mean 25) and 30 to 449 isolates (mean 250) were tested (Morales-Espinosa et al., 1999; Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Sheu et al., 2009; Norazah et al., 2009; Toita et al., 2013; Enroth et al., 1999; Marshall et al., 1995; Dore et al., 1998). With 42 patients and 375 isolated colonies tested, our study is among the most extensive. Typing and antimicrobial susceptibility testing applied on isolated colonies are more effective than when applied on pools of colonies from a primary culture. Wong et al. (2001), for example, have demonstrated better discrimination and higher prevalence of both multiple and mixed infections when working on isolated colonies rather than pools. In our study, genotyping was not, as has been the case in other works, applied exclusively to resistant or discordant susceptibility isolates, but rather to unselected isolates, thereby yielding more relevant results while avoiding recruitment bias. One weakness of our study is the relatively low number of gastric sites sampled. As one biopsy specimen from the antrum does not adequately represent the whole stomach surface area, our sample methodology could have underestimated multiple infections and heteroresistance. But the above-mentioned criticism can be addressed to all existing studies, regardless of the more or less large number of biopsies collected. In point of fact, exhaustive sampling of the whole

Fig. 2. Random amplified polymorphism DNA (RAPD) banding patterns of 10 isolates from the French patient n° 42 were obtained with the primer 1254. Isolate A: reproducibility control. Isolate B was clarithromycin resistant, metronidazole resistant, and ciprofloxacin susceptible. Isolates C and G were clarithromycin resistant, metronidazole resistant, and ciprofloxacin susceptible. Isolates D–F, H, J, and K were clarithromycin resistant, metronidazole resistant, and ciprofloxacin susceptible. Isolate I was clarithromycin susceptible, metronidazole susceptible, and ciprofloxacin susceptible.

K. Ben Mansour et al. / Infection, Genetics and Evolution 37 (2016) 43–48

gastric surface area is unachievable since it is estimated to range from 150 to 200 cm2 while a gastric biopsy covers only a few mm2 (Henry et al., 2007). Among several genotyping methods applied to H. pylori, RAPD-PCR is considered to be useful because it is a simple, rapid, and low-cost means of distinguishing H. pylori isolates (Burucoa et al., 1999). While MLST and whole genome sequencing are adapted to phylogenic studies, they are time-consuming and very expensive for emerging countries. And in any case, RAPD typing has conclusively evidenced high discriminatory power and made significant contributions to several studies (Morales-Espinosa et al., 1999; Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Ren et al., 2012; Sheu et al., 2009; Hua et al., 1999; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Toita et al., 2013; Enroth et al., 1999). Moreover, detection of the genetic diversity of isolates from patients with single and multiple infection has proven highly consistent using two different primers. In two previous studies performed in the same places (Tunis, Tunisia and Poitiers, France), the prevalence of multiple infection estimated by vacA genotyping was 31.4% in Tunisian patients and 2.6% in French patients compared to 48% and 5% respectively in the study using RAPD typing, highlighting the risk of underestimating multiple infection when virulence gene genotyping is used to discriminate isolates (Ben Mansour et al., 2010b; Audibert et al., 2000). Our study demonstrates a higher prevalence of multiple infections in a developing country compared to a developed country. This can be explained by the differing prevalence of the infection in the two countries. Thanks to improved hygiene and sanitation as well as antimicrobial treatments over the last decades, prevalence of H. pylori infection has declined (den Hollander et al., 2015). As regards Tunisia, the prevalence of H. pylori infection was evaluated in a multicenter serological study (Ben Mansour et al., 2010a). The seroprevalence in blood donors was 64% and 99% in patients referred for gastro-duodenal endoscopy. In France, on the other hand, prevalence of H. pylori infection is estimated at 30% (there has been no recently published study). As an estimation of the relative prevalence observed in the two countries (Tunisia and France), we can compare the prevalence of H. pylori-culture positive patients among patients referred for gastro-duodenal endoscopy and never treated for H. pylori infection in Tunis, Tunisia (47%) and in Poitiers, France (20%). As is the case in many other developing countries, in Tunisia the prevalence of H. pylori infection is high and the chances for a single Tunisian host to be infected or reinfected by different strains are greater than in France, a developed country where prevalence is low. Our results are in concordance with numerous reports on the prevalence of multiple infections reported in developing or in developed countries. The prevalence of 48% we observed in Tunisia can be compared to the prevalence comprised between 85 and 100% found in India, China or Mexico (Morales-Espinosa et al., 1999; Patra et al., 2012; Wong et al., 2001; Yakoob et al., 2001; Ren et al., 2012) and 20– 35% found in emerging countries Malaysia and Taiwan (Sheu et al., 2009; Norazah et al., 2009). The prevalence of 5% we observed in France is within the 0–10% range usually observed in developed countries such as Germany, The Netherlands, Japan, Sweden, Ireland, or Singapore (Hua et al., 1999; Selgrad et al., 2014; Van Der Ende et al., 2001; Toita et al., 2013; Enroth et al., 1999; Marshall et al., 1995; Dore et al., 1998). It is interesting to note that the only French patient with multiple infection was born in Senegal (patient n° 42 in Table 2). This observation is representative of the diversity of the French population, which comprises 11% of immigrants. Multiple infections with H. pylori facilitate interstrain gene transfer and the maintenance of genetic diversity (Kenneman et al., 2011; Krebes et al., 2014). As discussed by other authors, we would be inclined to predict in developed countries, given the decline in prevalence and the low frequency of multiple infections, a future diminution in the genetic diversity of H. pylori (Kenneman et al., 2011; Krebes et al., 2014).

47

Even with a small number of patients, we observed statistically different prevalences of primary resistance between France and Tunisia with higher prevalence of primary resistance in France than in Tunisia (29% vs. 14% for clarithromycin, 81% vs 19% for metronidazole, and 19% vs 0% for ciprofloxacin). These results are in accordance with previous results of independent studies in these two countries (Raymond et al., 2010; Ben Mansour et al., 2010c). These differences in antibiotic resistance levels between Tunisia and France could be explained by differences in antibiotic consumption. In Europe, a significant positive association between outpatient antibiotic usage and the level of primary resistance observed in H. pylori to key antimicrobial agents has been demonstrated (Megraud et al., 2013). In the absence of accurate knowledge on antibiotic consumption in Tunisia, we can only suspect a highest consumption in France. In our study we have demonstrated that mixed infection characterized by differing antibiotic susceptibilities among the different isolates of a single patient, is a common phenomenon in patients who are naive to eradication treatment (24%). Mixed infection is more prevalent in French patients (29%) than in Tunisian patients (19%) but without any statistically significant difference (p = 0.47). Mixed infection is likely to be more frequent in areas of higher prevalence of antibiotic resistance. It seems logical that the chance of detecting a mixed infection is higher in the case of high prevalence of antibiotic resistance. While previous studies have found comparably high discordant antibiotic susceptibility among different isolates of a single patient (Wong et al., 2001; Yakoob et al., 2001; Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001; Dore et al., 1998; Raymond et al., 2010; Ben Mansour et al., 2010c, 24–26), they mainly compared antibiotic susceptibility between antrum and corpus biopsies, without attempting to isolate single colony isolates, and were consequently likely to underestimate mixed infection prevalence (Wong et al., 2001). According to our results and those of previous studies, antibiotic resistance is indeed likely to be underestimated, or even overlooked, if the different isolates having been isolated from single colonies of a biopsy are not tested for antimicrobial susceptibility. All in all, this factor should impact the effectiveness of H. pylori therapy. But, from a practical standpoint and as regards economic considerations, this type of protocol is difficult to apply in routine practice. One interesting aspect of the study is the fact that mixed infection (susceptible and resistant isolates) occurs mainly among patients with single infection (unique RAPD fingerprint) (80%). Indeed, eight of the ten patients with mixed infection showed identical fingerprinting patterns, thereby suggesting an infection with a single H. pylori strain. Only in two patients were clear differences in the DNA pattern detectable, thereby suggesting multiple infection. This finding is in accordance with previous studies showing that DNA fingerprinting patterns of different isolates with different susceptibilities from a single patient are identical, which may mean that antibiotic-resistant H. pylori strains typically develop from a pre-existing susceptible strain rather than from co-infection with different strains (Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001). Metronidazole resistance, clarithromycin resistance and ciprofloxacin resistance are due to point mutations into rdxA and frxA, 23S rRNA, and gyrA genes respectively, which are not detected by RAPD-PCR (Garcia et al., 2012; Burucoa et al., 2005). That is the reason why resistant and susceptible isolates may present an identical RAPD profile. The mutations that confer resistance to different antibiotics used in eradication treatments occur spontaneously during replication errors at very low frequency. Antibiotic treatment with one of these molecules or a member of their antibiotic family will select a mutant that will occupy the stomach left vacant after the death of sensitive clones. The resistant clone will be responsible for a failure of eradication therapy. The selection pressure exerted at a population level with a high consumption of antibiotics will be responsible for the emergence and dissemination of resistance in the population and a high prevalence of primary resistance.

48

K. Ben Mansour et al. / Infection, Genetics and Evolution 37 (2016) 43–48

Mixed infection could be explained to a greater extent by the emergence under selective pressure of resistant subclones of the susceptible isolate infecting the patient than by co-infection of a susceptible and a genetically distinct and resistant strain. As in previous studies (Norazah et al., 2009; Kao et al., 2014; Selgrad et al., 2014; Van Der Ende et al., 2001), mixed infection was most common for metronidazole. This could be explained by the higher prevalence of metronidazole resistance (Raymond et al., 2010; Ben Mansour et al., 2010c). In conclusion, multiple H. pylori infection is more prevalent in Tunisia (48%), a high endemic country, than in France (5%), a low endemic country. Mixed infection is common (24%) and occurs mainly (80%) in genetically related clones. Acknowledgments We gratefully acknowledge Mr. Jeffrey Arsham, an American translator, for his rereading and revision of the original English-language manuscript. This work was supported in part by the CHU de Poitiers n° 78001-2013 and by a PHC Utique n° 30570XA. References Audibert, C., Janvier, B., Grignon, B., Salaun, L., Burucoa, C., Lecron, J.C., Fauchère, J.L., 2000. Correlation between IL-8 induction, cagA status and vacA genotypes in 153 French Helicobacter pylori isolates. Res. Microbiol. 151, 191–200. Ben Mansour, K., Burucoa, C., Zribi, M., Masmoudi, A., Karoui, S., Kallel, L., Chouaib, S., Matri, S., Fekih, M., Zarrouk, S., Labbene, M., Boubaker, J., Cheikh, I., Hriz, M.B., Siala, N., Ayadi, A., Filali, A., Mami, N.B., Najjar, T., Maherzi, A., Sfar, M.T., Fendri, C., 2010a. Primary resistance to clarithromycin, metronidazole and amoxicillin of Helicobacter pylori isolated from Tunisian patients with peptic ulcers and gastritis: a prospective multicentre study. Ann. Clin. Microbiol. Antimicrob. 9, 22. http://dx.doi.org/10. 1186/1476-0711-9-22. Ben Mansour, K., Fendri, C., Zribi, M., Masmoudi, A., Labbene, M., Fillali, A., Ben Mami, N., Najjar, T., Meherzi, A., Sfar, T., Burucoa, C., 2010b. Prevalence of Helicobacter pylori vacA, cagA, iceA and oipA genotypes in Tunisian patients. Ann. Clin. Microbiol. Antimicrob. 9, 1–7. Ben Mansour, K., Keita, A., Zribi, M., Masmoudi, A., Zarrouk, S., Labbene, M., Kallel, L., Karoui, S., Fekih, M., Matri, S., Boubaker, J., Cheikh, I., Chouaib, S., Filali, A., Mami, N.B., Najjar, T., Fendri, C., 2010c. Seroprevalence of Helicobacter pylori among Tunisian blood donors (outpatients), symptomatic patients and control subjects. Gastroenterol. Clin. Biol. 34, 75–82. Burucoa, C., Mégraud, F., 2011. Helicobacter pylori. In: Cornaglia, G., Courcol, R., Herrmann, J.L., Kahimeter, G., Pelgue-Lafeuille, H., Vila, J. (Eds.), European Manual of Clinical Microbiology. European Society of Clinical Microbiology and Infectious Diseases, Basel, pp. 283–286. Burucoa, C., Lhomme, V., Fauchere, J.L., 1999. Performance criteria of DNA fingerprinting methods for typing of Helicobacter pylori isolates: experimental results and metaanalysis. J. Clin. Microbiol. 37, 4071–4080. Burucoa, C., Landron, C., Garnier, M., Fauchere, J.L., 2005. T2182C mutation is not associated with clarithromycin resistance in Helicobacter pylori. Antimicrob. Agents Chemother. 49, 868–870. den Hollander, W.J., Holster, I.L., van Gilst, B., van Vuuren, A.J., Jaddoe, V.W., Hofman, A., Perez-Perez, G.I., Kuipers, E.J., Moll, H.A., Blaser, M.J., 2015. Intergenerational change in Helicobacter pylori colonization in children living in a multi-ethnic Western population. Gut 64, 1200–1208. Dore, M.P., Osato, M.S., Kwon, D.H., Graham, D.Y., El Zaatari, A.K., 1998. Demonstration of unexpected antibiotic resistance of genotypically identical Helicobacter pylori isolates. Clin. Infect. Dis. 27, 84–89. Enroth, H., Nyren, O., Engstrand, L., 1999. One stomach–one strain: does Helicobacter pylori strain variation influence disease outcome? Dig. Dis. Sci. 44, 102–107. Frenck, R.W., Clemens, J., 2003. Helicobacter in the developing world. Microbes Infect. 5, 705–713.

Garcia, M., Raymond, J., Garnier, M., Cremniter, J., Burucoa, C., 2012. Distribution of spontaneous gyrA mutations in 97 fluoroquinolone-resistant Helicobacter pylori isolates collected in France. Antimicrob. Agents Chemother. 56, 550–551. Henry, J.A., O'Sulivan, G., Pandit, A.S., 2007. Using computed tomography scans to develop an ex-vivo gastric model. World J. Gastroenterol. 13, 1372–1377. Hua, J., Ng, H.C., Yeoh, K.G., Ho, B., 1999. Predominance of a single strain of Helicobacter pylori in gastric antrum. Helicobacter 4, 28–32. Kao, C.Y., Lee, A.Y., Huang, A.H., Song, P.Y., Yang, Y.J., Sheu, S.M., Chang, W.L., Sheu, B.S., Wu, J.J., 2014. Heteroresistance of Helicobacter pylori from the same patient prior to antibiotic treatment. Infect. Genet. Evol. 23, 196–202. Kenneman, L., Didelot, X., Aebischer, T., Kuhn, S., Drescher, B., Droege, M., Reinhart, R., Correa, P., Meyer, T.F., Josenhans, C., Falush, D., Suerbaum, S., 2011. Helicobacter pylori genome evolution during human infection. Proc. Natl. Acad. Sci. U. S. A. 108, 5033–5038. Krebes, J., Didelot, X., Kenneman, L., Suerbaum, S., 2014. Bidirectional genomic exchange between Helicobacter pylori strains from a family in Coventry, United Kingdom. Int. J. Med. Microbiol. 304, 1135–1146. Marshall, D.G., Chua, A., Keeling, N.P.W., Sullivan, D.J., Colemen, D.C., Smyth, C.J., 1995. Molecular analysis of Helicobacter pylori populations in antral biopsies from individual patients using randomly amplified polymorphic DNA (RAPD) fingerprinting. FEMS Immunol. Med. Microbiol. 10, 317–324. Megraud, F., Coenen, S., Versporten, A., Kist, M., Lopez-Brea, M., Hirschl, A.M., Andersen, L.P., Goossens, H., Glupczynski, Y., Study Group participants, 2013. Helicobacter pylori resistance to antibiotics in Europe and its relationship to antibiotic consumption. Gut 62, 34–42. Morales-Espinosa, R., Castillo-Rojas, G., Gonzales-Valencia, G., Ponce de Leon, S., Cravioto, A., Atherton, J.C., Lopez-Vidal, Y., 1999. Colonization of Mexican patients by multiple Helicobacter pylori strains with different vacA and cagA genotypes. J. Clin. Microbiol. 37, 3001–3004. Norazah, A., Wan Rasinah, Z., Zaili, Z., Aminuddin, A., Ramelah, M., 2009. Analysis of PCRRAPD DNA and antibiotic susceptibility profiles of antrum and corpus isolates of Helicobacter pylori from Malaysian patients. Malays. J. Pathol. 31, 29–34. Patra, R., Chattopadhyay, S., De, R., Gosh, P., Ganguly, M., Chowdhury, A., Ramamurthy, T., Nair, G.B., Mukhopadhyay, A.K., 2012. Multiple infection and microdiversity among Helicobacter pylori isolates in a single host in India. PLoS One 7 (e43370), 0–11. Raymond, J., Lamarque, D., Kalach, N., Chaussade, S., Burucoa, C., 2010. High level of antimicrobial resistance in French Helicobacter pylori isolates. Helicobacter 15, 21–27. Ren, L., Liao, Y.L., Song, Y., Guo, Y., Mao, X.H., Xie, Q.H., Zhang, W.J., Guo, G., Zou, Q.M., 2012. High frequency variations of Helicobacter pylori isolates in individual hosts in a Chinese population. Int. J. Infect. Dis. 16, e358–e363. Selgrad, M., Tammer, I., Langner, C., Bornschein, J., Meisle, J., Kandulski, A., Varbanova, M., Wex, T., Schluter, D., Malfertheiner, P., 2014. Different antibiotic susceptibility between antrum and corpus of the stomach, a possible reason for treatment failure of Helicobacter infection. World J. Gastroenterol. 20, 16245–16251. Sheu, S.M., Sheu, B.S., Lu, C.C., Yang, H.B., Wu, J.J., 2009. Mixed infections of Helicobacter pylori: tissue tropism and histological significance. Clin. Microbiol. Infect. 15, 253–259. Suerbaum, S., Michetti, P., 2002. Helicobacter pylori infection. N. Engl. J. Med. 347, 1175–1186. Suerbaum, S., Maynard Smith, J., Bapumia, K., Morelli, G., Smith, N.H., Kunstmann, E., Dyrek, I., Achtman, M., 1998. Free recombination within Helicobacter pylori. Proc. Natl. Acad. Sci. U. S. A. 95, 12619–12624. Tenover, F.C., Arbeit, R.A., Goering, R.V., Mickelsen, P.A., Murray, B.E., Persing, D.H., Swaminathan, B., 1995. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J. Clin. Microbiol. 33, 2233–2239. Toita, N., Yokota, S., Fujii, N., Konno, M., 2013. Clonality analysis of Helicobacter pylori in patients isolated from several biopsy specimens and gastric juice in Japanese urban population by random amplified polymorphic DNA fingerprinting. Gastroenterol. Res. Pract. 2013, 721306. Van Der Ende, A., Van Doorn, L.J., Rooijakkers, S., Feller, M., Tytgat, G.N.J., Dankert, J., 2001. Clarithromycin-susceptible and-resistant Helicobacter pylori isolates with identical randomly amplified polymorphic DNA-PCR genotypes cultured from single gastric biopsy specimen prior to antibiotic therapy. J. Clin. Microbiol. 39, 2648–2651. Wong, B.C., Wang, W.H., Berg, D.E., Fung, F.M.Y., Wong, K.W., Wong, W.M., Lai, K.C., Cho, C.H., Hui, W.M., Lam, S.K., 2001. High prevalence of mixed infections by Helicobacter pylori in Hong Kong: metronidazole sensitivity and overall genotype. Aliment. Pharmacol. Ther. 15, 493–503. Yakoob, J., Fan, X.G., Hu, G.L., Yang, H.X., Liu, L., Liu, S.H., Tan, D.M., Li, T.G., Zhang, Z., 2001. Polycolonization of Helicobacter pylori among Chinese subjects. Clin. Microbiol. Infect. 7, 187–192.