Characterization of Colonizing Staphylococcus aureus Isolated from Surgical Wards' Patients in a Nigerian University Hospital

In contrast to developed countries, only limited data on the prevalence, resistance and clonal structure of Staphylococcus aureus are available for African countries. Since S. aureus carriage is a risk factor for postoperative wound infection, patients who had been hospitalized in surgical wards in a Nigerian University Teaching Hospital were screened for S. aureus carriage. All S. aureus isolates were genotyped (spa, agr) and assigned to multilocus sequence types (MLST). Species affiliation, methicillin-resistance, and the possession of pyrogenic toxin superantigens (PTSAg), exfoliative toxins (ETs) and Panton-Valentine Leukocidin (PVL) were analyzed. Of 192 patients screened, the S. aureus carrier rate was 31.8 % (n = 61). Of these isolates, 7 (11.5%) were methicillin-resistant (MRSA). The isolates comprised 24 spa types. The most frequent spa types were t064, t084, t311, and t1931, while the most prevalent MLST clonal complexes were CC5 and CC15. The most frequent PTSAg genes detected were seg/sei (41.0%) followed by seb (29.5%), sea (19.7%), seh (14.7%) and sec (11.5). The difference between the possession of classical and newly described PTSAg genes was not significant (63.9% versus 59.0% respectively; P = 0.602). PVL encoding genes were found in 39.3% isolates. All MRSA isolates were PVL negative, SCCmec types I and VI in MLST CC 5 and CC 30, respectively. Typing of the accessory gene regulator (agr) showed the following distribution: agr group 1 (n = 20), group II (n = 17), group III (n = 14) and group IV (n = 10). Compared to European data, enterotoxin gene seb and PVL-encoding genes were more prevalent in Nigerian methicillin-susceptible S. aureus isolates, which may therefore act as potential reservoir for PVL and PTSAg genes.


Introduction
Staphylococcus aureus is a widely recognized human pathogen that continues to represent a significant public health challenge globally. Owing to its broad spectrum of inherent virulence factors, it has been shown to cause a variety of infectious diseases such as superficial skin infections, endocarditis, septicemia, and toxin-mediated diseases, which are often difficult to treat [1,2]. In addition to its known virulence properties, the ability of methicillin-resistant S. aureus strains (MRSA) to consistently evolve resistance mechanisms has led to the emergence of hospitaladapted multi-resistant clones, which became a major cause of nosocomial infections world-wide [3,4,5]. The occurrence of MRSA strains in the community (community-associated MRSA) and in livestock (livestock-associated MRSA), as well as the potential risk of import of such strains into the hospital, are also matters of concern [6,7]. While some data are available on the occurrence of S. aureus implicated in various infections [1,4,5,6,8,9,10,13,14,15] there are only few reports describing S. aureus colonization pattern in patients from sub-Saharan Africa [11,12,13]. In particular, for surgical patients who are predisposed to develop a variety of post-operative events, including postsurgical toxic shock syndrome, data concerning the toxigenic equipment, MRSA prevalence, and clonal structure of S. aureus are missing in Africa. It is well known that antibiotic-resistance, toxigenic equipment carriage and clonal analyses in S. aureus can provide useful insights into the virulence potential and nature of S. aureus populations [1,2,16,17,18,19]. Among the several exotoxins secreted by S. aureus, Panton-Valentine Leukocidin (PVL), a bicomponent cytotoxin encoded by lukS-PV and lukF-PV genes of the PVL-encoding operon, has been observed in less than 5% of S. aureus isolates in Europe and can be associated with necrotic skin lesions including severe necrotizing pneumonia [20,21]. Moreover, the expression of most S. aureus virulence factors is known to be controlled by a complex regulatory network including the accessory gene regulator (agr) system which consists of an approximate 3kb genetic locus containing divergent transcription units [20,21,22,23]. The agr locus is characterized by a polymorphism of its auto-inducing peptide (AIP), and has been shown to reliably divide S. aureus into four major agr groups. These groups evolved early in staphylococcal evolution [20,21,22]. Additionally, molecular typing technologies such as staphylococcal protein A (spa) typing and multilocus sequence typing (MLST) provide informative typing results which enable the grouping of individual isolates in clonal lineages [20,21,24]. Therefore, the aims of the present study were (i) to prospectively collect S. aureus isolates from a hospitalized surgical patient cohort in a Nigerian University Teaching Hospital (ii) to investigate their toxigenic and resistance properties and (iii) to analyze their clonal composition.

Population
The Obafemi Awolowo University Teaching Hospitals Complex in Ile-Ife has approximately 500 available beds across seven major hospital wards (internal medicine, surgery, orthopedics, psychiatry, gynecology, obstetrics and pediatrics). The hospital serves the city of Ile-Ife in Osun State and is a major referral hospital in Southwestern Nigeria. Exclusion criteria were (i) referral from other hospitals, (ii) superficial skin and soft tissue infections, and (iii) hospitalization for more than 48 hours at the time of sampling. The most common causes of hospitalization on the basis of clinical records were internal fractures (38.8%), cancers (34.6%), peritonitis (8.3%), osteomyelitis (7.2%), and urinary tract infections (5.5%). Demographic and clinical information was collected by interviews, questionnaires and review of clinical records. Institutional Review Board approval was obtained from Obafemi Awolowo University Teaching Hospitals Complex and all patients gave written informed consent to participate in the study.

Bacterial strains
Between November 2008 and July 2010 all eligible patients hospitalized in the surgical wards (general surgery, pediatric surgery, orthopedics and gynecology) were prospectively screened for S. aureus within 48 hours of admittance. Swabs were collected from both anterior nares and from the skin (back of the wrist) using sterile cotton-tipped swab sticks (MicroPoint Diagnostics). A total of 48 strains were isolated from the nares, and 13 strains were recovered from cutaneous specimens. Identification of S. aureus and antimicrobial susceptibility testing was performed using VITEK-2 automated systems (BioMérieux, Marcy l'Etoile, France).
spa and agr typing To determine the spa type of S. aureus strains, the polymorphic X-region of the S. aureus protein A gene (spa) was sequence-typed [18]. Cluster formation of spa types was performed by the ''based upon repeat patterns'' (BURP) algorithm of the StaphType software (Ridom, Münster, Germany) using the default parameters described by Mellmann et al. [19]. Additionally, subtypes of the accessory gene regulator were detected by multiplex PCR as previously described [20,30]. Multilocus sequence typing (MLST) was performed for spa types detected for the first time in this study [24].

Statistical analysis
Categorical data were compared using the chi-square test. In addition, we report proportions of categorical variables (i.e. virulence factors and antimicrobial resistance). All computations were performed using Epi Info TM 3.5.3 (Centers for Disease Control and Prevention, Atlanta, USA). P values .0.05 were considered not statistically significant.

Patient characteristics and S. aureus carriage
Among 192 persons tested, the overall prevalence of S. aureus carriage was 31.8% (n = 61) with 27.1% (n = 28) among males and 37.1 % (n = 33) among females (P = 0.105) ( Table 1). The mean age was 39.7 years among carriers of methicillin-susceptible S. aureus (MSSA), 31.4 years among MRSA carriers and 44.2 years among S. aureus non-carriers (P = 0.30). All S. aureus were tested nuc positive. Seven of the 192 patients (3.7%) were colonized by MRSA and 11.5% of all S. aureus were MRSA. The seven MRSA were recovered within the period of sampling from seven patients variously hospitalized in the orthopedic (n = 2), general surgery (n = 4) and pediatric surgery (n = 1) wards. All MRSA isolates were tested positive for mecA and negative for mecC. Of all 61 S. aureus isolates, 73.7% were resistant to penicillin, 42.6% to tetracycline, 3.3% to clindamycin, 16.4% to gentamicin, 9.8% to levofloxacin

Exotoxin gene detection
Overall, 70.5% (n = 43) of all isolates tested were PTSAg genepositive. We did not observe significant differences between the prevalence of classical PTSAg genes and the subsequently described enterotoxin and enterotoxin-like genes tested (63.9% vs 59.0% repspectively; P = 0.602). Regarding the classical enterotoxin genes, seb gene was the most prevalent followed by sea gene and sec gene. The sed gene was detected in two isolates, while see gene was not observed. Altogether, three isolates harbored the tst gene. The seg-sei genes occurred most frequently and strictly in combination with one another while the seh gene was observed in nine isolates ( Table 2). Several enterotoxin gene combinations were observed including isolates with a combination of two (n = 14, 22.9%), three (n = 20, 32.7%), four (n = 2, 3.6%) and five (n = 1, 1.6%) different genes. The most frequent combination of genes detected was seg -seiseb genes occurring in 19.6% (n = 12) of all isolates. In contrast, seven isolates (11.4%) harbored a single PTSAg gene. Overall, three (4.9%) of the isolates were positive for the exfoliative toxin genes; etb gene was not detected. While 99% of isolates tested PCR positive to chemolysin encoding genes, the lukF-PV and lukS-PV genes of the PVL operon were detected in 24 isolates (39%), all of which were MSSA strains. Moreover, 75.9% (n = 41) of the MSSA population were PTSAg gene positive (Table 3): 13% were positive for the classical PTSAg genes (n = 7); 13% exclusively encoded the newly described PTSAg genes (n = 7) while 50% of the MSSA population encoded both classical and newly described PTSAg genes. The seb gene was the most frequently observed classical PTSAg gene in the MSSA group (n = 17, 31.4%) while seg-sei genes occupied this position (n = 24, 44.4%) among the subsequently described PTSAg genes. The percentage of exfoliative toxin gene-positive MSSA strains was 5.5% (n = 3).
Regarding the MRSA subset, two isolates were PTSAg gene positive (28.5%). One of the MRSA isolates was associated with five PTSAg genes, namely: seb, sed-sej and seg-sei genes. Meanwhile the second MRSA isolate harbored the sea gene alone. Exfoliative toxin and PVL encoding genes were not observed in any of the MRSA isolates tested.

SCCmec typing
Two distinct SCCmec types (I and VI) were observed among the seven MRSA isolates. However, five of the seven MRSA isolates were not typeable (NT, Table 3).

Discussion
We analyzed the clonal structure, toxin gene equipment and resistance pattern of 61S. aureus carriage isolates obtained from a surgical patient cohort (n = 192) in Nigeria. While 3.9% of the patients in the cohort studied were colonized by MRSA, Heysell et al. [11] reported on a high prevalence of multidrug-resistant MRSA nasal carriage (29%) in a district hospital in South Africa. Interestingly, in epidemiological studies involving persons on the Cape Verde islands as well as Central African Babongo Pygmies, no MRSA isolate was found [12,32]. However, in studies characterizing infection-related S. aureus isolates, higher and increasing MRSA rates (11.1-47.4%) have been reported in the North and sub-Saharan African region [9,13,33]. Reports on the prevalence of PTSAg genes and the detection of respective toxin genes differ greatly depending on the geographic affiliation, the population structure tested, and the range of staphylococcal PTSAg genes included. In particular, the recent detection of enterotoxins and enterotoxin gene-like toxin genes beyond the classical spectrum (such as enterotoxins SEG -SEJ, SEK -SER and SEU) has increased the percentage of those S. aureus isolates bearing at least one of the PTSAg encoding genes [28,34]. Moreover, like SEC (staphylococcal enterotoxin C), several of the newly described enterotoxins are polymorphic, being characterized by the occurrence of nucleotide variants, additionally challenging their precise detection [29]. In this study, 11.4% of the S. aureus isolates were shown to harbor at least one of the PTSAg gene family members. This is in accordance with other studies of different geographic origin [23,28,35]. However, these studies differ from our investigation because they included S. aureus isolates obtained from the clinical care setting. It is well documented that the smallest amounts of staphylococcal enterotoxins may induce T-cell stimulation resulting in systemic illness such as staphylococcal enterotoxin-induced shock and autoimmunity [28,36]. Given the detection of a significant amount of toxin genes including tst gene in the present study, post-operative patients hospitalized in the surgical wards may be at risk of postsurgical toxic shock syndrome [16,17]. For SEH, emetic activity has been shown and this PTSAg is considered a potential causative agent for food poisoning [37]. Importantly, a high percentage of seh gene-positive isolates (14.7%) was found in our study relative to a low prevalence reported for this enterotoxin gene in other studies investigating non-food borne-related S. aureus isolates: 5.4% [28], 4% [35], and 4.3% [38]. Peck et al. [39] reported significant differences in the prevalence of selected enterotoxin genes, including seh gene in S. aureus isolates obtained from blood compared to nasal isolates (7.2% vs. 30.5%). However for food borne isolates, higher percentages of seh gene-positive isolates have been observed. Aydin et al. [40] reported on 16.3% seh genepositive S. aureus isolates obtained from food samples, particularly from meat.
We have found a high prevalence (44%) of lukS-PV and lukF-PV genes of the PVL operon in MSSA carriage isolates. For S. aureus isolates obtained from wound infection, similar results have been reported [15,37]. Similarly, Schaumburg et al. [13] reported on Gabonese S. aureus isolates from both, asymptomatic carriers and infections with high percentages of PVL-positive MSSA isolates. Moreover, for a remote Central African population with limited access to health care facilities, a high prevalence of PVL-positive MSSA was also detected [12]. Since PVL-encoding genes are carried on prophages as mobile genetic elements, which enable incorporation into S. aureus lineages via horizontal transfer, the Central and West African region may act as a potential major reservoir for the PVL virulence factor with considerable impact on regional as well as global health care systems. Molecular typing of the MRSA carriage isolates characterized in this study revealed that they belonged to spa types indicative for the clonal complexes spa-CC5 (t002, t037, t194, t1842), spa-CC30 (t007) and spa-CC88 (t729), all of which have been previously described in African countries [12,15,41,42].
In contrast, among MSSA isolates, spa types indicative for MLST CC15 (ST1/ST215; mostly spa types t084, t1931) and MLST CC5 (ST1/ST8; mostly spa types t064, t311) were predominant. While similar findings have been reported from other studies on the molecular structure of African MSSA [10,43], we did not observe S. aureus strains of MLST sequence types ST2250, and ST2277 in the present study. Of note, four hitherto undescribed spa types: t7834, t7835, t8034 and t8037, associated with ST2355, ST1, ST72 and ST152 respectively, were detected in this study. Furthermore, a high prevalence of PVL-positive MSSA isolates (16.4%) of spa types t159 and t2304 (clonal lineage MLST ST121/CC121), which belonged to agr group IV, was observed in our study. These findings are consistent with previous reports from Africa and other regions of the world including Europe and arctic Russia [10,15,44].
In summary, we documented a 3.9% prevalence of MRSA colonization among 192 patients screened. Analysis of the toxin gene content of the Nigerian S. aureus isolates revealed a relatively high overall prevalence of PTSAg genes. In particular, the seh gene encoding for an enterotoxin which can induce emetic disease was found in association with ST1. While no PVL-positive MRSA isolates were detected, PVL-encoding genes were highly frequent among MSSA isolates of different clonal lineages. Since we characterized carriage isolates of S. aureus obtained from surgery wards' patients in Nigeria, this study represents an important reference point for understanding the virulence potential and clonal epidemiology of the strains studied. While the role of the accessory gene regulator in S. aureus colonization is yet unclear, our study nonetheless provides further insight in to the clonal nature of the strains studied given the well-demonstrated relationship between accessory gene regulator groups and S. aureus clonal evolution [20,22].