Abstract
The aim of this study was to characterize a broad collection of isolates of Listeria monocytogenes, of different serotypes, recovered in Portugal between 2003 and 2007 from foods (n=353) and from clinical cases of human listeriosis (n=95), in terms of antimicrobial susceptibility. All the isolates were susceptible to ampicillin, the preferred agent to treat listeriosis. Resistances to nitrofurantoin (n=99), to ciprofloxacin (n=18), to erythromycin (n=10), to tetracycline (n=2), to gentamicin (n=1) and to rifampicin (n=1) were observed. One hundred (28.3%) and 20 (21.0%) food and clinical isolates, respectively, were resistant to at least one antibiotic. Eight isolates (1.8%) were resistant to two or more antimicrobials of different classes, and all were collected from foods. Serogroup IVb included the highest percentage of isolates resistant to erythromycin. The highest percentages of isolates resistant to nitrofurantoin were of serogroup IVb and IIc. It was demonstrated that the incidence of antibiotic-resistant isolates of L. monocytogenes, during the period 2003 to 2007, was low in Portugal but still higher than that observed in other countries. Given the increasing population at greater risk of listeriosis, namely, the elderly, the high mortality rate of the infection and the detection of resistant isolates, monitoring for antibiotic resistance in strains of L. monocytogenes on a large scale, and assessing the risk of infection by these strains, is highly recommended.
Introduction
L
Although invasive listeriosis is often associated with individuals who belong to risk groups (older adults, pregnant women, newborns, and immunocompromised individuals), listeriosis is a bacterial foodborne illness. L. monocytogenes can cause illness—listeriosis among healthy individuals (Ooi and Lorber, 2005).
Listeriosis often requires antimicrobial therapy. The treatment of choice consists of a β-lactam antibiotic, normally ampicillin, alone or in combination with an aminoglycoside, classically gentamicin. Second-line agents in cases of allergy to β-lactams or in certain disease states include trimethoprim/sulfamethoxazole, erythromycin, vancomycin, and the fluoroquinolones (Temple and Nahata, 2000; Allerberger and Wagner, 2010).
Although Listeria species were thought to be uniformly susceptible to all antibiotics active against Gram-positive bacteria for many years, since the late 1980s, an increasing number of reports have documented the existence of antibiotic-resistant strains of L. monocytogenes isolated from foods (Antunes et al., 2002; Conter et al., 2009; Ayaz and Erol, 2010; O'Connor et al., 2010; Pesavento et al., 2010; Lungu et al., 2011; Ruiz-Bolivar et al., 2011; Wang et al., 2011; Alonso-Hernando et al., 2012) and to a lesser extent from clinical samples (Poyart-Salmeron et al., 1990; Tsakris et al., 1997; Godreuil et al., 2003; Safdar and Armstrong, 2003; Morvan et al., 2010). The mechanisms responsible for the acquisition of antibiotic resistance in L. monocytogenes are still not fully explained. It was suggested that enterococci and streptococci might transmit mobile plasmids encoding antibiotic resistance determinants to Listeria spp. (Poyart-Salmeron et al., 1990; Charpentier and Courvalin, 1999). In addition, exposure to biocides (e.g., disinfectants, food and feed preservatives, or decontaminants) may promote strains with increasing resistance to clinically relevant antibiotics (Christensen et al., 2011; Capita and Alonso-Calleja, 2013; Kovacevic et al., 2013).
There is a dearth of comprehensive studies focused on the evaluation of antimicrobial resistance of L. monocytogenes, with relatively few publications reporting results for a large number of clinical isolates (e.g., Larsson et al., 1985; MacGowan et al., 1990; Safdar and Armstrong, 2003; Hansen et al., 2005; Morvan et al., 2010). Moreover, few studies have evaluated and compared antimicrobial susceptibility of food and clinical isolates recovered from the same geographic regions in the same period of time (Charpentier et al., 1995; Okada et al., 2011). In Portugal, the only information available concerning antibiotic resistance of L. monocytogenes was published by Antunes et al. (2002) and only refers to isolates from retail poultry carcasses commercialized in Porto. Considering the association between consumption of contaminated foods and listeriosis, data describing antimicrobial resistance in food and clinical isolates are of utmost importance. The spread of antibiotic resistance could be of considerable significance for successful clinical treatment of listeriosis, and may also be of epidemiological and ecological significance, as has occurred for other pathogens.
The aim of this study, therefore, was to characterize a broad collection of isolates of L. monocytogenes, of different serotypes, recovered in Portugal between 2003 and 2007 from various food categories and from clinical cases of human listeriosis in terms of antimicrobial susceptibility.
Materials and Methods
Origin of isolates
Four hundred and forty-eight isolates of L. monocytogenes—353 isolated from various food products (Table 1) and 95 from human cases of listeriosis (Table 2), isolated between 2003 and 2007—were confirmed as recommended by the ISO standards (Anonymous, 1996). These isolates were deposited and stored at −80°C in tryptone soya broth supplemented with 0.6% (wt/vol) of yeast extract containing 30% (vol/vol) glycerol in the Listeria culture collection of CBQF-Escola Superior de Biotecnologia (Porto, Portugal), and used in the current investigation.
Molecular serotypes were determined as reported by Doumith et al. (2004). Group IVb includes serotypes 4b or 4d or 4e; polymerase chain reaction (PCR) group IIb includes serotypes 1/2b or 3b; PCR group IIa includes serotypes 1/2a or 3a; PCR group IIc includes serotypes 1/2c and 3c.
All of the isolates were sensitive to ampicillin, penicillin, chloramphenicol, vancomycin, and trimethoprim/sulfamethoxazole.
Traditional fermented meat sausage produced in the North of Portugal.
Collected in canteens.
Molecular serotypes were determined as reported by Doumith et al. (2004). Group IVb includes serotypes 4b or 4d or 4e; polymerase chain reaction (PCR) group IIb includes serotypes 1/2b or 3b; PCR group IIa includes serotypes 1/2a or 3a; PCR group IIc includes serotypes 1/2c and 3c.
All of the isolates were sensitive to ampicillin, penicillin, chloramphenicol, vancomycin, tetracycline, trimethoprim/sulfamethoxazole, erythromycin, rifampicin, and gentamicin.
Growth and storage conditions
Isolates were grown on tryptone soya agar (Pronadisa, Madrid, Spain) supplemented with 0.6% (wt/vol) of yeast extract (TSAYE; Lab M, Lancashire, UK) at 37°C for 24 h and subcultured twice before use in assays.
Genoserotyping or polymerase chain reaction (PCR) grouping
Genoserotyping was determined by PCR grouping with a multiplex PCR as described by Doumith et al. (2004) using primers targeting fragments of genes lmo0737, ORF2819, ORF 2110, lmo1118, and prs (MWG-Biotech, Muenchenstein, Switzerland).
Determination of minimum inhibitory concentration (MIC)
Culture preparation
All L. monocytogenes isolates as well as control strains (Enterococcus faecalis ATCC 29212 and Escherichia coli ATCC 25922) were maintained on TSAYE. The inocula were prepared by suspending some colonies (grown for 24 h at 37°C on TSAYE) in sterile Ringer's solution (Lab M) in order to obtain turbidity equivalent to 0.5 McFarland standards and used for the tests, according to the Clinical and Laboratory Standard Institute (CLSI, 2007).
Antibiotics used and plates preparation
Representatives of the main classes of antibiotics used in both human and animal medicine were selected for this study. Each test was carried out on Mueller-Hinton agar (MHA; BioMérieux, Marcy l'Etoile, France) with cation adjusted for penicillin G (Sigma, Steinheim, Germany) and ampicillin (Fluka, Steinheim, Germany) and on MHA for the other nine antibiotics: vancomycin and chloramphenicol (both from Fluka), nitrofurantoin and trimethoprim/sulfamethoxazole (both from Sigma), erythromycin, tetracycline, ciprofloxacin, gentamicin, and rifampicin (kindly supplied by the company Labesfal, Tondela, Portugal). A volume of 2 mL of each dilution of antibiotic solutions, prepared from serial twofold dilutions according to CLSI (2007), and 18 mL of molten MHA cooled to 48°C, with 3% (vol/vol) of lysed horse blood (Oxoid, Basingstoke, UK) were placed into Petri dishes. Antibiotic concentrations ranged from 0.015 to 512 μg/mL.
Agar microdilution method
According to CLSI, Listeria spp. should be tested with the broth microdilution. However, agar dilution testing, generally not performed in routine clinical laboratories, is ideal for research dealing with large numbers of isolates (Rankin, 2005). Therefore, in the present study, the MIC (μg/mL) for all isolates was determined by the agar microdilution method (CLSI, 2007). A volume of 1 μL of the inoculum was spotted using a micropipette on each plate with antibiotic and incubated at 37°C for 24 h. The absence of growth in the minimal concentration of an antibiotic was interpreted as the MIC. All the isolates were grown on plates of MHA and MHA with cation adjusted (both supplemented with 3% [vol/vol] of lysed horse blood) with no antibiotic as controls. For all antimicrobials, E. coli ATCC 25922 and E. faecalis ATCC 29212 were used as quality control bacteria for MIC as recommended by CLSI (2007). Each experiment was performed in duplicate.
Classification of isolates as susceptible or resistant to antibiotics used
Apart from penicillin and ampicillin, for which specific breakpoints for Listeria susceptibility testing are defined by the CLSI, in the present study, breakpoints used for the agar dilution method were those recommended by the CLSI, formerly National Committee for Clinical Laboratory Standards (NCCLS, 2002) for veterinary pathogens or CLSI criteria for staphylococci as previously reported by other authors (Conter et al., 2009).
Statistical analysis
An analysis of variance was carried out to test the effect of time (2003–2007) on the MICs of the antibiotics investigated. All calculations were carried out using the software Kaleidagraph (version 4.04, Synergy Software, Reading, PA).
Data were analyzed according to contingency tables (cross-tabulation) to assess dependency between serotypes of the isolates and resistance/susceptibility. The analyses were performed for each of the antibiotics. All analyses were performed using IBM SPSS Statistics, 20 (IBM Corporation, USA).
Results and Discussion
One hundred (28.3%) and 20 (21.0%) food and clinical isolates, respectively, were resistant to at least one antibiotic. All the clinical and food isolates were susceptible to ampicillin (generally considered the preferred agent to treat human listeriosis), penicillin, chloramphenicol, vancomycin, and trimethoprim/sulfamethoxazole (Tables 1 and 2). Although resistance to these antibiotics has been previously demonstrated (Safdar and Armstrong, 2003; Conter et al., 2009; Ayaz and Erol, 2010; O'Connor et al., 2010; Pesavento et al., 2010), the current results are in agreement with most previously published reports when testing L. monocytogenes strains isolated from foods (Okada et al., 2011; Ruiz-Bolivar et al., 2011; Sakaridis et al., 2011; Alonso-Hernando et al., 2012; Korsak et al., 2012) and from human patients (Hansen et al., 2005; Okada et al., 2011). Although all isolates were classified as susceptible to ampicillin, a significant increase in the average MIC values (p<0.05) was observed during the years of testing (0.25±0.0, 2003; 0.27±0.0, 2004; 0.27±0.0, 2005; 0.33±0.1, 2006; 0.45±0.1, 2007). This tendency was not observed for any of the other antibiotics investigated (p>0.05), and may represent a response to treatment of cases of listeriosis with this antibiotic. It would be important to assess any continuing increase in resistance in the next few years. The highest percentages of resistant isolates were those with resistance to nitrofurantoin (23.8% from food and 20.0% from clinical cases), ciprofloxacin (4.5% from food and 1.0% from clinical cases) and erythromycin (2.8% from food) (Tables 1 and 2). It is also important to note that the percentage of clinical isolates resistant to nitrofurantoin significantly (p<0.05) increased from 7% in 2003 to 50% in 2007 (data not shown). A high percentage of resistance to nitrofurantoin had already been reported for food (Wang et al., 2011; Alonso-Hernando et al., 2012) and clinical isolates (Safdar and Armstrong, 2003). In the European Union, the use of nitrofuran antibiotics for the treatment of bacterial diseases in animal production was banned in 1995 (EC, 1995). Since January 2006, the use of antimicrobials as growth-promoting agents has been banned within the European Union in order to reduce the numbers of resistant bacteria in farm animals (EC, 2003). Cross-resistance or co-resistance mechanisms could be associated with the high resistance of poultry isolates to furazolidone, as suggested by Capita and Alonso-Calleja (2013). It will be important to assess any further increase in resistance during the next few years.
Although L. monocytogenes is considered naturally susceptible to fluoroquinolones (Troxler et al., 2000) a significant increase in its resistance of to ciprofloxacin has been reported recently (O'Connor et al., 2010; Alonso-Hernando et al., 2012). This increase was attributed by Alonso-Hernando et al. (2012) to the widespread use of fluoroquinolones in poultry production. Fluoroquinolones are often used to treat respiratory infections in flocks. Antunes et al. (2002) had previously reported that 58% of L. monocytogenes isolates recovered from poultry carcasses were resistant to enrofloxacin and also attributed this high frequency to extensive use of fluoroquinolones in animal feeds.
The resistance of 10 (2.8%) isolates to erythromycin (Table 1) is a reason for concern, since erythromycin may be used as a second-choice drug in cases of β-lactam allergy (Temple and Nahata, 2000). Resistance to erythromycin has been previously reported by, among others, Miranda et al. (2008), Osaili et al. (2011), Pesavento et al. (2010), and Ruiz-Bolivar et al. (2011).
Eight isolates (1.8%) were resistant to two or more antimicrobials of different classes, and all were collected from foods (Table 3). It is important to note that, with the exception of nitrofurantoin and tetracycline, these antibiotics are recommended as second-line options for the treatment of some cases of listeriosis (Temple and Nahata, 2000). As previously demonstrated (Morvan et al., 2010), multidrug-resistant isolates appear to be rare among clinical isolates.
Traditional fermented meat sausage produced in the North of Portugal.
S, susceptible.
Listeria monocytogenes has been differentiated into 13 serotypes; serotypes1/2a, 1/2b, and 4b have been involved in the majority of reported human listeriosis cases (Farber and Peterkin, 1991; Almeida et al., 2010). In this work, isolates belonging to serogroup IVb (includes serotypes 4b or 4d or 4e) were the most prevalent among both food (34.5%) and clinical (71.5%) isolates (Tables 1 and 2). Resistance to erythromycin and nitrofurantoin was correlated (p<0.05) with the serogroup of the isolates. Serogroup IVb included the highest percentage of isolates resistant to erythromycin. The highest percentages of isolates resistant to nitrofurantoin were of serogroup IVb and IIc (includes serotypes 1/2c and 3c). This correlation was not observed for ciprofloxacin (p>0.05). Kovacevic et al. (2013), however, found that serotype 1/2a isolates were more frequently resistant to ciprofloxacin compared to serotype 4b. Ayaz and Erol (2010) found that isolates of serotype 1/2c were more resistant to penicillin and ampicillin than those of serotypes 1/2a, 1/2b, and 4b. To our knowledge, correlations between serotypes and antimicrobial resistance have only been reported in the present work and in the two works mentioned above, and no explanation for these observations has been proposed.
Conclusions
The results presented in this study demonstrated that the incidence of antibiotic-resistant isolates of L. monocytogenes, during the period 2003–2007, was low in Portugal but still higher than that observed in other countries including Japan (Okada et al., 2011) France (Granier et al., 2011), and Poland (Korsak et al., 2012). Given the increased population at greater risk of listeriosis (i.e., the elderly), the high mortality rate from the infection, and the continuing detection of resistant isolates, monitoring for antibiotic resistance of L. monocytogenes on a large scale and assessing the risk of infection by these strains is highly recommended (Charpentier and Courvalin, 1999). According to Lungu et al. (2011), resistance in clinical human isolates may emerge in the near future.
Footnotes
Acknowledgments
Editing of the manuscript by Dr. P. A. Gibbs is gratefully acknowledged. This work was supported by National Funds from FCT—Fundação para a Ciência e a Tecnologia—through projects PTDC/AGR-ALI/64662/2006 and PEst-OE/EQB/LA0016/2011.
Disclosure Statement
No competing financial interests exist.
