Molecular epidemiology of Panton-Valentine Leukocidin-positive community-acquired methicillin resistant Staphylococcus aureus isolates in pastoral communities of rural south western Uganda

About a third of healthy individuals in the community are asymptomatically colonized with Staphylococcus aures (S. aureus) in the nostrils [1], a very important finding considering the fact that nasal carriage of S. aureus has been associated with subsequent infection [2], and carriers are an important source of spread of infection in communities. A major concern is the world-wide emergence of methicillin resistant S. aureus (MRSA) in the community [3]. In contrast with health care associated MRSA (HA-MRSA) infections, community associated MRSA (CA-MRSA) infections can occur in healthy individuals [4], suggesting that these strains have greater virulence. Skin and soft-tissue infections represent about 90% of cases of CA-MRSA infection, mostly characterized by abscesses or cellulitis with purulent drainage [3].

A predominant feature of CA-MRSA is the presence of the Panton-Valentine Leukocidin (PVL) genes that encode a S. aureus exotoxin that induces lysis of monocytes and neutrophil granulocytes [5]. Additionally, there is evidence that PVL-positive S. aureus strains susceptible to methicillin (MSSA) may be reservoirs for the development of PVL-positive MRSA via the integration of the staphylococcal cassette chromosome mec (SCCmec) elements including the mecA gene conferring methicillin resistance [6]. Another key feature of CA-MRSA is that the strains mainly harbor SCCmec types IV and V [7‐9], and a relationship between CA-MRSA, SCCmec type IV and V and PVL has been confirmed in some studies [10, 11].

While efforts have been made to map out the geographical predominance and spread of different CA-MRSA clones worldwide [3], little is known about its magnitude and genetic composition in developing countries, especially in Africa. Molecular analytical techniques such as pulsed field gel electrophoresis (PFGE), multilocus sequence typing (MLST), spa and SCCmec typing have been used to show both the spread and evolution of MRSA. PFGE is still considered the gold standard for typing MRSA isolates, and is one of the most discriminative typing methods [12]. While MLST is an excellent tool for investigating clonal evolution of MRSA, it is rather expensive, labour intensive and time consuming. Sequences from the polymorphic region of the S. aureus protein A (spa) gene have been used to develop a single-locus sequence typing technique for MRSA [13] with a discriminatory power between PFGE and MLST, and ability to investigate both molecular evolution and outbreak situations [14], while remaining simple.

The resistance of S. aureus to methicillin is caused by the mecA gene, located on a mobile genetic element, the staphylococcal cassette chromosome mec (SCCmec) [7]. Currently, eleven main SCCmec types (I to XI) are known [15], and Zhang et al. developed a multiplex PCR for the characterization of SCCmec type I to V [16].

In African countries, CA-MRSA has been reported in a few studies in Mali [17], Nigeria [18], Algeria [19], Egypt [20] and Gabon [21]. In Uganda, a few studies have been done at tertiary health care centers [22‐25], and these have focused on HA-MRSA. Community studies have focused on S. aureus carriage in raw milk and its products as well as in urban milk vendors [26, 27]. We set out to establish the prevalence and molecular epidemiology of PVL-positive CA-MRSA isolates from pastoral communities of rural south-western Uganda so as to inform public health on how to develop effective strategies to prevent spread MRSA in these and surrounding communities.

This study isolated 132 Staphylococci isolates from the nares of 253 participants, of which 73 were coagulase positive S. aureus (carriage rate of 29%). The 73 isolates of S. aureus in our study were obtained from 48 male and 25 female participants, with an average age of 13 years, with 56/73 (76. 7%) of the participants being between 7 and 15 years old. This is similar to observations elsewhere that CA-MRSA infections tend to occur in previously healthy younger patients [21, 32, 33]. Our finding is of public health importance because this school-going age group has potential of disseminating the strains far and wide in the communities. The carriage rate in our study is more than double that in a study on milk men in and around Kampala city, Uganda, where only 11 Staphylococci were isolated from 31 individuals, of which only 4/31 (13%) were S. aureus [26]. Our finding, however, is in agreement with statistics from literature suggesting that between 25 to 35% of healthy individuals are asymptomatically colonized with S. aureus in the nostrils [1, 21, 34, 35].

In the current study, the proportion of S. aureus isolates carrying the mecA gene, hence MRSA, was 48/73 (65.8%). This is generally high in comparison to community studies elsewhere. In the urban and peri-urban Kampala, the four isolates from milk men were all MSSA. In a study of indigenous remote African Babongo pygmies living in Waka National Parc, Central Gabon, all 34 S. aureus isolates were susceptible to Oxacillin/Methicillin, and did not amplify for the mecA gene by PCR. The authors hypothesized that the result could be due to limited use of antibiotics in that population. Studies in North America and Australia, however, have shown that native and indigenous populations have been associated with a high risk of colonization and infection with CA-MRSA which may be related to many of these groups being disadvantaged, due to their association with low socio-economic status, crowded living conditions and frequent use of antibiotics [36, 37]. There have been reports of absolute resistance to penicillin and high percentage resistance to other antibiotics in milk from similar settings in central Uganda, hence risk of spread to humans through the food chain [27]. In our study community there is frequent usage, by farmers, of veterinary antibiotics to treat nearly all ailments in their livestock, due to poor outreach services by the veterinary department. It is therefore more likely that constant contact with antibiotics for animal use, as well as consumption of raw milk and its products without observing drug withdrawal periods, as is the culture in this setting, are limiting future options for the management of multidrug resistant microorganisms in both humans and animals in pastoral communities of Uganda. Among the 48 isolates carrying the mecA gene, 39 (81.3%) were type V, the other being type IV or its subtypes (Tables 1 and 2). Type IV and V SCCmec elements have strongly been associated with stains causing MRSA infections in persons with no history of hospitalization, hence thought to be more related to CA-MRSA [38]. Furthermore, it has been shown that children may be at a higher risk of infection with SCCmec types IV and V, as well as PVL carrying strains, compared to adults [38, 39]. SCCmec type V, which comprised of 81.3% of the mecA positive strains in our study, is known to be rare in Europe and the United States [33], and only recently seen in Greece [40]. Because SCCmec type IV and V are known to be small and highly mobile elements, their dissemination in a community population may be most commonly by transfers of strains from carriers to other individuals or from MRSA strains to Methicillin Sensitive Staphylococcus aureus (MSSA strains), or even from coagulase-negative staphylococci strain to an MSSA strain [41]. However, a recent study on patients with surgical site infections (SSI) at the National Referral hospital in Mulago, Kampala, showed that SCCmec type V was the most predominant type [22], suggesting the presence of mixed CA-MRSA and HA-MRSA genotypes in hospital settings in Uganda.

A further characterization of the 36 isolates carrying the PVL gene in this study was done. SCCmecIV and PVL are known to be molecular markers associated with the emergence of CA-MRSA worldwide [10]. The proportion of strains carrying both the PVL and mecA genes was 25/36 (69.4%), while the mecA gene was detected in 23/37 (62.2%) of the isolates without the PVL gene. While there was no significant difference in mecA carriage, PFGE profiles for both PVL positive and negative isolates show that there were bigger clusters in PVL positive strains (Table 1) compared to PVL negative strains (Table 2). In fact, 18/37 (48.6%) of the PVL negative isolates were in clusters (Table 1) compared to 21/36 (58.3%) of PVL positive isolates, with the largest (B1) comprising of eight isolates (Table 1). These results suggest that PVL positive isolates in the study were more likely to be involved in chains of transmission compared to the PVL negative isolates. There has been recent evidence of familial spread of PVL carrying MSSA strains in Israel [42] and MRSA strains in Italy [43], while a previous study in Greece revealed that a unique clone of PVL-positive MRSA had spread in both the community and hospital settings, and was replacing older clonal types [44]. In Central Gabon, Africa, PVL-encoding genes were detected in 55.9% of study isolates, with authors concluding that the pygmies in that study faced a risk of developing necrotizing infections, due to the virulence characteristic of the PVL. The finding of a high proportion of isolates carrying both the PVL and mecA genes in our study may have considerable implications on future strategies for infection control in these underprivileged communities.

Among the PVL positive isolates, there was a predominant circulating Spa type, t318, comprising of nine of the 34 (26.5%) typable isolates. This is the first time the Spa type has been identified in a Ugandan setting. Moreover, seven strains (77.8%) of this Spa type carried the mecA gene, pointing to the presence of a potentially virulent methicillin resistant strain circulating in the community. This strain has been found to be pandemic, and mostly PVL positive, also denoted ST30 by multilocus sequence typing. It has recently been mapped to have originated from Australia and disseminated to Brazil, United States, South Africa and Western Europe [3]. However, it is not related to the USA300, belonging to ST8 and leading cause of CA-MRSA in the USA. It has been revealed that PVL is most frequent in pandemic CA-MRSA strains and certain MSSA lineages, including ST30, appear to be a reservoir of CA-MRSA [6]. Surprisingly, ST30 has also been isolated from Babongo pygmies of Gabon in Africa, who are known to have been separated from other humans over millennia ago [21]. In a study of MRSA in five African cities, only one strain of this type was isolated from Antananarivo, and none from Cassablanca (Moroco), Niamey (Niger), Dakar (Senegal) and Yaoundé (Cameroon). While SCCmec type V strains have been isolated from hospital settings in Uganda, none were Spa type t318 [22]. However, Spa type t645, the second most frequent type in our collection (20.6%), also common in Western Europe and the Middle East (http://​spaserver.​ridom.​de), was found to be the most frequent type isolated from SSI at Mulago National referral hospital [22], supporting the notion that there is a changing epidemiology reflected by community associated SCCmec genotypes being now more associated with hospital infections as observed elsewhere [23, 45]. However, it is noteworthy that some spa types of PVL+ S. aureus in the present study (t186, t729 and t355) have been previously identified in studies from Africa, and they all share the MLST type ST88 [46].