Implementation of a novel in vitro model of infection of reconstituted human epithelium for expression of virulence genes in methicillin-resistant Staphylococcus aureus strains isolated from catheter-related infections in Mexico

Background Methicillin-resistant Staphylococcus aureus (MRSA) are clinically relevant pathogens that cause severe catheter-related nosocomial infections driven by several virulence factors. Methods We implemented a novel model of infection in vitro of reconstituted human epithelium (RHE) to analyze the expression patterns of virulence genes in 21 MRSA strains isolated from catheter-related infections in Mexican patients undergoing haemodialysis. We also determined the phenotypic and genotypic co-occurrence of antibiotic- and disinfectant-resistance traits in the S. aureus strains, which were also analysed by pulsed-field-gel electrophoresis (PFGE). Results In this study, MRSA strains isolated from haemodialysis catheter-related infections expressed virulence markers that mediate adhesion to, and invasion of, RHE. The most frequent pattern of expression (present in 47.6% of the strains) was as follows: fnbA, fnbB, spa, clfA, clfB, cna, bbp, ebps, eap, sdrC, sdrD, sdrE, efb, icaA, and agr. Seventy-one percent of the strains harboured the antibiotic- and disinfectant-resistance genes ermA, ermB, tet(M), tet(K), blaZ, qacA, qacB, and qacC. PFGE of the isolated MRSA revealed three identical strains and two pairs of identical strains. The strains with identical PFGE patterns showed the same phenotypes and genotypes, including the same spa type (t895), suggesting hospital personnel manipulating the haemodialysis equipment could be the source of catheter contamination. Conclusion These findings help define the prevalence of MRSA virulence factors in catheter-related infections. Some of the products of the expressed genes that we detected in this work may serve as potential antigens for inclusion in a vaccine for the prevention of MRSA-catheter-related infections.


Introduction
Staphylococcus aureus is a bacterial pathogen that causes multiple infections in humans, ranging from superficial skin infections to endocarditis, bone and joint infections, septic shock [1], and severe haemodialysis catheter-related infections [2]. S. aureus produce a broad spectrum of extracellular and cell wall-associated virulence determinants that contribute to the severity of infection [3]. Microbial adherence to cells and extracellular matrix is an essential first step in the process of colonization and infection [4], for which S. aureus express numerous surface adhesins known as microbial surface components recognizing adhesive matrix molecules (MSCRAMMs). These adhesins mediate adherence to host extracellular matrix components such as fibrinogen, fibronectin, and collagen [5]. Pathogenic adhesins include Fibronectin-binding protein A and B (FnBPA and FnBPB); Staphylococcus protein A (Spa); clumping factor A (ClfA); clumping factor B (ClfB); collagen adhesion (Cna); sialoprotein-binding protein (Bbp); elastinbinding protein of Staphylococcus aureus (EbpS); extracellular adhesion protein (Eap); serine aspartate repeat proteins C, D, and E (SdrC, SdrD, and SdrE; [3]); and extracellular fibrinogen-binding protein (Efb; [6]). Furthermore, pathogenic S. aureus strains exhibit a great capacity for biofilm formation on surfaces, making endovascular catheters a favourable niche for infection. Biofilm formation requires synthesis of PNAG (polymeric N-acetylglucosamine); the enzymes responsible for its synthesis are encoded by the icaADBC operon [7]. Expression of most virulence factors in S. aureus is controlled by the agr locus [8].
The number of methicillin-resistant S. aureus (MRSA) strains has steadily increased and nosocomial infections caused by MRSA have become a serious problem worldwide, as its presence has dramatically reduced the number of effective antibiotics available for the prevention and treatment of infections in hospitals and communities [9]. The main mechanism of methicillin resistance involves expression of the mecA gene, which encodes penicillinbinding protein 2a (PBP2a), a transpeptidase with low affinity for β-lactams [10]. MRSA strains frequently carry genes for multidrug resistance such as blaZ, which codes β−lactamases and confers resistance to β-lactams; ermA, ermB, and ermC confer erythromycin resistance; aac(69)-Ie-aph(20)-Ia confers aminoglycoside resistance; tet(M), tet(O), and tet(K) confer resistance to tetracycline; and vanA and vanB confer vancomycin resistance [11]. Widespread use of quaternary ammonium compounds (QAC) in hospitals contributes to the selection of disinfectantresistant S. aureus [12]. In several staphylococcal species, the qacA, qacB, and qacC genes have been identified in plasmids that also contain antibiotic-resistance genes [13][14][15].
The expression of virulence factors of S. aureus has been studied in vivo in animal models of infection [16,17], but immune cellular factors and nutritional conditions have affected the expression of virulence determinants. To circumvent these problems, in this study we propose a novel model of infection in vitro of reconstituted human epithelium (RHE) to analyze the expression patterns of virulence genotypes of MRSA strains isolated from catheter-related infections in Mexican patients undergoing haemodialysis. We also determined the phenotypic and genotypic combinations of antibiotic-and disinfectantresistance in the S. aureus strains, which also were analysed by pulsed-field-gel electrophoresis (PFGE).

Bacterial strains
Twenty-one S. aureus strains were donated by the Laboratorio Clínico de la Facultad de Estudios Superiores Iztacala (Universidad Nacional Autónoma de México) for phenotypic and molecular analysis. These strains were isolated from catheter-related infections of ambulatory patients (n = 21) undergoing treatment at a public hospital in Estado de México, México from January to May 2013. The bacterial strains were identified by classical microbiologic methods: Gram staining; mannitol, catalase, and coagulase activity (Bactident-coagulase, Merck); and the Api 32 Staph test (BioMerieux). S. aureus strains were molecularly characterized by PCR amplification of 23S rRNA, coagulase (coa), thermonuclease (nuc), clumping factor (clfA), protein A region X (spa), femA, and femB [18,19]. Bacterial DNA was extracted with the Wizard Genomic DNA Purification Kit (Promega). Resistance to methicillin was determined by the cefoxitin disc diffusion test (Becton Dickinson; inhibition zone, ≤21 mm) [20]. β-Lactamase enzymes were detected by using paper discs impregnated with the chromogenic cephalosporin nitrocephin (Becton Dickinson, USA). This substrate changes from yellow to red after the amide bond of the β-lactam ring is hydrolysed by β-lactamase. The change in colour was observed from 5 min-1 h. The mecA gene, which encodes methicillin resistance, was identified by conventional PCR [21]. S. aureus ATCC 25923 (mecA -) and ATCC 33592 (mecA+) were used as controls in each test.

SCCmec types and agr groups
A multiplex PCR with four primer pairs was used to identify the five main known SCCmec types [22] and another multiplex PCR with five primers was used to identify the agr groups [23].

RHE inoculation with S. aureus
Reconstituted human epithelium (RHE; SkinEthic Laboratories, Nice, France) consists of human epithelial cell lines cultured on polycarbonate filters in vitro at the air-liquid interface in serum-free chemically defined medium. A total of 2 × 10 6 S. aureus cells suspended in 50 μL PBS were inoculated onto the surface of the RHE and incubated at 37°C for 72 h with 5% CO 2 and saturated humidity. The maintenance media was changed every 24 h.
S. aureus RNA purification and reverse transcription S. aureus cells were harvested from the RHE and suspended in 1 mL RNA Protect Bacteria Reagent (Qiagen). The sample was vortexed 30 s and incubated at room temperature 5 min. After centrifugation at 9400 rpm for 10 min, the cells were resuspended in 200 μL TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) containing 10 mg/mL lysozyme and 40 mg/mL lysostaphin. The sample was vortexed 10 s and incubated at room temperature 5 min. Total RNA was purified with the RN easy Mini Kit (Qiagen) according to the manufacturer's instructions, including DNase treatment. The concentration and purity of total RNA were analysed with a NanoDrop 2000 spectrophotometer (Thermo Scientific). cDNA synthesis was performed with the QuantiTec Reverse transcription kit (Qiagen) according to the manufacturer's instructions.

PFGE typing
MRSA isolates were PFGE typed by preparation of DNA and resolution of the SmaI-digested fragments as described by McDougal et al. [27]. Samples were separated on a CHEF-DR II system (Bio-Rad). Gels were photographed and digitized using a Bio-Rad Gel Doc. PFGE patterns were analysed as described by Tenover et al. [28] for bacterial strain typing.

Data analysis
PFGE patterns were analysed with Gene Tool and Gene Directory software (Syngene). Reference standard S. aureus NCTC 8325 was included in each gel for band normalization. Percent similarities were obtained from the weighted pair group with mathematical average (UPMGA) based on Dice coefficients. Band position tolerance was set at 1.25%. A similarity coefficient of 80% was selected to define the pulsed-field type clusters.

spa typing
The polymorphic X region of the protein A coding gene (spa) was amplified and sequenced as described [29]. Corresponding spa-types were assigned using the SPA Searcher website (http://seqtools.com). Ridom spa types were subsequently assigned using the spa-typing website (http://www.spaserver.ridom.de/) developed by Ridom Gmb.

Results
All S. aureus strains (n = 21) were resistant to methicillin ( Without considering the 7 genes expressed by all MRSA strains, ten distinct expression patterns of virulence markers were found during MRSA infection of RHE in vitro (Table 3). Pattern 1, formed by the other 8 studied genes (fnbA, fnbB, clfA, ebps, eap, sdrD, sdrE, ebf) was present in ten (n = 47.7%) MRSA isolates, whereas patterns 2 and 3, both composed of 7 genes each, were expressed by two strains. The other seven patterns were constituted by 4-7 genes and were expressed only by one MRSA strain (Table 3).
PFGE analysis showed that MRSA strains were distributed in 15 distinct electrophoretic patterns (data not shown). MRSA strains S-22 and S-36 (Table 2), isolated from the catheters of different patients, have identical electrophoretic patterns (data not shown) and the same spa type as the S-59 and S-66 MRSA strains (spa type t895). Three strains isolated from different patients showed 100% similarity by PFGE (S-77, S-79, S-82; Table 2) and belonged to the same spa type t895. Two closely related strains (S-8 and S108) were isolated from the catheters of two different patients and two possibly related strains (S-16 and S-19) were isolated from two different patients. The strains with identical PFGE patterns (S-22 and S-36; S-59 and S-66; S-77, S-79 and S-82) showed the same phenotypes and genotypes ( Table 1). The most frequent spa types were t895 (76.2%; n = 16) and t008 (14.2%; n = 3; Table 2).

Discussion
Haemodialysis patients who are infected with methicillinresistant Staphylococcus aureus (MRSA) are considered to have healthcare-associated (HA) infections [30,31]. For colonization and infection, bacterial adhesion to host extracellular matrix components like fibrinogen, fibronectin, and collagen is essential [5]. Therefore, there has been a strong interest in studying the involvement of proteins of the MSCRAMMs family of S. aureus using in vivo and in vitro models of infection [32][33][34][35]. In this study we implemented a novel model of infection in vitro of reconstituted human epithelium (RHE) to analyze the expression patterns of the MSCRAMMs family adhesion genes, and icaA and agr in S. aureus strains isolated from catheter-related infections in Mexican patients subjected to haemodialysis. Our data show that most of the genes that we studied were expressed by MRSA after infection of RHE (Table 2), which reflects the pathogenic behaviour of these strains. We identified ten different patterns of expression (Table 3), of which pattern No. 1, represented by the 15 genes studied (fnbA, fnbB, spa, clfA, clfB, cna, bbp, ebps, eap, sdrC, sdrD, sdrE, efb, icaA, agr), was present in ten MRSA strains (47.7%; Table 3). These results show that during infection of RHE, the MRSA strains expressed 11 to 13 genes coding for bacterial surface proteins; icaA, which participates in biofilm formation [7]; and the agr locus, which is a global regulator of multiple virulence factors [8]. These findings are consistent with the notion that pathogenesis of most S. aureus infections cannot be explained by the action of an unique virulence factor, but by several distinct factors acting in concert during the infective process [36]. Cna has been associated with endocarditis [37] and keratitis [38]. Fibronectin-binding proteins mediate adherence to human airway epithelium [4]. Clumping factor A (ClfA) plays an antiphagocytic role in neutrophils and macrophages [39] and is necessary for infection/pathogenesis in in vivo models of arthritis, sepsis, and endocarditis [40,41]. Clumping factor B (ClfB) mediates respiratory infection and attachment to cytokeratins on nasal epithelial cells [42] and the role of protein A (Spa) in S. aureus virulence has been demonstrated in models of sepsis and pneumonia [43]. Although the precise role of Sdr adhesins in staphylococcal infection is unknown, a strong correlation between the sdr genes of S. aureus and certain diseases has been reported. There is a significantly increased prevalence of the sdrE gene in invasive S. aureus strains [44], in S. aureus strains responsible for osteomyelitis [45] and in S. aureus isolates responsible for bone infections [46]. A recent report noted that Eap is associated with invasive diseases [33].
Earlier findings suggested a possible relationship between particular agr groups with the capacity of MRSA to cause specific illnesses. The agr I and agr II strains are associated with suppurative infections; agr III is associated with toxic shock syndrome toxin (TSST-1) mediated disease, and agr IV is associated with exfoliative toxin-and impetigo-producing strains [47]. Consistent with these findings, the strains reported in this work were isolated from suppurative catheter-related infections, or initiating suppurative catheter-related infections, and carried agrII (85.7%, n = 18) or agr I (14.3%, n = 3; Table 2).
Most of the strains reported here carried the SCCmec type IV (57%, n = 12; Table 1), an allele initially considered to be a characteristic of community-acquired MRSA (CA-MRSA; [48]). However, MRSA infections in dialysisdependent patients have been considered to be mainly HA (healthcare-associated) according to epidemiologic classifications [49], and there are several reports of MRSA strains harbouring SCCmec type IV in HA infections [30,31,[50][51][52]. The increase in multiple drug-resistant (MDR) MRSA has become a major challenge for the treatment of infectious diseases caused by what are known as superbugs. Strikingly, all the strains reported here were multidrug resistant β-lactamase producers. They were resistant to 5 to 12 antibiotics. None of the strains was sensitive to erythromycin, ampicillin, tetracycline, or penicillin (Table 1). On the other hand, only three strains (14.3%) were resistant to sulfamethoxasole/ Trimethoprim or gentamycin. The high frequency of resistance to these antibiotics may reflect the fact that physicians of the public health service in Mexico prescribe all of these antibiotics, which are considered essential drugs in this sector.
The most frequent combination of antibiotic-and disinfectant-resistance phenotype/genotype in the MRSA strains (71.4%, n = 15) was: resistance to erythromycin, ampicillin, tetracycline, cefotaxime, penicillin, and βlactamase production in association with ermA, ermB, tet(M), tet(K), blaz, qacA, qacB, and qacC genotypes (Table 1). These strains are not only resistant to five antibiotics, but also harbour three of the six plasmidencoded MDR efflux pumps which mediate resistance to several biocides such as antimicrobial organic cations, including intercalating dyes (e.g., acriflavine and ethidium bromide), and quaternary ammonium compounds (e.g., benzalkonium chloride; [53]). Staphylococcal strains resistant to disinfectant have been identified in clinical isolates of MRSA from China (62%; [54]), Taiwan (55.4%; [55]), and Hong Kong (42%; [56]). All the erythromycin-resistant MRSA strains reported here carried the ermA and ermB genes and two strains (S-73 and S-108) carried the ermC gene. The msrA and mef genes were not detected in either strain (Table 1). We did not find discordances between the presence of erm genes and phenotypic resistance to erythromycin, as has been reported in other studies [12,57]. The incidence of ermA genes in our MRSA strains (100%) is higher than those reported in erythromycin-resistant S. aureus: 7.7% in Tunisia [12] and 16% in Denmark [58].
All MRSA strains were resistant to tetracycline and possessed the tet(M) and tet(K) genes (Table 1). Tetracycline is an antibiotic commonly used in Mexico to treat human and animal bacterial infections, contributing to the selection and propagation of resistant strains. Tetracycline resistance in S. aureus is encoded by the pT181 plasmid [59] and the tet (M) gene can be found in combination with ermB in the Tn916 transposon. The simultaneous presence of these genes has been found frequently in enterococci, as well as in streptococci and staphylococci [60]. Consistent with this, all the strains studied here carried these two genes, suggesting the presence of these genetic elements in the MRSA we analysed.
The gene aac(69)-Ie-aph(20)-Ia was detected in three strains (S-19, S-58, and S-108) that were also resistant to gentamycin; the vanA and vanB genes were not detected in any MRSA strain.
The strains with identical PFGE patterns showed the same phenotypes, genotypes, and spa type, suggesting that hospital personnel manipulating the haemodialysis equipment could be the cause of catheter contamination by these strains.
Our results are relevant because they demonstrate that MRSA strains isolated from catheter-related infections in haemodialysis patients express several virulence markers involved in the adhesion and invasion of RHE. We also analysed the phenotypes and genotypes of antibiotics and disinfectant resistance. These results will help define the incidence of virulence factors in MRSA associated with catheter-related infections and improve therapies in haemodialysis patients. In addition, some of the products of the expressed genes that we detected in this work may serve as potential antigens for inclusion in a vaccine for the prevention of MRSA-catheter-related infections.